from:"Christian Gagneraud"

On Tue, 7 Apr 2020 at 02:00, Richard Weickelt  wrote:
> Hi,
>
> > Stop
>
> I think you might get more support from others if you stop spamming the
> mailing list right now. Looks like anybody sending emails to the list gets
> these kind of replies (me included). This way you just cause even more
> annoyance and make people just extending their "Roland Hughes filter".

Your message is spam according to gmail.
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Fwd: Fwd: Fwd: Fwd: 100% CPU core on a README.md

-- Forwarded message -
From: 6138848841 <6138848...@pcs.rogers.com>
Date: Tue, 7 Apr 2020 at 01:33
Subject: Re: [Qt-creator] Fwd: Fwd: Fwd: 100% CPU core on a README.md
To: 


Stop


Original Message, sent Tue, 7 Apr 2020 01:17:41 +1200:
> I won't stop forwarding every single messages until this
> is solved.
> -- Forwarded message -
> From:
> 6138848841 <6138848...@pcs.rogers.com>
> Date: Tue, 7 Apr 2
> 020 at 01:15
> Subject: Re: [Qt-creator] Fwd: Fwd: 100% CPU
>  core on a README.md
> To: 
> Stop
> Orig
> inal Message, sent Tue, 7 Apr 2020 01:04:41 +1200:
> > 
> -- Forwarded message -
> > From: 6138848841 <
> >
>  6138848...@pcs.rogers.com>
> > Date: Tue, 7 Apr 2020 at 00
> :57
> > Subject: Re: [Qt-creator] Fwd: 100% CPU core on a R
> EADME
> > .md
> > To: 
> > Stop
> > Original Me
> ssage, sen
> > t Tue, 7 Apr 2020 00:29:12 +1200:
> > > --
>  Forwarded
> > message -
> > > From: 6138848841 <
> > > 6138848841@pcs.r
> > ogers.com>
> > > Date: Tue, 7 Apr 2
> 020 at 00:03
> > > Subject: Re
> > : [Qt-creator] 100% CPU co
> re on a README.md
> > > T
> > > o:  > a...@gmail.com>
> > >
> Stop
> > > Original Message, sent Mon
> > > , 6
> > Apr 2020 23
> :34:36 +1200:
> > > > On Mon, 6 Apr 2020 at 17:4
> > >
> >  2,
> Eike Ziller  > > > io> wrote:
> > > > >
> > >
>  >
> >  > Whic
> > > h version of Qt Creator are you using?
> >
>  > > > E.
> > g. Qt Creato
> > > r 4.11.0 and .1 had https://
> bugreports.q
> > >
> >  > t.io/browse/QT
> > > CREATORBUG-23647
>  , fixed in 4.11.2
> > >
> > > 4.11.2.
> > > > Could be d
> > > u
> e to my clang-tidy settings,
> > defaults are 'th
> > > > oro
> ugh c
> > > hecks'
> > > > and 'build b/f
> >  analysis' is on.
> > > > Maybe i sh
> > > > o
> > > uld have a proje
> > ct overr
> ide.
> > > > Need to do more testing.
> > > >
> > >  C
> > > > hr
> > is
> > > > >
> > > > > Br, Eike
> > > > >
> > > > > > On Apr 4
> , 2020, at
> > 1
> > > 3:32, Chri
> > > > stian Gagneraud  n...@gmail.com> wro
> > te:
> > > >
> > > > >
> > > > > > On Sun,
> >
>  > >  5 Apr 2020 at 00:05, C
> > hristian Gagne
> > > raud  hgans@gmail.c
> > > > om> wrote:
> > > >
> > > >> Any known iss
> ue
> > > with MD? Or maybe with th
> > > > e la
> > nguage serv
> er (I use that
> > > > > >> for python)?
> > > > > >
> > >
> >  >
>  > > O
> > > > ops, sorry, the issu
> > > e is not caused by
> th
> > at file. I had a
> > > > nother QtC
> > > > > >
> > > open
>  w/ a 'brok
> > en' project.
> > > > > >
> > > > > > I need
> > >
>  >  to unde
> > > rstand
> >  a project that contains a USB f
> irmware and
> > > > > > W
> > > i
> > ndows dlls and exes sourc
> es, and the whole thing i
> > > > s
> > f
> > > or Visual
> > > >
>  > > studio.
> > > > > > So I created a top
> > level qm
> > >
> a
> > > > ke file, with ONLY
> > > > > > INCLUDEPATH
> > >
> >  >
>  > > QMAKE_CFLA
> > > GS
> > > > > > S
> > > > OURCES
> > > > > >
> > > > >
> >  > None of this can build, b
> > > /c it require
> s wi
> > > > ndow
> > s and a custom ARM
> > > > > > toolchai
> >
>  > n (I have none). Bu
> > t a
> > > > t least i can use QtC,
>  and some
> > > of the
> > > > > >
> > benefit of the
> > > >  co
> de model, which is what
> > > I only ne
> > ed.
> > > > > >
> > >
>  > > > So, is it
> > > >  expected/acceptable
> > >
> > , that
> QtC needs 100% of a core to co
> > > > pe
> > > > > > with
> >
>  unb
> > > uildable source code?
> > > > > >
> > > > > > I thin
> k that
> > > >  being a
> > > ble to use QtC for retro-engine
> ering from
> >  source
> > > > > > is
> > >  a nice capability.
> > > > > >
> > > > > >
> > Chris
> > > > > > _
> > >
> > _
> > _
> > > > > > Qt-crea
> tor ma
> > > iling
> > > >  list
> > > > > >
> >  Qt-creator@qt-p
> roject.org
> > > > > > https
> > > ://lists.qt-
> > >
> >  > proj
> ect.org/listinfo/qt-creator
> > > > >
> > > > > --
> > > > > E
> > ike Ziller
> > > > > Pr
> > > > incipal Software Engineer
> >
>  > > >
> > >
> >  >
> > >  > The Qt Company GmbH
> > > > > Erich
> > > > -Thilo-Straß
> > e 10
> > > > >
> > > D-12489 Berlin
> > >
>  > > eike.zil...@qt.io
> > > > >
> > > > http://qt.io
> > > >
> >
>  >  > Geschäftsführer: Mika Pälsi
> > ,
> > > > > Juha Var
> > > > elius,
> > > Mika Harjuaho
> > > > > Sitz de
> > r Gesel
> lschaft: Berlin, Reg
> > > > is
> > > tergericht: Amtsgeri
> >
> cht Charlottenburg, HRB 144331 B
> > > > >
> > > > __
> 
> > _
> > > > Qt-cr
> > >
> eato
> > > > r
> > mailing list
> > > > Qt-creator@qt-project.o
> rg
> > > > https
> > > :/
> > /lists.qt
> > > > -project.org/list
> info/qt-creator
> > > --
> > > Sen
> > t
> > > from my Rogers Wir
> eless Cellular Phone
> > > 
> > __
> > > __
> ___
> > > Qt-creator maili
> > ng list
> >
> > Qt-
> > > crea...@qt-project.org
> > > https://lists.qt
> > -
> project.org/listi
> > > nfo/qt-creator
> > --
> > Sent from my
> Rog
> > ers Wireless Cellular Phone

Re: [Qt-creator] Fwd: Fwd: Fwd: 100% CPU core on a README.md

On Tue, 7 Apr 2020 at 01:23, Sébastien Le Ray  wrote:
> Or you could just look at the List-Unsubscribe header to unsubscribe a
> list you asked to receive messages from…

No that's not the problem, i've been on this list for years (and i feel good).
The problem is that someone on this ml thinks it is fun to spam people
with unrelated messages (eg forward Qt mailing list message to random
people).

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Fwd: Fwd: Fwd: 100% CPU core on a README.md

I won't stop forwarding every single messages until this is solved.

-- Forwarded message -
From: 6138848841 <6138848...@pcs.rogers.com>
Date: Tue, 7 Apr 2020 at 01:15
Subject: Re: [Qt-creator] Fwd: Fwd: 100% CPU core on a README.md
To: 


Stop


Original Message, sent Tue, 7 Apr 2020 01:04:41 +1200:
> -- Forwarded message -
> From: 6138848841 <
> 6138848...@pcs.rogers.com>
> Date: Tue, 7 Apr 2020 at 00:57
> Subject: Re: [Qt-creator] Fwd: 100% CPU core on a README
> .md
> To: 
> Stop
> Original Message, sen
> t Tue, 7 Apr 2020 00:29:12 +1200:
> > -- Forwarded
> message -
> > From: 6138848841 <
> > 6138848841@pcs.r
> ogers.com>
> > Date: Tue, 7 Apr 2020 at 00:03
> > Subject: Re
> : [Qt-creator] 100% CPU core on a README.md
> > T
> > o:  a...@gmail.com>
> > Stop
> > Original Message, sent Mon
> > , 6
> Apr 2020 23:34:36 +1200:
> > > On Mon, 6 Apr 2020 at 17:4
> >
>  2, Eike Ziller  > > io> wrote:
> > > >
> > >
>  > Whic
> > h version of Qt Creator are you using?
> > > > E.
> g. Qt Creato
> > r 4.11.0 and .1 had https://bugreports.q
> >
>  > t.io/browse/QT
> > CREATORBUG-23647 , fixed in 4.11.2
> >
> > 4.11.2.
> > > Could be d
> > ue to my clang-tidy settings,
> defaults are 'th
> > > orough c
> > hecks'
> > > and 'build b/f
>  analysis' is on.
> > > Maybe i sh
> > > o
> > uld have a proje
> ct override.
> > > Need to do more testing.
> > >
> >  C
> > > hr
> is
> > > >
> > > > Br, Eike
> > > >
> > > > > On Apr 4, 2020, at
> 1
> > 3:32, Chri
> > > stian Gagneraud  wro
> te:
> > >
> > > >
> > > > > On Sun,
> > >  5 Apr 2020 at 00:05, C
> hristian Gagne
> > raud  > > om> wrote:
> > >
> > >> Any known issue
> > with MD? Or maybe with th
> > > e la
> nguage server (I use that
> > > > >> for python)?
> > > > >
> >
>  > > > O
> > > ops, sorry, the issu
> > e is not caused by th
> at file. I had a
> > > nother QtC
> > > > >
> > open w/ a 'brok
> en' project.
> > > > >
> > > > > I need
> > >  to unde
> > rstand
>  a project that contains a USB firmware and
> > > > > W
> > i
> ndows dlls and exes sources, and the whole thing i
> > > s
> f
> > or Visual
> > > > > studio.
> > > > > So I created a top
> level qm
> > a
> > > ke file, with ONLY
> > > > > INCLUDEPATH
> >
>  > > > QMAKE_CFLA
> > GS
> > > > > S
> > > OURCES
> > > > >
> > > >
>  > None of this can build, b
> > /c it requires wi
> > > ndow
> s and a custom ARM
> > > > > toolchai
> > n (I have none). Bu
> t a
> > > t least i can use QtC, and some
> > of the
> > > > >
> benefit of the
> > >  code model, which is what
> > I only ne
> ed.
> > > > >
> > > > > So, is it
> > >  expected/acceptable
> >
> , that QtC needs 100% of a core to co
> > > pe
> > > > > with
>  unb
> > uildable source code?
> > > > >
> > > > > I think that
> > >  being a
> > ble to use QtC for retro-engineering from
>  source
> > > > > is
> >  a nice capability.
> > > > >
> > > > >
> Chris
> > > > > _
> > > _
> _
> > > > > Qt-creator ma
> > iling
> > >  list
> > > > >
>  Qt-creator@qt-project.org
> > > > > https
> > ://lists.qt-
> >
>  > project.org/listinfo/qt-creator
> > > >
> > > > --
> > > > E
> ike Ziller
> > > > Pr
> > > incipal Software Engineer
> > > >
> >
>  >
> >  > The Qt Company GmbH
> > > > Erich
> > > -Thilo-Straß
> e 10
> > > >
> > D-12489 Berlin
> > > > eike.zil...@qt.io
> > > >
> > > http://qt.io
> > >
> >  > Geschäftsführer: Mika Pälsi
> ,
> > > > Juha Var
> > > elius,
> > Mika Harjuaho
> > > > Sitz de
> r Gesellschaft: Berlin, Reg
> > > is
> > tergericht: Amtsgeri
> cht Charlottenburg, HRB 144331 B
> > > >
> > > __
> _
> > > Qt-cr
> > eato
> > > r
> mailing list
> > > Qt-creator@qt-project.org
> > > https
> > :/
> /lists.qt
> > > -project.org/listinfo/qt-creator
> > --
> > Sen
> t
> > from my Rogers Wireless Cellular Phone
> > 
> __
> > _
> > Qt-creator maili
> ng list
> > Qt-
> > crea...@qt-project.org
> > https://lists.qt
> -project.org/listi
> > nfo/qt-creator
> --
> Sent from my Rog
> ers Wireless Cellular Phone
> _
> __
> Qt-creator mailing list
> Qt-creator@qt-
> project.org
> https://lists.qt-project.org/listinfo/qt-crea
> tor


--
Sent from my Rogers Wireless Cellular Phone
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Fwd: Fwd: 100% CPU core on a README.md

-- Forwarded message -
From: 6138848841 <6138848...@pcs.rogers.com>
Date: Tue, 7 Apr 2020 at 00:57
Subject: Re: [Qt-creator] Fwd: 100% CPU core on a README.md
To: 


Stop


Original Message, sent Tue, 7 Apr 2020 00:29:12 +1200:
> -- Forwarded message -
> From: 6138848841 <
> 6138848...@pcs.rogers.com>
> Date: Tue, 7 Apr 2020 at 00:03
> Subject: Re: [Qt-creator] 100% CPU core on a README.md
> T
> o: 
> Stop
> Original Message, sent Mon
> , 6 Apr 2020 23:34:36 +1200:
> > On Mon, 6 Apr 2020 at 17:4
> 2, Eike Ziller  > io> wrote:
> > >
> > > Whic
> h version of Qt Creator are you using?
> > > E.g. Qt Creato
> r 4.11.0 and .1 had https://bugreports.q
> > t.io/browse/QT
> CREATORBUG-23647 , fixed in 4.11.2
> > 4.11.2.
> > Could be d
> ue to my clang-tidy settings, defaults are 'th
> > orough c
> hecks'
> > and 'build b/f analysis' is on.
> > Maybe i sh
> > o
> uld have a project override.
> > Need to do more testing.
> >
>  C
> > hris
> > >
> > > Br, Eike
> > >
> > > > On Apr 4, 2020, at 1
> 3:32, Chri
> > stian Gagneraud  wrote:
> >
> > >
> > > > On Sun,
> >  5 Apr 2020 at 00:05, Christian Gagne
> raud  > om> wrote:
> > > >> Any known issue
> with MD? Or maybe with th
> > e language server (I use that
> > > >> for python)?
> > > >
> > > > O
> > ops, sorry, the issu
> e is not caused by that file. I had a
> > nother QtC
> > > >
> open w/ a 'broken' project.
> > > >
> > > > I need
> >  to unde
> rstand a project that contains a USB firmware and
> > > > W
> indows dlls and exes sources, and the whole thing i
> > s f
> or Visual
> > > > studio.
> > > > So I created a top level qm
> a
> > ke file, with ONLY
> > > > INCLUDEPATH
> > > > QMAKE_CFLA
> GS
> > > > S
> > OURCES
> > > >
> > > > None of this can build, b
> /c it requires wi
> > ndows and a custom ARM
> > > > toolchai
> n (I have none). But a
> > t least i can use QtC, and some
> of the
> > > > benefit of the
> >  code model, which is what
> I only need.
> > > >
> > > > So, is it
> >  expected/acceptable
> , that QtC needs 100% of a core to co
> > pe
> > > > with unb
> uildable source code?
> > > >
> > > > I think that
> >  being a
> ble to use QtC for retro-engineering from source
> > > > is
>  a nice capability.
> > > >
> > > > Chris
> > > > _
> > __
> > > > Qt-creator ma
> iling
> >  list
> > > > Qt-creator@qt-project.org
> > > > https
> ://lists.qt-
> > project.org/listinfo/qt-creator
> > >
> > > --
> > > Eike Ziller
> > > Pr
> > incipal Software Engineer
> > >
> >
>  > The Qt Company GmbH
> > > Erich
> > -Thilo-Straße 10
> > >
> D-12489 Berlin
> > > eike.zil...@qt.io
> > >
> > http://qt.io
> >
>  > Geschäftsführer: Mika Pälsi,
> > > Juha Var
> > elius,
> Mika Harjuaho
> > > Sitz der Gesellschaft: Berlin, Reg
> > is
> tergericht: Amtsgericht Charlottenburg, HRB 144331 B
> > >
> > ___
> > Qt-cr
> eato
> > r mailing list
> > Qt-creator@qt-project.org
> > https
> ://lists.qt
> > -project.org/listinfo/qt-creator
> --
> Sent
> from my Rogers Wireless Cellular Phone
> __
> _
> Qt-creator mailing list
> Qt-
> crea...@qt-project.org
> https://lists.qt-project.org/listi
> nfo/qt-creator


--
Sent from my Rogers Wireless Cellular Phone
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Fwd: 100% CPU core on a README.md

-- Forwarded message -
From: 6138848841 <6138848...@pcs.rogers.com>
Date: Tue, 7 Apr 2020 at 00:03
Subject: Re: [Qt-creator] 100% CPU core on a README.md
To: 


Stop


Original Message, sent Mon, 6 Apr 2020 23:34:36 +1200:
> On Mon, 6 Apr 2020 at 17:42, Eike Ziller  io> wrote:
> >
> > Which version of Qt Creator are you using?
> > E.g. Qt Creator 4.11.0 and .1 had https://bugreports.q
> t.io/browse/QTCREATORBUG-23647 , fixed in 4.11.2
> 4.11.2.
> Could be due to my clang-tidy settings, defaults are 'th
> orough checks'
> and 'build b/f analysis' is on.
> Maybe i sh
> ould have a project override.
> Need to do more testing.
> C
> hris
> >
> > Br, Eike
> >
> > > On Apr 4, 2020, at 13:32, Chri
> stian Gagneraud  wrote:
> > >
> > > On Sun,
>  5 Apr 2020 at 00:05, Christian Gagneraud  om> wrote:
> > >> Any known issue with MD? Or maybe with th
> e language server (I use that
> > >> for python)?
> > >
> > > O
> ops, sorry, the issue is not caused by that file. I had a
> nother QtC
> > > open w/ a 'broken' project.
> > >
> > > I need
>  to understand a project that contains a USB firmware and
> > > Windows dlls and exes sources, and the whole thing i
> s for Visual
> > > studio.
> > > So I created a top level qma
> ke file, with ONLY
> > > INCLUDEPATH
> > > QMAKE_CFLAGS
> > > S
> OURCES
> > >
> > > None of this can build, b/c it requires wi
> ndows and a custom ARM
> > > toolchain (I have none). But a
> t least i can use QtC, and some of the
> > > benefit of the
>  code model, which is what I only need.
> > >
> > > So, is it
>  expected/acceptable, that QtC needs 100% of a core to co
> pe
> > > with unbuildable source code?
> > >
> > > I think that
>  being able to use QtC for retro-engineering from source
> > > is a nice capability.
> > >
> > > Chris
> > > _
> __
> > > Qt-creator mailing
>  list
> > > Qt-creator@qt-project.org
> > > https://lists.qt-
> project.org/listinfo/qt-creator
> >
> > --
> > Eike Ziller
> > Pr
> incipal Software Engineer
> >
> > The Qt Company GmbH
> > Erich
> -Thilo-Straße 10
> > D-12489 Berlin
> > eike.zil...@qt.io
> >
> http://qt.io
> > Geschäftsführer: Mika Pälsi,
> > Juha Var
> elius, Mika Harjuaho
> > Sitz der Gesellschaft: Berlin, Reg
> istergericht: Amtsgericht Charlottenburg, HRB 144331 B
> >
> ___
> Qt-creato
> r mailing list
> Qt-creator@qt-project.org
> https://lists.qt
> -project.org/listinfo/qt-creator


--
Sent from my Rogers Wireless Cellular Phone
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] 100% CPU core on a README.md

On Mon, 6 Apr 2020 at 17:42, Eike Ziller  wrote:
>
> Which version of Qt Creator are you using?
> E.g. Qt Creator 4.11.0 and .1 had 
> https://bugreports.qt.io/browse/QTCREATORBUG-23647 , fixed in 4.11.2

4.11.2.
Could be due to my clang-tidy settings, defaults are 'thorough checks'
and 'build b/f analysis' is on.
Maybe i should have a project override.
Need to do more testing.

Chris



>
> Br, Eike
>
> > On Apr 4, 2020, at 13:32, Christian Gagneraud  wrote:
> >
> > On Sun, 5 Apr 2020 at 00:05, Christian Gagneraud  wrote:
> >> Any known issue with MD? Or maybe with the language server (I use that
> >> for python)?
> >
> > Oops, sorry, the issue is not caused by that file. I had another QtC
> > open w/ a 'broken' project.
> >
> > I need to understand a project that contains a USB firmware and
> > Windows dlls and exes sources, and the whole thing is for Visual
> > studio.
> > So I created a top level qmake file, with ONLY
> > INCLUDEPATH
> > QMAKE_CFLAGS
> > SOURCES
> >
> > None of this can build, b/c it requires windows and a custom ARM
> > toolchain (I have none). But at least i can use QtC, and some of the
> > benefit of the code model, which is what I only need.
> >
> > So, is it expected/acceptable, that QtC needs 100% of a core to cope
> > with unbuildable source code?
> >
> > I think that being able to use QtC for retro-engineering from source
> > is a nice capability.
> >
> > Chris
> > ___
> > Qt-creator mailing list
> > Qt-creator@qt-project.org
> > https://lists.qt-project.org/listinfo/qt-creator
>
> --
> Eike Ziller
> Principal Software Engineer
>
> The Qt Company GmbH
> Erich-Thilo-Straße 10
> D-12489 Berlin
> eike.zil...@qt.io
> http://qt.io
> Geschäftsführer: Mika Pälsi,
> Juha Varelius, Mika Harjuaho
> Sitz der Gesellschaft: Berlin, Registergericht: Amtsgericht Charlottenburg, 
> HRB 144331 B
>
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] 100% CPU core on a README.md

On Sun, 5 Apr 2020 at 03:56, André Pönitz  wrote:
> On Sun, Apr 05, 2020 at 01:57:39AM +1300, Christian Gagneraud wrote:
> > On Sun, 5 Apr 2020 at 01:44, ID-FOX -mike <1mike2...@gmail.com> wrote:
> > > I never seen thia site, why i get that fucking email?
> > > Il sab 4 apr 2020, 14:37 Christian Gagneraud  ha 
> > > scritto:
> > >> On Sun, 5 Apr 2020 at 01:24, ID-FOX -mike <1mike2...@gmail.com> wrote:
> > >> > Can you get out of my fucking mail?
> > >> https://lists.qt-project.org/listinfo/qt-creator
> >
> > Somebody is using your email account.
> > This is the act of a junevile wanna-be hacker (aka. script kiddie),
> > immature people with social problems.
> >
> > Can any QtC admin take care of this?
>
> The subscriptions work with confirmation mails.
>
> While the original subscribtion request can be (and are... triggered)
> by evil people, there's not much room for them to actually fake the
> confirmation mail.
>
> So the working hypothesis here is that '1mike2...@gmail.com' is one
> of those people who click at random links in random mails and now
> wonders why he got what he not asked for, but confirmed to have asked.

Since i am the only that have reported this kind of problem twice on
Qt MLs, what's possible is that someone forward my emails to random
'poor guys'.
If the goal is to annoy me, then i will simply ban any messages from
the 'poor guys'.
This way, whoever is trying to make a 'joke' just to annoy me, will
end up upsetting some random people *but not me*.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] 100% CPU core on a README.md

On Sun, 5 Apr 2020 at 01:44, ID-FOX -mike <1mike2...@gmail.com> wrote:
> I never seen thia site, why i get that fucking email?
> Il sab 4 apr 2020, 14:37 Christian Gagneraud  ha scritto:
>> On Sun, 5 Apr 2020 at 01:24, ID-FOX -mike <1mike2...@gmail.com> wrote:
>> > Can you get out of my fucking mail?
>> https://lists.qt-project.org/listinfo/qt-creator

Somebody is using your email account.
This is the act of a junevile wanna-be hacker (aka. script kiddie),
immature people with social problems.

Can any QtC admin take care of this?

Chris

PS:
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] 100% CPU core on a README.md

On Sun, 5 Apr 2020 at 01:24, ID-FOX -mike <1mike2...@gmail.com> wrote:
> Can you get out of my fucking mail?

https://lists.qt-project.org/listinfo/qt-creator
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] 100% CPU core on a README.md

On Sun, 5 Apr 2020 at 01:01, ID-FOX -mike <1mike2...@gmail.com> wrote:
> GET THE FUCK OUT MY FUCKING ACCPUNT BITCH
> Il sab 4 apr 2020, 13:05 Christian Gagneraud  ha scritto:
>> Hi there,

Hi 1mike2mad!

How are you today, did you take your pills this morning?

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] 100% CPU core on a README.md

On Sun, 5 Apr 2020 at 00:05, Christian Gagneraud  wrote:
> Any known issue with MD? Or maybe with the language server (I use that
> for python)?

Oops, sorry, the issue is not caused by that file. I had another QtC
open w/ a 'broken' project.

I need to understand a project that contains a USB firmware and
Windows dlls and exes sources, and the whole thing is for Visual
studio.
So I created a top level qmake file, with ONLY
INCLUDEPATH
QMAKE_CFLAGS
SOURCES

None of this can build, b/c it requires windows and a custom ARM
toolchain (I have none). But at least i can use QtC, and some of the
benefit of the code model, which is what I only need.

So, is it expected/acceptable, that QtC needs 100% of a core to cope
with unbuildable source code?

I think that being able to use QtC for retro-engineering from source
is a nice capability.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] 100% CPU core on a README.md

Hi there,

I have a markdown read me  open in QtC. After a minute, my (noisy) fan
started to go. It turns out that apparently it's due to me having a
'diagnostic' issue:
I'm using a triple back quote for a raw block, these backquotes and
the block content his yellow underlined, and QtC is burning a full CPU
core, non-stop.

Any known issue with MD? Or maybe with the language server (I use that
for python)?

Thanks,
Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] C++ indexing stucked

2020-04-02 Thread Christian Gagneraud

Hi there,

I'm having problems with code model and indexer, once a sqlite DB was
corrupted (I don't know how), but even after removing this DB I still
have issues with this sqlite DB.

Full log:
-
chgans@chgans-pc:~$ rm -f .local-qtc/QtProject/QtCreator.db
chgans@chgans-pc:~$ rm -Rf .cache/QtProject/
chgans@chgans-pc:~$ rm -Rf
Projects/build-neon-testapp-Desktop_Qt_Clang_9_64bit-Debug/
chgans@chgans-pc:~$
~/Programs/QtCommercial/Tools/QtCreator/bin/qtcreator
Projects/neon-testapp/CMakeLists.txt
chgans@chgans-pc:~$
~/Programs/QtCommercial/Tools/QtCreator/bin/qtcreator
Projects/neon-testapp/CMakeLists.txt
qtaccount.license: Reading Qt Account settings from disk:
"/home/chgans/.local/share/Qt/qtaccount.ini"
qtaccount.license: Reading licenses from file:
"/home/chgans/.local/share/Qt/qtlicenses.ini"
qtaccount.license: License added. 
qtaccount.license: Valid licenses found: 1
SOFT ASSERT: "current->id() != tc->id()" in file
/home/qt/work/build/qt-creator/src/plugins/projectexplorer/toolchainmanager.cpp,
line 211
SOFT ASSERT: "current->id() != tc->id()" in file
/home/qt/work/build/qt-creator/src/plugins/projectexplorer/toolchainmanager.cpp,
line 211
qtaccount.license: License added. 
qtaccount.license: Saving licenses to file:
"/home/chgans/.local/share/Qt/qtlicenses.ini"
SqliteStatement::stepStatement: contraint prevent insert or update!
"UNIQUE constraint failed: locations.sourceId, locations.line,
locations.column"
terminate called after throwing an instance of 'std::future_error'
  what():  std::future_error: No associated state
"RefactoringConnectionClient" LocalSocket Error: "QLocalSocket: Remote closed"
SqliteStatement::stepStatement: contraint prevent insert or update!
"UNIQUE constraint failed: locations.sourceId, locations.line,
locations.column"
-

This is a relatively small project and it builds fine for Android and
Linux desktop, both using clang.
As well, QtC takes ages to exit.

OS: Ubuntu-18.04
QtC:  4.11.2
Experimental plugins:
 - ClangRefactoring
Code model:
- ambiguous headers: no
- ignore pch: yes
- 5MB limit: yes
- diags: questionable constructs

Disabling ClangRefactoring plugin seems to fix the problem. Is this a
known issue?

Thanks,
Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Project explorer (was Re: Annoyance with 4.11)

2020-02-20 Thread Christian Gagneraud

On Fri, 21 Feb 2020 at 01:03, Tobias Hunger  wrote:
>
> On Thu, 2020-02-20 at 22:45 +1300, Christian Gagneraud wrote:
> 
>
> > The project explorer is a GUI component that assists the user with
> > graphically exploring a project.
>
> Actually the project view is meant to show what the build system thinks your
> project is. Such a view is necessary to allow for UI operations to manipulate
> the project.

I guess the problem comes from the convention we use, 1 build target
per directory/CMake file, hence why i have this feeling of un-needed
nesting.
A sub-dir = a project = a build target. but i get 3 nested nodes out
of that. It's logical on one hand, but feels overloaded on the other
hand.

Then there's the 'simplified' option, interesting at first, except
that it doesn't scale with big projects, I think it's flatten too
much, it starts with all the CMakefiles, i have over 100. And then the
build targets (which is good), but then it's followed by a massive
list of 'loose' headers. Which I think are headers used by the code,
but not mentioned in any CMake files. I really don't know how this
list is computed, on another (small) project, i have all QtSerialBus
related headers, including examples?!? I just discovered that these
files go away with the 'hide disabled files'.
Another thing i don't like, is that the list of source files
associated to a target is completely flatten, which is annoying when a
build target has a lot of files, organised in sub-directories, you
loose this natural grouping property.

My preference for the simplified view, would be to not show the build
files, and provide a list of build target at level1 (as currently
done), each expanding into a source tree that follows the storage
structure. And no extra loose files at the end.
that would be both simple and usable (in my case).

I would use this simplified view when i 'write code', and use the full
view when i work on the build system files.
When you think about it, generally developers don't care about the
build system files, they just want to add a class here and there and
use it somewhere else.

And there's a third use-case i think, which is the code review mode,
but in that mode, locator and code navigation shortcuts are king, so i
don't need the explorer much.

Hum... I'm sure someone already came with the idea of code review from
within QtC...

My 2 cents.
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Annoyance with 4.11

2020-02-20 Thread Christian Gagneraud

On Wed, 19 Feb 2020 at 21:52, Christian Kandeler
 wrote:
>
> On Wed, 19 Feb 2020 11:01:32 +1300
> Christian Gagneraud  wrote:
>
> > But my point was quite simple:
> > QtC 4.11 is broken, and has been released untested (or was knowingly
> > buggy), it doesn't seem to be anywhere near release quality (or i'm
> > really unlucky).
>
> This sounds quite hyperbolic, considering that your problem is just that the 
> project tree structure looks slightly different from what you'd prefer.

This sounds quite reverse hyperbolic. Like you're over-minimising the
impact of this bug.

For personal reason, i had to switch from a friendly 24/7 open source
user to a 8h/day, 5d/w commercial user.
The project explorer is a GUI component that assists the user with
graphically exploring a project.
The screenshots i sent recently clearly show a regression, not a
'preference issue'.

Please as well take a note of my comment about cmake plugin being
immature, my hate of this build system is orthogonal with the problem.
If it's not ready, it should be marked as such. There's no exemption.
Some of QtC plugins have been marked as experimental for 5+ years.

And IMHO the regression nature of this problem (what i refer to as an
'annoyance' in my original email) is backed by the fact that it was
fixed with QtC-4.12. You wouldn't fix a non-problem, would you?

What i'm surprised, and unhappy with is that there's no bug fix
available with the now-commercial-only Qt SDK installer, and my
current Qt install is broken (I'm paid to use QtC 8 hours a day, and
right now i cannot rely on your commercial offering, they just don't
work) . You rely on me to be an 'advanced' user. Unfortunately this
time is now gone.
The only option i got from the installer was to install an obsolete
QtC 4.11 preview version. WTF?!?
This doesn't look serious to me. You guys failed on that one, and it's
OK, shit happens.

If you're serious with getting the business up to scratch, you better
have to deliver commercial solutions up to scratch.

I am highly unimpressed with the 'commercial' solution so far.

With all due respect, please always take my emails with a pinch of
salt - But not too much, i really mean what i said! :P

Chris

PS: I was told about the fact that none of the 'business' guy listen
to this ML. Unfortunately my commercial leverage is close to minus
infinity.
We're a minority using QtC here, and given the 'bug' reported above,
and the reaction form the the Qt Company, my job of helping you with
getting developers adopt QtC in favour of, say MS VS or Android studio
is quite slim.

 I would look ridiculous to show them the screenshot i sent to you ***

Do you understand?
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Annoyance with 4.11

On Wed, 19 Feb 2020 at 11:01, Christian Gagneraud  wrote:
> Downloading http://download.qt.io/snapshots/qtcreator/4.12/4.12.0-beta1/248/
> right now, will see how far i can go with that.

Yes, this looks better (I ran qtcreator -temporarycleansettings). I
still see that some projects (build targets to be precise) have
'Header Files' and 'Source Files' as direct children, but some other
sub-projects don't. At least i don't have all the other crazy things
anymore.

That's goo news, but doesn't really solve my problem, i'm stuck with a
broken Qt SDK now. And it seems the only way to fix that is to install
QtC-4.10 separately, or run an separate QtC experimental build or
re-install everything :(

Will report on Jira with customer account.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Annoyance with 4.11

On Wed, 19 Feb 2020 at 11:01, Christian Gagneraud  wrote:
>
> > Could you please consider to try a 4.12 snapshot and check how things look
> > there?
>
> QtC-4.12 is not available in the 'Preview' components of the
> installer, the only option there is 4.11.0-rc1. I guess that's
> outdated stuff.
>
> I might try it, but I feel nervous, switching back and forth b/w QtC
> versions tend to break your config files. I can run QtC with a custom
> config dir, but then i'll have to configure all my toolchains, etc..
> which is a (bit of a) pain in my case.
>
> Downloading http://download.qt.io/snapshots/qtcreator/4.12/4.12.0-beta1/248/
> right now, will see how far i can go with that.
>
> But my point was quite simple:
> QtC 4.11 is broken, and has been released untested (or was knowingly
> buggy), it doesn't seem to be anywhere near release quality (or i'm
> really unlucky).

2 more points:
- The Qt installer doesn't provide roll-back methods, so now i'm stuck
with this new QtC version.
- If cmake integration is a feature in development, then the plugin
should be marked as experimental and be disabled by default. And so by
activating it, i would agree with a potential beta quality. I've done
that a lot, for everything clang related, and i never ranted when the
code model was pulling my machine down to it's knees (remember the
early days?).

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Annoyance with 4.11

> Could you please consider to try a 4.12 snapshot and check how things look
> there?

QtC-4.12 is not available in the 'Preview' components of the
installer, the only option there is 4.11.0-rc1. I guess that's
outdated stuff.

I might try it, but I feel nervous, switching back and forth b/w QtC
versions tend to break your config files. I can run QtC with a custom
config dir, but then i'll have to configure all my toolchains, etc..
which is a (bit of a) pain in my case.

Downloading http://download.qt.io/snapshots/qtcreator/4.12/4.12.0-beta1/248/
right now, will see how far i can go with that.

But my point was quite simple:
QtC 4.11 is broken, and has been released untested (or was knowingly
buggy), it doesn't seem to be anywhere near release quality (or i'm
really unlucky).
As an Open Source user i would accept that, but this is not acceptable
as a commercial user.
I am not a beta-tester, I am a client (well my employer is the client
really) *buying* a product and associated services. Which entitled me
(i think it's fair) to have quality expectation.

My apologies to the whole technical team for all the bad words, but
i'm quite upset on that one.
I'm pretty sure the problem lies in the business/management layer, not
the engineering one.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Annoyance with 4.11

Thanks Eike.

On Tue, 18 Feb 2020 at 23:33, Eike Ziller  wrote:

> Thanks for the notice, I’ve removed them.
>
> > On 18. Feb 2020, at 02:01, Christian Gagneraud  wrote:
> >
> > I received this weird email as an answer to my previous post. Apparently
> sent via the ML.
> > Not sure if other received it.
> >
> > See raw original email attached.
> >
> > -- Forwarded message -
> > From: HOAKA SWIMWEAR 
> > Date: Tue, 18 Feb 2020 at 13:55
> > Subject: Re: [Qt-creator] Annoyance with 4.11
> > To: Christian Gagneraud 
> >
> >
> > Hello Christian,
> > We are currently offline. We have received your message and one of our
> customer representatives will get back to you as soon as possible. Do not
> hesitate to leave more messages with more details.
> >
> > Allo Christian Gagneraud,
> > Nous sommes présentement hors ligne. Nous avons reçu ton message et une
> de nos responsables du service à la clientèle va vous répondre dès que
> possible. N'hésite pas à laisser plus de messages ci-bas.
> >
> > --
> >
> > HOAKA TEAM ❤️
> >
> > Christian Gagneraud wrote:
> >
> > Hi there,
> >
> > Just updated to 4.11, and with a CMake project, i've noticed (what is
> > for me) regressions in the project explorer.
> > 1. path to subdirectories are shown using backslash, which looks
> > really ugly on Linux
> > 2. this sub-dir path contains nested directory that you need to expand
> > one by one, so there's now one extra node to expand (the path node)
> > 3. now there are 'Header files' and 'Source files' sub nodes, that's
> > yet another nesting level that need to be expanded.
> > 4. sometimes, there's nested duplicated sub-nodes, see attachment
> > 5. All of the above seem to be random, on my project i have all of the
> > above but not always
> >
> > The above seems to be related to the introduction of 'CodeBlocks'
> generator?
> >
> > Is this a known issue? Anyone else having these issues? Am I the only
> > one to think that this is a bad UX regression?
> >
> > Thanks,
> > Chris
> > ___
> > Qt-creator mailing list
> > Qt-creator@qt-project.org
> > https://lists.qt-project.org/listinfo/qt-creator
> >
> >
> > <[Qt-creator] Annoyance with
> 4.11.eml>___
> > Qt-creator mailing list
> > Qt-creator@qt-project.org
> > https://lists.qt-project.org/listinfo/qt-creator
>
> --
> Eike Ziller
> Principal Software Engineer
>
> The Qt Company GmbH
> Erich-Thilo-Straße 10
> D-12489 Berlin
> eike.zil...@qt.io
> http://qt.io
> Geschäftsführer: Mika Pälsi,
> Juha Varelius, Mika Harjuaho
> Sitz der Gesellschaft: Berlin, Registergericht: Amtsgericht
> Charlottenburg, HRB 144331 B
>
>
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Annoyance with 4.11

On Tue, 18 Feb 2020 at 22:05, Tobias Hunger  wrote:
>
> On Tue, 2020-02-18 at 13:54 +1300, Christian Gagneraud wrote:
> > Hi there,
> >
> > Just updated to 4.11, and with a CMake project, i've noticed (what is
> > for me) regressions in the project explorer.
> > 1. path to subdirectories are shown using backslash, which looks
> > really ugly on Linux
> > 2. this sub-dir path contains nested directory that you need to expand
> > one by one, so there's now one extra node to expand (the path node)
> > 3. now there are 'Header files' and 'Source files' sub nodes, that's
> > yet another nesting level that need to be expanded.
> > 4. sometimes, there's nested duplicated sub-nodes, see attachment
> > 5. All of the above seem to be random, on my project i have all of the
> > above but not always

Hi Tobias,

First i would like to start positive: thanks a lot for all the effort,
for driving QtCreator to the edge of the technology, I enjoy QtCreator
every single day, and i'm always waiting impatiently for the next
release.

**BUT**, It seems that you guys at Qt (and KDAB, etc, ...) went
completely oblivion with CMake.
You decided to adopt the biggest piece of crap ever written by human
kind since the invention of the Perl language.
Just because: Hey! everyone is doing it, let's follow the rotten
rabbit hole that stink like a fish died there 3 years ago.
And now you kind-of telling me that i'm using QtC+CMake the wrong way!
I'm pretty tired with these 10x engineers comments.

OK, i'm done with it, so please find honest answers to your questions below.

> I would guess that you define source groups in your CMakeLists.txt files.

I don't, but as a recent developer survey in my company showed. 97% of
devs use Visual studio (the other 3% are not all using QtCreator BTW).
So yes *they* started to 'define source groups'.
I hate CMake from the bottom of my heart, and i don't understand how a
build system can mix FS layout (git repo), logic Layout (dev's view)
and build dependency (build system view). You got none of them right.
Obviously you can (or not, hence my email), but the question is: Are
you doing it right? (obviously you don't)

QtCreator Project explorer should be independent of these "views", I
should be presented with a 'brain friendly' view of my project, not
random rubbish dictated by cmake.

> Could you please consider to try a 4.12 snapshot and check how things look
> there?

Could you please consider releasing thoroughly tested SW to your
(paid) customers?
Do you have a SW QA team?
Have you heard of the UX (user experience) concept?

> A screenshot from 4.12 would help a lot since that changed the icon for source
> groups so that they are actually distinguishable from folders;-)

Thanks for the smiley, but in case you didn't noticed, i'm not in a
mood for that. You released a version that completely wreck havock my
productivity.
And now you're asking me to do beta testing?!?
How much are you willing to pay me for that? I've wasted nearly an
hour to update the installer (Qt licensing stuff), so that i can
update QtCreator, so that it can completely brake my 'developer
experience'.

My apologies, but i'm really loosing faith in the Qt eco-system. We're
going from bad to worth, every single day it seems.
Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Fwd: Annoyance with 4.11

2020-02-17 Thread Christian Gagneraud

I received this weird email as an answer to my previous post. Apparently
sent via the ML.
Not sure if other received it.

See raw original email attached.

-- Forwarded message -
From: HOAKA SWIMWEAR 
Date: Tue, 18 Feb 2020 at 13:55
Subject: Re: [Qt-creator] Annoyance with 4.11
To: Christian Gagneraud 

Hello Christian,
*We are currently offline.* We have received your message and one of our
customer representatives will get back to you as soon as possible. Do not
hesitate to leave more messages with more details.

Allo Christian Gagneraud,
*Nous sommes présentement hors ligne.* Nous avons reçu ton message et une
de nos responsables du service à la clientèle va vous répondre dès que
possible. N'hésite pas à laisser plus de messages ci-bas.

--

--

HOAKA TEAM ❤️
Christian Gagneraud wrote:

Hi there,

Just updated to 4.11, and with a CMake project, i've noticed (what is
for me) regressions in the project explorer.
1. path to subdirectories are shown using backslash, which looks
really ugly on Linux
2. this sub-dir path contains nested directory that you need to expand
one by one, so there's now one extra node to expand (the path node)
3. now there are 'Header files' and 'Source files' sub nodes, that's
yet another nesting level that need to be expanded.
4. sometimes, there's nested duplicated sub-nodes, see attachment
5. All of the above seem to be random, on my project i have all of the
above but not always

The above seems to be related to the introduction of 'CodeBlocks' generator?

Is this a known issue? Anyone else having these issues? Am I the only
one to think that this is a bad UX regression?

Thanks,
Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator
--- Begin Message ---
Hello Christian,
**We are currently offline.** We have received your message and one of our 
customer representatives will get back to you as soon as possible. Do not 
hesitate to leave more messages with more details.

Allo Christian Gagneraud,
**Nous sommes présentement hors ligne.** Nous avons reçu ton message et une de 
nos responsables du service à la clientèle va vous répondre dès que possible. 
N'hésite pas à laisser plus de messages ci-bas.
--

HOAKA TEAM ❤️

--

Christian Gagneraud wrote: | a écrit:

> Hi there,
> 
> Just updated to 4.11, and with a CMake project, i've noticed (what is
> for me) regressions in the project explorer.
> 1. path to subdirectories are shown using backslash, which looks
> really ugly on Linux
> 2. this sub-dir path contains nested directory that you need to expand
> one by one, so there's now one extra node to expand (the path node)
> 3. now there are 'Header files' and 'Source files' sub nodes, that's
> yet another nesting level that need to be expanded.
> 4. sometimes, there's nested duplicated sub-nodes, see attachment
> 5. All of the above seem to be random, on my project i have all of the
> above but not always
> 
> The above seems to be related to the introduction of 'CodeBlocks' generator?
> 
> Is this a known issue? Anyone else having these issues? Am I the only
> one to think that this is a bad UX regression?
> 
> Thanks,
> Chris
> 
> Qt-creator mailing list
> Qt-creator@qt-project.org
> https://lists.qt-project.org/listinfo/qt-creator

--- End Message ---
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Annoyance with 4.11

2020-02-17 Thread Christian Gagneraud

Hi there,

Just updated to 4.11, and with a CMake project, i've noticed (what is
for me) regressions in the project explorer.
1. path to subdirectories are shown using backslash, which looks
really ugly on Linux
2. this sub-dir path contains nested directory that you need to expand
one by one, so there's now one extra node to expand (the path node)
3. now there are 'Header files' and 'Source files' sub nodes, that's
yet another nesting level that need to be expanded.
4. sometimes, there's nested duplicated sub-nodes, see attachment
5. All of the above seem to be random, on my project i have all of the
above but not always

The above seems to be related to the introduction of 'CodeBlocks' generator?

Is this a known issue? Anyone else having these issues? Am I the only
one to think that this is a bad UX regression?

Thanks,
Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Announcement: New UI Git plugin for QtCreator

2019-12-16 Thread Christian Gagneraud

On Mon, 16 Dec 2019 at 22:01, Francesc Martinez  wrote:
>
> Hi all,
>
> Today I've announced in the Planet that I've released a plugin to integrate 
> Git in QtCreator. I know that a long while ago there was someone asking to 
> integrate one of the Qt Git apps into the IDE and I'm confident to said that 
> this is done. And with GPLv2 license, of course!
>
> I've you have not seen the post in the planet you can find it here:
> https://www.francescmm.com/new-git-plugin-for-qtcreator/
>
> In addition, you can download the plugin directly from GitHub:
> https://github.com/francescmm/GitQlientPlugin/releases/tag/v1.0.0
>
> I've compiled and tested it in Windows and Linux (main OS).
>
> I'll be happy if you can provide feedback about improvements, fixes, etc.
>
> I hope you don't see this as pure spam since I just want to give back 
> something to the Qt community after all this years of learning.

Looking at the screenshots, this looks interesting, will try it tomorrow.
I didn't quite get, what's the common code base with? Is it related to 'qgit'?
I tend to prefer git in command line (eg. 'git lola' gives you this
'guitar hero' feeling [1]), but i sometimes use qgit, and QtC's own
git plugin.
My main grief with QtC git integration is the python language server
kicking in with error messages each time i open a git textual view.
I actually rarely use QtC to deal with my git workflow ( git fetch +
git rebase, or just git rebase -i), cmake is causing so much churn on
each rebase (PS: This is not a cmake rant email).
I usually prefer to switch to console mode for my git operations and
then wait for a minute or two, and switch back to QtC when the rebase
got 'digested'.
Frozen UI is the worst UX. I got to a point where i prefer to manage
my important builds manually, and only switch to QtC once the checkout
stabilise (cmake, clang code model, unit test scanner, ...).

Just random feedback really (and a bit OT)... will give it a try.

Chris

[1] http://blog.kfish.org/2010/04/git-lola.html

>
> Francesc.
> ___
> Qt-creator mailing list
> Qt-creator@qt-project.org
> https://lists.qt-project.org/listinfo/qt-creator
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Weird version reporting

2019-08-29 Thread Christian Gagneraud

Hi there,

Not a blocker at all for me, not even an annoyance, but i just wonder
what is going on.

I have Qt SDK installed, with the latest stable QtC along the current "preview".
I have 2 4.9.2 instances open, from the task bar (KUbuntu-18.04), one
instance is reported as "QtC 4.10.0-rc1 (see
https://imgur.com/jT2cHpl)".
But when i go to check the version in the about dialog, both instances
report as 4.9.2 (Built on Jun 26 2019, rev 8181...).
I launch QtC from the shell, and i indeed launch the stable version.

Has anyone seen this kind of behaviour?

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Turn off scan for tests per project?

2019-07-17 Thread Christian Gagneraud

On Thu, 18 Jul 2019 at 11:39, Andy  wrote:
>
> Thanks Chris.
>
> I neglected to give version info - I'm using 4.9.2 on macOS 10.14.
>
> The project I'm trying to work with uses Unreal Engine which seems to have 
> problems in most IDEs. I just can't work with XCode, so I thought I'd try to 
> get it working in Qt Creator...
>
> If they want to stress test Qt Creator, they should grab the Unreal source 
> code :-)

Do you have a URL?
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Turn off scan for tests per project?

2019-07-17 Thread Christian Gagneraud

On Thu, 18 Jul 2019 at 10:18, Andy  wrote:
>
> Is there a way to turn off the scan for tests for a specific project?
>
> I'm opening a huge cmake project and it takes minutes to do this scan on 
> every single cmake change.
>
> (I can turn it off completely, but I'm looking to just turn it off on this 
> one project.)

AFAIK, there's no way to disable it on a per project basis. I use to
have perf issue with it, but i think the problem was solved with QtC
4.8.
I find QtC to be pretty bad when handling CMake based projects where
you need to edit the CMake files. This constant reloading and UI
freeze is really irritating.
Hope QtC 4.10 will bring improvement on that.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Sanitizer integration

2019-06-27 Thread Christian Gagneraud

Hi there,

I recently gave a try at sanitzers
(https://github.com/google/sanitizers), and it's pretty awesome.
Most (all?) problems are reported in the same format as compiler
diagnostics, some come with complementary report that looks very
similar to valgrind ones.
enabling sanitizer is normally as simple as adding eg
-fsanitize=address,undefined to compiler and linker flags ("undefined"
stands for undefined behaviour sanitizer).

Would be nice to have this integrated into the "Analyze" framework, so
that when i run my instrumented app, i get a nice sanitizer issue view
of the "debug" mode. with source location lookup.
The only problem, i met so far, was source location points to inline
code in system header, with no stack trace. sometimes stack trace are
reported, so it would be nice to have a view similar to the debugger
one.

Has anyone thought about this?
It shouldn't be that hard (tm)

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Problem with clang code model when cross-compiling with clang

2019-06-27 Thread Christian Gagneraud

Hi there,

QtCreator v4.9.1, from online installer.
x86_64 Ubuntu 18.04
clang-8 from https://apt.llvm.org
target: armv7a-custom-linux (g++-5.3.0)

As an experiment, i'm currently cross-compiling our project with clang-8.
Since then, the clang-code model is broken.
I've noticed these asserts and error messages:
"The command \"/usr/bin/clang++-8\" terminated with exit code 1."
SOFT ASSERT: "false && "__cplusplus is not predefined, assuming latest
C++ we support."" in file
/home/qt/work/build/qt-creator/src/plugins/projectexplorer/toolchain.cpp,
line 327

Followed by a few of these (that we've always had, so here for the
sake of completeness)
C++ Indexer: Skipping file "..." because it is too big.

And Finally:
Parsing /current/source/file/in/editor.cpp failed: CXError_Failure
Reparsing /current/source/file/in/editor.cpp failed: 3

And the whole code model is broken (no completion, type hierarchy,
file/class outline, symbol lookup, etc...)

I'm using cmake, with a custom (but simple) toolchain file. And i can
cross build our project (from command line and fromQtC), and run the
exe on the target.

For QtCreator i had to do one "special" things.
Since clang is multi target, the compiler is auto detected as x86_64,
so i had to select the "" ABI, and set it to
arm-linux-generic-elf-32bits.
I then cloned the original kit (a gcc-5 cross kit), and simply change
the compiler settings and the toolchain file.

The cmake toolchain file doesn't do anything fancy apart from setting
CMAKE_CROSSCOMPILING to true and CMAKE_CXX_COMPILER_TARGET to our gnu
triplet.
cmake sets clang's -target and --sysroot, which works out of the box.
I do not need to do crazy things with include/lib search path.
The only trick in there is forcing the linker, since it defaults to
/usr/bin/ld, and ld is not clever enough to pick up our cross env, so
i ended up adding "-luse-ld=lld-8"  which is "sysroot aware"
(/usr/bin/lld-8).
I can build an hello world with just
clang++-8 -target arm-navico-linux-gnueabi -march=armv7-a
--sysroot=$SYSROOT test.cpp

Point is: this toolchain is valid and functional.

For the sake of completeness i have (and had for a while) these other
soft asserts:
SOFT ASSERT: "type == "INTERNAL" || type == "UNINITIALIZED"" in file
/home/qt/work/build/qt-creator/src/plugins/cmakeprojectmanager/cmakeconfigitem.cpp,
line 153


Any idea what is going on? Is it a problem with my setup, is this
setup supported by QtCreator? any special log domain i could enable to
troubleshoot that?

Thanks,
Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] JSON wizard: build systems and sub-projects

2019-05-13 Thread Christian Gagneraud

On Mon, 13 May 2019 at 20:55, Tobias Hunger  wrote:
>
> Hi Christian,
>
> reading my last mail again it was rather short and sounds harsher than 
> intended.
> Sorry for that. I did not intend to slight you or your work, it is just that I
> get very lazy when composing mails on my phone.

I never felt offended or whatever. I tend to be short and sharp and
I'm too impatient sometimes.

> On Mon, 2019-05-13 at 10:38 +1200, Christian Gagneraud wrote:
> > On Mon, 13 May 2019 at 05:27, Tobias Hunger  wrote:
> I wanted to point out that the wizard has a hard-coded list of project types 
> it
> supports. Creator also has a list of build systems it supports, which is 
> dynamic
> and depends on the plugins that are loaded.
>
> At the "Define Build System" page that dynamic part is the important one: The
> wizard should not offer build systems that are not supported by the current
> Creator instance, even though it would support these build systems in other
> creator instances.
>
> So IMHO the wizard does the right thing here by relying on the loaded plugins
> and not on the supportedProjectTypes.

I might be wrong, but maybe you're into your generic world where every
kind of plugin can offer every kind of feature.
I'm a bit more pragmatic, all that I see is that there is a one to one
relationship b/w the project managers, their project type and their
underlying build systems.
QmakeProjectManager plugin -> QmakeProjectManager.QmakeProject project
-> "qmake" build tool
QbsProjectManager plugin -> QbsProjectManager.QbsProject project ->
"qbs" build tool
CMakeProjectManager plugin -> CMakeProjectManager.CMakeProject project
-> "cmake" build tool

These 3 combinations are the only valid ones out of a 3x3x3 permutations.

I haven't look into the other ones, autotools, bare metal, ...

I'm fine with giving up this idea, you guys know your tool more than I
do. I'm not familiar enough with the code base.

> > Currently JSON wizards for subprojects are not working, only the
> > built-in qmake one work (and AFAIU they need to go).
> > So something has to be done to solve this issue.
>
> I am aware of that and appreciate you looking into this problem.
>
> This will need more than just the wizard part to work though. E.g. CMake has 
> no
> way to add new files into existing CMakeLists.txt files, so a wizard will not 
> be
> able to do much in that case.

Troll Detector: WARNING! WARNING!

Well, so be it! We shouldn't hold qmake and qbs b/c cmake is inferior

Troll Detector: All Clear!

I'm sure the cmake guys will come with a solution later on, in the
mean time: let's move on.
If a particular project template only support Qbs and Qmake, it just
needs to declare it in it's supportedProjectType/enable properties,
isn't it how it is supposed to work? :)

> > What i'm proposing here, is to forget about these 3 facets, and use
> > only one way to talk about projects: The ProjectTypeId.
>
> I got that.
>
> But you will need to check whether a plugin is currently loaded and you will
> need to have a user-visible name for that build system and you need the. Where
> do you get that information from?
>
> Do you offer a registry in one of the plugins to hold that information
> somewhere?

availableProjectTypes and parentProjectTypes (when applicable) are the
user-visible project-type *providers*.

In my ideal world, i do not care about plugin loaded or not, if a
plugin is not loaded, it's associated project type is not available.
Is this a correct assumption?

Then, the "registry" is actually implicit, all the json wizards
already hard-code all of these:
 - plugin names (CMakeProjectManager, ...)
 - plugin namespace'd project types (QmakeProjectManager.QmakeProject)
 - local namespace'd build tools (qbs, ...)

Again, i might be wrong.

Maybe i should come with an ideal subdirs.json and discuss how realistic it is.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] JSON wizard: build systems and sub-projects

2019-05-12 Thread Christian Gagneraud

On Mon, 13 May 2019 at 10:38, Christian Gagneraud  wrote:
> One last comments, sorry to sound negative, but currently all project
> templates are a massive pile of copy/paste, with complex and error
> prone macro/javascript expressions.
> That is clearly a sign that something is "far from ideal".
> I'm not moaning or anything, i'm really just trying to help this
> wizard thing going. I can't wait to be able to do with qbs/cmake what
> i can currently do only with qmake.

I'm sure i'm not the first one to think that the wizard could use
python instead of JSON.
I guess the idea with JSON was to have something declarative (and
decoupled from QtC internal?), but the introduction of macro and
javascript shows that there is a need for logic too.
Python could help here.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] JSON wizard: build systems and sub-projects

2019-05-12 Thread Christian Gagneraud

On Mon, 13 May 2019 at 05:27, Tobias Hunger  wrote:
>> 2. "Define Build System" page.
>> Currently the boiler plate attempt to identify the "Project type" to
>> generate, but doesn't rely/use the semantic of the
>> "supportedProjectType" above.
>> Instead they use the HasPlugin concept to map the build system tool
>> (aka, "qmake", "qbs", "cmake"), which is actually equivalent to the
>> project type (the wizard doesn't care about the tool, it cares about
>> the project manager it targets: CMakeProjectManager.CMakeProject,
>> Qbs.QbsProject, ...).
>
>
> Sure. Disable qmake and qbs support in Creator and see what happens:-)

!?!

What i'm trying to say here, is that for each project manager, the
wizard plays with 3 information, let's take qmake:
PluginId => QmakeProjectManager
ProjectTypeId => QmakeProjectManager.QmakeProject (as defined by the plugin)
BuildSystem => "qmake" (which could be "some random alias" since this
is internal to the wizard, and is just used to multiplex the
generator)

Currently JSON wizards for subprojects are not working, only the
built-in qmake one work (and AFAIU they need to go).
So something has to be done to solve this issue.

What i'm proposing here, is to forget about these 3 facets, and use
only one way to talk about projects: The ProjectTypeId.
I do not care about the fact that 'QmakeProjectManager' plugin
provides 'QmakeProject', and that it uses "qmake" as a build system.
What i care, is that my JSON wizard can generate stuff for
ProjectTypeIdX, ProjectTypeIdY and ProjectTypeIdZ, all the rest is
implementation details.

Injected into wizard:
 - availableProjectTypes
 - parentProjectType (in case of sub-project)
Selected by user:
 - ProjectType (in case of top project, aka. %{BuildSystem})
Defined by wizard:
 - supportedProjectType
 - enabled = !intersection(availableProjectTypes,
supportedProjectType).isEmpty()

This would make the writing of project wizards IMHO more consistent,
but would require quite some changes, i agree with that.

And concerning the case "Sub project from main menu", Wouldn't it be
better to ask for the parent project on the first page, not the last
one?
Actually, maybe the first page should only be about project parenting
and build systems, since they are strongly coupled

>
> 
>
>> 5. Ideas to achieve that
>> I did some quick hack in the ProjectNode, there is a "buildKey()"
>> there, that returns the build target key, so i added a
>> "projectTypeKey()" that returns the project manager's project ID
>> (CMakeProjectManager.CMakeProject, ...), and which is injected wizard
>> environment.
>
>
> That could help when starting the wizard via the context menu. That is 
> unfortunately not the only way to trigger this wizard.

Yes, there are already variables that are only defined when started
via the context menu. we could add more, couldn't we?

>
>> With the above and proper macro nesting, it should be possible to
>> define the following wizard options:
>>  - IsSubProject, IsTopProject
>>  - HasCMake, HasQmake, HasQbs
>>
>>  - IsCMake, IsQmake, IsQbs
>
>
> Please do not hardcore the list of possible build systems! There are more 
> plugins out there 5han those that are shipped with Creator.

I'm talking about the "options" section of a wizard, this information
is already hard coded into the wizard, the above list is just a
convenience naming sheme, eg:
options: [
 {key: "HasCMake", "value": "%{JS:
value('Plugins').indexof('CMakeProjectManager') >= 0}"},
 {key: "IsSubProject", value: "%{Exists:ProjectExplorer.Profile.Ids}"}
 {key: "IsTopProject", value: "%{JS: !value('IsSubProject')}"}
 ...
]

the IsQmakeProject, etc... case is a bit trickier, as there are 2 sub-cases:
- Triggered from main menu: User's choice if top project, parent's
project type in case of sub-project
- Triggered from context menu: injected by the project explorer (as
described above)

One last comments, sorry to sound negative, but currently all project
templates are a massive pile of copy/paste, with complex and error
prone macro/javascript expressions.
That is clearly a sign that something is "far from ideal".
I'm not moaning or anything, i'm really just trying to help this
wizard thing going. I can't wait to be able to do with qbs/cmake what
i can currently do only with qmake.

chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] JSON wizard: build systems and sub-projects

2019-05-12 Thread Christian Gagneraud

Hi there,

In my attempt to add a unified "subdirs" project and make all projects
templates "subproject" friendly, I discovered a few grey area, and
would like to discuss what they are how we could fix them.

1. "supportedProjectType" and "enabled" properties are currently
redundant (in the way they are used), consider:
"supportedProjectTypes": [ "CMakeProjectManager.CMakeProject",
"Qbs.QbsProject", "Qt4ProjectManager.Qt4Project" ],
and
"enabled": "%{JS: value('Plugins').indexOf('QbsProjectManager') >= 0
|| value('Plugins').indexOf('QmakeProjectManager') >= 0 ||
value('Plugins').indexOf('CMakeProjectManager') >= 0}",

They both express the same thing, but using different approaches. IMHO
enabled should be used to express further conditions on top of the
supported project types.

2. "Define Build System" page.
Currently the boiler plate attempt to identify the "Project type" to
generate, but doesn't rely/use the semantic of the
"supportedProjectType" above.
Instead they use the HasPlugin concept to map the build system tool
(aka, "qmake", "qbs", "cmake"), which is actually equivalent to the
project type (the wizard doesn't care about the tool, it cares about
the project manager it targets: CMakeProjectManager.CMakeProject,
Qbs.QbsProject, ...).
This page also suffers from selecting an index as a default, instead of a value.
It would be nice to be able to write:
"default": "%{ProjectTypeKey},
"items":
[
{
"trKey": "qmake",
"value": "%{QmakeProjectTypeKey}",
"condition": "%{HasQmake}"
},
...
]


3 "generator"
They both need, in the case of the "subdirs" project, to know if it is
a top/sub project, and which project type (build system) it is for.
Ideally, i would like to be able to write:
{
"source": "subdirs.qbs",
"target": "%{QbsProjectFile}",
"openAsProject": "%{IsTopProject}",
"openInEditor": "%{IsSubProject}",
"condition": "%{IsQbsProject}"
},
...

4. "Summary" page

Similar to the "build system" page, there is a need to select the
default sub-project, and disable the combo box.

5. Ideas to achieve that
I did some quick hack in the ProjectNode, there is a "buildKey()"
there, that returns the build target key, so i added a
"projectTypeKey()" that returns the project manager's project ID
(CMakeProjectManager.CMakeProject, ...), and which is injected wizard
environment.

With the above and proper macro nesting, it should be possible to
define the following wizard options:
 - IsSubProject, IsTopProject
 - HasCMake, HasQmake, HasQbs
 - IsCMake, IsQmake, IsQbs
  - top project:
by comparing the "Build system" page choice
  - sub-project
by comparing the injected "projectType" (same values as above)

The HasFoo, IsFoo and ProjectType could all rely on the project
manager key. No need to deal with project manager key, plugin names
and build tool names.

My 2 cents.
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] work-flow/gerrit: branch with mixed reviews

2019-05-10 Thread Christian Gagneraud

On Thu, 9 May 2019 at 11:21, Christian Gagneraud  wrote:
>
> Hi there,
>
> I'm working on a couple of fix/improvement, and i need to work on a
> branch (from master) that require a cherry-pick from 4.9, and on top
> of that 3 patches that need their own review.
> So far, i'm doing lot of manual work to keep my work-flow and gerrit
> happy, but surely there are better way to work. For example is it
> possible to push only one commit to gerrit?
> Say my stack is:
> - master
> - cherry-pick1-from-4.9
> - review1
> - review2
> - wip
>
> I'm currently using interactive rebase and/or transient branches to
> place the correct commit on top of master before pushing to gerrit.
> But that is a pain.
>
> Any tip on how to manage that in an efficient and convenient way?

I just found out how that works! (Sorry for the dumb question)

Actually gerrit track all the reviews with Change-Id, so it's possible
to have a stack of commits, and push the branch to gerrit, all
associated reviews will get updated. Neat!

Now i have to sort out the reverse operation: How to pull the
said-branch, when it has changed on gerrit.
Typical use case is with 2 clones: say, one at work and one at home.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] work-flow/gerrit: branch with mixed reviews

2019-05-10 Thread Christian Gagneraud

On Fri, 10 May 2019 at 22:37, Christian Gagneraud  wrote:
>
> On Thu, 9 May 2019 at 11:21, Christian Gagneraud  wrote:
> >
> > Hi there,
> >
> > I'm working on a couple of fix/improvement, and i need to work on a
> > branch (from master) that require a cherry-pick from 4.9, and on top
> > of that 3 patches that need their own review.
> > So far, i'm doing lot of manual work to keep my work-flow and gerrit
> > happy, but surely there are better way to work. For example is it
> > possible to push only one commit to gerrit?
> > Say my stack is:
> > - master
> > - cherry-pick1-from-4.9
> > - review1
> > - review2
> > - wip
> >
> > I'm currently using interactive rebase and/or transient branches to
> > place the correct commit on top of master before pushing to gerrit.
> > But that is a pain.
> >
> > Any tip on how to manage that in an efficient and convenient way?
>
> I just found out how that works! (Sorry for the dumb question)
>
> Actually gerrit track all the reviews with Change-Id, so it's possible
> to have a stack of commits, and push the branch to gerrit, all
> associated reviews will get updated. Neat!
>
> Now i have to sort out the reverse operation: How to pull the
> said-branch, when it has changed on gerrit.
> Typical use case is with 2 clones: say, one at work and one at home.

As well, i still have to discover how to track dependencies b/w
qt-creator and the qbs submodule, i have a qt-creator commit that
depends on a qbs commit...

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] work-flow/gerrit: branch with mixed reviews

2019-05-08 Thread Christian Gagneraud

Hi there,

I'm working on a couple of fix/improvement, and i need to work on a
branch (from master) that require a cherry-pick from 4.9, and on top
of that 3 patches that need their own review.
So far, i'm doing lot of manual work to keep my work-flow and gerrit
happy, but surely there are better way to work. For example is it
possible to push only one commit to gerrit?
Say my stack is:
- master
- cherry-pick1-from-4.9
- review1
- review2
- wip

I'm currently using interactive rebase and/or transient branches to
place the correct commit on top of master before pushing to gerrit.
But that is a pain.

Any tip on how to manage that in an efficient and convenient way?

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] JSON Wizard broken?

2019-05-06 Thread Christian Gagneraud

On Mon, 6 May 2019 at 22:57, Christian Gagneraud  wrote:
> On Mon, 6 May 2019 at 19:32, Eike Ziller  wrote:
> > > On 4. May 2019, at 06:40, Christian Gagneraud  wrote:
> > > Some json wizard rely on this to deduce top level project vs subproject.
> > > eg.
> > > IsTopLevelProject: "%{JS: !'%{Exists:ProjectExplorer.Profile.Ids}'"
> > >
> > > And then, this deducing is actually broken (even w/o the above bug).
> > > Since ProjectExplorer.Profile.Ids would be set only from the context
> > > menu, and from the context menu on a project node you cannot create a
> > > top level project, only subprojects.
> >
> > Without the above bug, that would work. IsTopLevelProject would correctly 
> > be false when the wizard was triggered from the context menu, and true 
> > otherwise.
>
> "would work", "would correctly be false", the reality is that "would"
> doesn't happen for a lot of reasons.
> The main issue is Qt'c JSON Wizard's "bool from variant" and QtC's
> macro expansion system doesn't mix well with javascript concept of
> true or false.
> Using the debugger i saw things such as a string like "true || (some
> more unexpanded/unevaluated stuff related to qtc plugins)" being
> evaluating as Wizard's True b/c the it is not an empty string or not
> equal to the string "false".
> I even saw strings like "JS Error: blah, blah, blah" being evaluated
> as true b/c it's not empty/null/"false"...

https://bugreports.qt.io/browse/QTCREATORBUG-22411
https://codereview.qt-project.org/#/c/260852/

It's a one character change, yet it was not easy to find which json
file it was. :(

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] JSON Wizard broken?

2019-05-06 Thread Christian Gagneraud

On Mon, 6 May 2019 at 19:32, Eike Ziller  wrote:
> > On 4. May 2019, at 06:40, Christian Gagneraud  wrote:
> > Some json wizard rely on this to deduce top level project vs subproject.
> > eg.
> > IsTopLevelProject: "%{JS: !'%{Exists:ProjectExplorer.Profile.Ids}'"
> >
> > And then, this deducing is actually broken (even w/o the above bug).
> > Since ProjectExplorer.Profile.Ids would be set only from the context
> > menu, and from the context menu on a project node you cannot create a
> > top level project, only subprojects.
>
> Without the above bug, that would work. IsTopLevelProject would correctly be 
> false when the wizard was triggered from the context menu, and true otherwise.

"would work", "would correctly be false", the reality is that "would"
doesn't happen for a lot of reasons.
The main issue is Qt'c JSON Wizard's "bool from variant" and QtC's
macro expansion system doesn't mix well with javascript concept of
true or false.
Using the debugger i saw things such as a string like "true || (some
more unexpanded/unevaluated stuff related to qtc plugins)" being
evaluating as Wizard's True b/c the it is not an empty string or not
equal to the string "false".
I even saw strings like "JS Error: blah, blah, blah" being evaluated
as true b/c it's not empty/null/"false"...

> So, yes the behavior (and which variables are available etc) of the JSON 
> wizards in case of subdirs is broken (afaict since quite a while).
> It is on the list of things I want to investigate in the nearer future.
>
> Note that there are more issues, like
> - If you start the wizard from the main menu you can still add it as a 
> subproject on the summary page - but you where asked for the build system 
> before, and a list of kits. As far as I can tell, if multiple build systems 
> support subdirs (e.g. qmake + qbs), then you’ll be suggested e.g. qmake 
> projects even if you selected qbs as the build system for your new project.

I'm fine with that one, this is minor annoyance (UX repetition). As
long as the end goal is achieved (project manager implementation being
notified of the "sub-project addition/removal")

> - If the Build System page is hidden in case of a subdir project wizard, no 
> build system is selected, and the filters in the file generation section are 
> broken. So e.g. if the wizard was triggered for creating a subproject, it 
> must be given some information about the build system that it should create 
> files for.

I naively tried to "solve" that one, and gave up rather quickly. This
is IMHO more worrisome.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] JSON Wizard broken?

2019-05-03 Thread Christian Gagneraud

Hi there,

According to the documentation:

The following information is only available when the wizard was
triggered via the context menu of a node in the Projects view:
- InitialPath with the path to the selected node.
- ProjectExplorer.Profile.Ids contains a list of Kits configured for
the project of the selected node.

But it turns out that this is true for 'InitialPath' but not for
'ProjectExplorer.Profile.Ids'. (Confirmed by debugging the
m_expander.registerPrefix("Exists",...) in JsonWizard::JsonWizard())

Some json wizard rely on this to deduce top level project vs subproject.
eg.
IsTopLevelProject: "%{JS: !'%{Exists:ProjectExplorer.Profile.Ids}'"

And then, this deducing is actually broken (even w/o the above bug).
Since ProjectExplorer.Profile.Ids would be set only from the context
menu, and from the context menu on a project node you cannot create a
top level project, only subprojects.

As well some wizards use '"enabled": "%{JS: ! %{IsSubproject}}", but
it seems that IsSubproject is only set in the summary page.

I now have my "subdirs" project template sorted, i now just need to
know if the project is top-level or not, so that i can set either
"openAsProject" or "openInEditor" to true in my generator data.

For a subproject, openInEditor has to be set to true, other wise this
will trigger
https://github.com/qt-creator/qt-creator/blob/master/src/plugins/projectexplorer/jsonwizard/jsonwizard.cpp#L477

Can't we have IsTopLevelProject/IsSubproject set all the time, for any
page, and both from context menu or main menu?

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Custom wizard vs qmake's wizards

2019-05-01 Thread Christian Gagneraud

On Tue, 30 Apr 2019 at 18:30, André Pönitz  wrote:
>
>
> I did not say anything about "banning" build system specific wizards,
> on the contrary, I said

Sorry that was my feeling reading your answers.

>
>  "It doesn't dictate that. If there's a need to add a build system specific
>features, suitable callback/virtual function/property/whatever can
>be added on demand."
>
> This includes build-system specific wizards.

All good then. :)

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Custom wizard vs qmake's wizards

2019-05-01 Thread Christian Gagneraud

On Tue, 30 Apr 2019 at 19:57, Christian Kandeler
 wrote:
>
> On Tue, 30 Apr 2019 12:02:24 +1200
> Christian Gagneraud  wrote:
>
> > Christian, any thoughts on adding this capability to Qbs core API?
>
> Sure, that was the idea all along. We just didn't get around to do it.
>
> > How hard is it to handle adding/removing references?
>
> Modifying the project file is easy. Updating the build graph internals 
> directly (like we do for adding/removing files, for performance reasons) is 
> probably tricky, but we could (and perhaps should, to be on the safe side) 
> "cheat" and simply re-resolve the project.
>

Looking for reviewers:
- Qbs:
https://codereview.qt-project.org/#/c/260217/
I definitely need some help on that one
- QtC/Qbs integration:
https://codereview.qt-project.org/#/c/260218/
- QtC Subdirs json/custom new wizard template:
https://codereview.qt-project.org/#/c/260219/

Chris

>
> Christian
>
> ___
> Qt-creator mailing list
> Qt-creator@qt-project.org
> https://lists.qt-project.org/listinfo/qt-creator
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Custom wizard vs qmake's wizards

On Tue, 30 Apr 2019, 00:19 André Pönitz,  wrote:

> Christian Gagneraud  wrote:
>
> > you risk to end up will all supported build systems being limited
> > (least common denominator).
>
> That risk is manageable by looking for (and finding) abstraction on
> the right level, and it doesn't have to be common denominator.
>

I don't think that such ideas should be banned, they should just be
monitored tightly. I agree with the risk of fragmentation and duplication,
but there are valid usecases for build system specific wizards.
And I don't think it applies to qbs only. I'm sure cmake has it's own needs
too.
Just think about new commers for example, they start with point and click,
and then as they learn the underlying system they might choose to edit bs
files manually.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Custom wizard vs qmake's wizards

On Mon, 29 Apr 2019 at 21:50, Eike Ziller  wrote:
> > On 29. Apr 2019, at 09:43, Christian Kandeler  
> > wrote:
> >
> > On Mon, 29 Apr 2019 18:34:09 +1200
> > Christian Gagneraud  wrote:
> >
> >>> If I register a new "Qbs.AddLibrary" action, i end up with duplicates
> >>> in the context menu (qmake disabled, and qbs enabled),
>
> Calling setVisible instead of setEnabled for these actions would of course be 
> possible, but it would probably be better to try and not duplicate all the 
> work.

I think that was the wrong track, i should ignore all these "Qt4" qmake things.
This "Add Library" is to add a dependency, and at this stage, i can
manage that manually.
I just want helpers that would create qbs skeleton code for a
"composite" project.
I wrote a "Qbs Subdir Project" custom wizard, I now just need to be
able to chain that.

> >> Actually, I just saw the supportsAction in the project node. That's a good
> >> start. But there doesn't seem to be anything generic to handle adding
> >> library, etc
> >> Maybe it's not hard to add, maybe not...
>
> That might be the way to go, but I also can’t really tell how much work that 
> would be. E.g. would it make sense to share (parts of) the actual wizard?

AFAIU, it should be managed in the ProjectExplorer::ProjectNode implementation.
https://github.com/qt-creator/qt-creator/blob/master/src/plugins/projectexplorer/projectnodes.h#L63
https://github.com/qt-creator/qt-creator/blob/master/src/plugins/projectexplorer/projectnodes.h#L326

It is hooked up on the context menu "New Subproject...". "Add new..."
is already supported by qbs project manager, ...

The difficulty is to add support for editing the "references" property
of a Project item:
https://github.com/qbs/qbs/blob/master/src/lib/corelib/api/project.h#L159

I quickly hacked the QtCreator side (easiest part), and added stubs to
the Qbs API. the "New subproject..." is now enabled and gives me the
"New project" wizard, it seems that the list of available "Custom
Wizard" based items is filtered by "supported build system", this is
pretty cool so far. I can even see my custom "Qbs Subdir Project"
wizard.

Christian, any thoughts on adding this capability to Qbs core API? How
hard is it to handle adding/removing references?
Given 
https://github.com/qbs/qbs/blob/master/src/lib/corelib/api/project.cpp#L496,
this doesn't look trivial to me, but shouldn't be that hard for
someone who knows the internals very well... ;)

Chris

Attached are a diff for Qtc that just enable the feature and forward
stuff to Qbs API and another one for Qbs that just add stubs.
diff --git a/src/lib/corelib/api/project.cpp b/src/lib/corelib/api/project.cpp
index 3ffd6b2e9..590c52e60 100644
--- a/src/lib/corelib/api/project.cpp
+++ b/src/lib/corelib/api/project.cpp
@@ -569,6 +569,8 @@ void ProjectPrivate::addFiles(const ProductData &product, const GroupData &group
 }
 }

+void ProjectPrivate::addReference(const ProjectData &project, const QString &filePath) {}
+
 void ProjectPrivate::removeFiles(const ProductData &product, const GroupData &group,
  const QStringList &filePaths)
 {
@@ -638,6 +640,9 @@ void ProjectPrivate::removeGroup(const ProductData &product, const GroupData &gr
 QBS_CHECK(removed);
 }
 }
+
+void ProjectPrivate::removeReference(const ProjectData &project, const QString &filePath) {}
+
 #endif // QBS_ENABLE_PROJECT_FILE_UPDATES

 void ProjectPrivate::removeFilesFromBuildGraph(const ResolvedProductConstPtr &product,
@@ -1303,6 +1308,21 @@ ErrorInfo Project::addFiles(const ProductData &product, const GroupData &group,
 }
 }

+ErrorInfo Project::addReference(const ProjectData &project, const QString &filePath)
+{
+try {
+QBS_CHECK(isValid());
+d->prepareChangeToProject();
+d->addReference(project, filePath);
+d->internalProject->lastStartResolveTime = FileTime::currentTime();
+d->internalProject->store(d->logger);
+return {};
+} catch (ErrorInfo errorInfo) {
+errorInfo.prepend(Tr::tr("Failure adding references to project."));
+return errorInfo;
+}
+}
+
 /*!
  * \brief Removes the given files from the given product.
  * If \c group is a default-constructed object, the files will be removed from the product's
@@ -1350,6 +1370,22 @@ ErrorInfo Project::removeGroup(const ProductData &product, const GroupData &grou
 return errorInfo;
 }
 }
+
+ErrorInfo Project::removeReference(const ProjectData &project, const QString &filePath)
+{
+try {
+QBS_CHECK(isValid());
+d->prepareCh

Re: [Qt-creator] Custom wizard vs qmake's wizards

On Mon, 29 Apr 2019 at 21:50, Eike Ziller  wrote:
> > On 29. Apr 2019, at 09:43, Christian Kandeler  
> > wrote:
> > On Mon, 29 Apr 2019 18:34:09 +1200
> > Christian Gagneraud  wrote:
> >>> If I register a new "Qbs.AddLibrary" action, i end up with duplicates
> >>> in the context menu (qmake disabled, and qbs enabled),
> Calling setVisible instead of setEnabled for these actions would of course be 
> possible, but it would probably be better to try and not duplicate all the 
> work.
>
> >>> if instead if
> >>> fetch the "Qt4.AddLibrary" and enable it, i get the (qmake) "Add
> >>> library wizard", but when i go till the end of the wizard i get qmake
> >>> stuff added to my qbs project file ... (that was funny, i expected a
> >>> crash or something :))
> >>>
> >>
> >> Actually, I just saw the supportsAction in the project node. That's a good
> >> start. But there doesn't seem to be anything generic to handle adding
> >> library, etc
> >> Maybe it's not hard to add, maybe not...
>
> That might be the way to go, but I also can’t really tell how much work that 
> would be. E.g. would it make sense to share (parts of) the actual wizard?

Well, maybe the project explorer shouldn't bother about implementation
details of the underlying project managers.
Why hard-code the concept of sub-projects, library, plugin, unit test,
file, project, non-qt project, ... into the project explorer?
Or even hard-code the qmake view of what a "library" is (internal,
external, system and pkg-config).
Both Qbs and CMake have different view of this and they handle that in
a better way.
So maybe the first page of the wizard could be shared (i'm not even
conviced of that), but then the next pages are very likely to be BS
dependent (BS=build-system, not bullshit :)).
For Qbs, an internal library is "as simple" as 'Depends { name:
"myInternalLibrary"}', no need to specify include path, etc (as
required by qmake and present in the current wizard), CMake work in a
very similar way AFAIU (eg public/private export).
qmake is the worst common denominator of "better" BS.

I'm not that familiar with CMake and other build system, but if the
project explorer attempts to dictate hard abstraction concepts, you
risk to end up will all supported build systems being limited (least
common denominator).
Currently the list of "Build system" plugins (project managers) are:
- autotools
- cmake
- clang compile db
- generic (does that include python projects?)
- qbs
- qmake

We can safely rule out clang db (as it is read-only AFAIK), autotools
(welcome to 21st century) and maybe generic (too generic?).
qmake is EOL, but will need support for years to come. CMake is
widespread and qbs a challenger with interesting potential (personal
opinion). And python is about to bring it's own lot of weird
requirements.

Just a wild idea, but what about having only one context menu entry:
"Add new item...", that would yield a re-organised "Add new file
and/or project and/or what-have-you" wizard under tight control of the
project manager implementation?
Maybe it's a good idea to share small bits, but not whole wizards.
Custom JSON wizards are cool for adding project manager independent
(re)sources, but i'm not convinced that they are good to handle
project manager/build system dependent features.

As far as Qbs is concerned here are the most likely candidates for
"Add new.." that are not files (source or not):
https://doc.qt.io/qbs/qbsconvenienceitems-qmlmodule.html
Note: I'm not saying we need support for all of these

My 2 cents.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Custom wizard vs qmake's wizards

On Mon, 29 Apr 2019 at 19:44, Christian Kandeler
 wrote:
>
> On Mon, 29 Apr 2019 18:34:09 +1200
> Christian Gagneraud  wrote:
>
> > > If I register a new "Qbs.AddLibrary" action, i end up with duplicates
> > > in the context menu (qmake disabled, and qbs enabled), if instead if
> > > fetch the "Qt4.AddLibrary" and enable it, i get the (qmake) "Add
> > > library wizard", but when i go till the end of the wizard i get qmake
> > > stuff added to my qbs project file ... (that was funny, i expected a
> > > crash or something :))
> > >
> >
> > Actually, I just saw the supportsAction in the project node. That's a good
> > start. But there doesn't seem to be anything generic to handle adding
> > library, etc
> > Maybe it's not hard to add, maybe not...
>
> I'll let people more knowledgeable about wizards answer the main question, 
> but to provide some context:
> Keep in mind that Qt Creator started with only qmake support, and things were 
> generalized one by one "on demand". What you are seeing here is something 
> that's probably still in the original state.

Yes, I get that, and it's good to see things moving in the right direction.

> Note also that "New Subproject..." is quite weird these days because it only 
> exists for qmake, yet most of the wizards that pop up let you choose the 
> build system for the sub-project, which makes no sense. (Ok, in theory you 
> could mix build systems, but let's not get too silly.)

True, i was thinking in the bus, what about adding a Qbs only entry to
the project pane context menu, like "Add a new Qbs Item...".
This would yield to a wizard alike of "add a new C++ class". Item type
being a list of the Qbs convenience items (and later on, with
possibility for custom item).
I like the generic (qbs-wise) approach, but there would be a need to
filter out the choice list (eg, cannot add a product to a product,
...)
Maybe the parameter action could help here (not familiar with that code).

Thinking loud really.

My use case is simple and easy to understand: right now if i start a
project from scratch, i need to create all the initial qbs skeletons
(just Project/Product w/ references). Then i can open the top qbs file
and start customising the whole thing, by editing Qbs files from
within QtC and use add new, add existing, ... I'm really after
something simple, just to bootstrap a "complex" project without having
to fire emacs.

I have virtually no use of the "Create an app/lib/whatever project",
most of the time, i want to create a "composite" project, with a
layout to my tastes (eg. w/ apps, libs, plugins, tests).

Would you accept a new "Subdirs Project" custom wizard template? It's
really just a few lines of json, along with "Project{}" as a qbs
template. Then i would have to check how to handle adding subprojects
and products.
I haven't figure out the Qbs file parser/writer, maybe that is the
blocking point, but QtC can already edit the files property of a
product, i have hope to find code that can do more than that.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Custom wizard vs qmake's wizards

2019-04-28 Thread Christian Gagneraud

On Mon, 29 Apr 2019, 18:01 Christian Gagneraud,  wrote:

>
> If I register a new "Qbs.AddLibrary" action, i end up with duplicates
> in the context menu (qmake disabled, and qbs enabled), if instead if
> fetch the "Qt4.AddLibrary" and enable it, i get the (qmake) "Add
> library wizard", but when i go till the end of the wizard i get qmake
> stuff added to my qbs project file ... (that was funny, i expected a
> crash or something :))
>

Actually, I just saw the supportsAction in the project node. That's a good
start. But there doesn't seem to be anything generic to handle adding
library, etc
Maybe it's not hard to add, maybe not...

Any thoughts?

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Custom wizard vs qmake's wizards

2019-04-28 Thread Christian Gagneraud

Hi there,

I was looking at the new custom wizards thing (QtC master), and i saw
that there's provision for build system selection, looks good.

I would like to investigate how to add this thing for Qbs, along with
add Application/Library/Plugin. I'm not after anything fancy, the idea
is that i should be able to bootstrap a project without having to
write the initial (almost empty) qbs files.
As well i'm not planning to address corner cases, and other "battery included".

Adding a qbs subdir to the "New project" wizard is easy with custom
wizard, but i can't make my head around how to hook on the
project/subproject context menu/command.

As well could someone clarify the mechanism behind the (greyed out for
abs) context menu entries of a "Project" qbs item in the project
explorer? ("New subproject", "Add library").

If I register a new "Qbs.AddLibrary" action, i end up with duplicates
in the context menu (qmake disabled, and qbs enabled), if instead if
fetch the "Qt4.AddLibrary" and enable it, i get the (qmake) "Add
library wizard", but when i go till the end of the wizard i get qmake
stuff added to my qbs project file ... (that was funny, i expected a
crash or something :))

Qmake project manager install it's own wizards using
IWizardFactory::registerFactoryCreator, is this the only way to go?
Any reason why there's no templates/wizards/projects/subdirproject templates?

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Object pool is gone?

2019-04-28 Thread Christian Gagneraud

Hi there,

I just noticed that the object pool is gone, QtC seems to use
singletong all around the place now.

Could someone share the rational behind getting rid of the object pool
and replace it with "almost-under-control" singletons.
Just an example, the `FileIconProviderImplementation *instance()`.
Isn't it bad that the first one who calls one of the FileIconProvider
namespaced function will control it's creation time?

OK, this example is bad as the FileIconProvider doesn't depends on
anything else.
How do you deal with the risk of inter-dependencies and initialisation
order (and loops, ...)?

I really would love to hear about the design decision, from a
technical point of view.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Dealing with python projects

2019-04-25 Thread Christian Gagneraud

On Thu, 25 Apr 2019 at 21:05, Cristián Maureira-Fredes
 wrote:
> On 4/25/19 1:18 AM, Christian Gagneraud wrote:
> > On Thu, 25 Apr 2019 at 04:29, Jason H  wrote:
> >>
> >> Check out this thread: 
> >> https://lists.qt-project.org/pipermail/qt-creator/2019-March/007865.html
> > Thanks, missed that one.
> >
> >> I also requested documentation on .pyproject and why a standard python 
> >> project format was not used.
> > Do you know a "standard python project" format?
>
> AFAIK the closer "Python project file" accepted by the community is the
> one introduced by PEP-518 https://www.python.org/dev/peps/pep-0518/
> I have been playing around with .toml files, but I don't know if that
> feature will be included in QtCreator soon.

I'm not familiar with that PEP, but it seems to be related to
packaging/deployment, not source, tests, execution, format/style, ...

> > I had a quick look at Visual code documentation [1], they support
> > choosing your interpreter (inc. from your venv), debugging, ...
> > But it seems to require you to write "supporting" JSON files.
>
> At the moment you can change the interpreter in QtCreator
> in the Run Configuration, but we are working on a nicer selection
> process, and maybe some auto-detection.

Interesting, it's only available for "Qt for python projects" (aka
.pythonproject).

May I suggest the idea that it should be part of
"Tools->Options...->Python", and be customisable in "project mode ->
project settings"?
That would be IMHO more powerful. But might require lot of work

The risk with the current approach is to put all python related stuff
into a god module that dictates everything.
That would be better to make the python language a first-class citizen
first. And make sure it can deal with project specific virtualenv.
/me dreaming again

> > And BTW, about my static analyser thing (pycodestyle, flake8), it is
> > very likely coming from the language server, creator "just" display
> > the issues reported by pyls.
>
> IMHO QtCreator needs to provide an option to select between flake8,
> black, and yapf, but at the moment it's just flake8.

Not sure. Maybe users could decide of their configuration if QtC fully
supports virtualenvs.
I can fully control my python language server via requirements.txt and
setup.cfg (inc. flake vs black vs yapf).
But wait! There's more! :)
McCabe, Rope, ... QtC cannot catch-up with this fast moving system,
better let the user decide what he/she wants.

If only i could teach QtC to "load" all the python-related settings
from my venv+pip+setup.cfg...

I think it would be a mistake to lead users into thinking that with
QtC there's no need to use virtualenv.
Even for a pyside project, virtualenv should be made mandatory. it's
IMHO a question of good practice (and sanity)

> > And I think that using Qbs to manage a python project has some
> > interesting potential for now
>
> Agree, but it's just a matter of spending some time working on it :/

I've been playing around this afternoon, and one of the main issue is
that Qbs was designed for compiled languages.
For a moment i thought about making Qbs understand python bytecode compilation.
I think that this source/build/install paradigm is both a strength and
a weakness when it comes to a python project.
It's all about how much integration you can make.
In python, build = make a package (eg, setup.py) and install means
install that package along dependencies (venv/pip), so the concept
still apply at the macro level.

>
> > Chris
> > [1] https://code.visualstudio.com/docs/python/python-tutorial
>
>
> Thanks for your feedback Chris, and Jason,
> hopefully you can help QtCreator to improve on the Python support.
> I will be more than happy to review your patches.

The python project i'm working on has already dragged me into adding a
new feature to clazy, i can't afford to get dragged into QtC and/or
Qbs. Never say never tho.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Dealing with python projects

2019-04-24 Thread Christian Gagneraud

On Thu, 25 Apr 2019 at 04:29, Jason H  wrote:
>
> Check out this thread: 
> https://lists.qt-project.org/pipermail/qt-creator/2019-March/007865.html
Thanks, missed that one.

> I also requested documentation on .pyproject and why a standard python 
> project format was not used.
Do you know a "standard python project" format?

> I think it isn't meant to be a proper python IDE, just one good enough for 
> PySide. Which is a shame.
That's IMHO a good start. I hope we'll have support for debugging
soon. As well - might be dreaming -, maybe the class and include
hierarchies will work someday. I'm not familiar with the language
server thing, but hopefully the work can be done there, and creator
just need to query it.

I had a quick look at Visual code documentation [1], they support
choosing your interpreter (inc. from your venv), debugging, ...
But it seems to require you to write "supporting" JSON files.

And BTW, about my static analyser thing (pycodestyle, flake8), it is
very likely coming from the language server, creator "just" display
the issues reported by pyls.

And I think that using Qbs to manage a python project has some
interesting potential for now

Chris
[1] https://code.visualstudio.com/docs/python/python-tutorial
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Dealing with python projects

2019-04-24 Thread Christian Gagneraud

On Wed, 24 Apr 2019 at 21:45, Christian Gagneraud  wrote:
> On Wed, 24 Apr 2019 at 19:20, Christian Kandeler  
> wrote:
> > On Wed, 24 Apr 2019 12:49:19 +1200 Christian Gagneraud  
> > wrote:
> As well i've noticed some sort of "static analyser" behaviour for
> python within QtC, I know QtC picks up my .clang-tidy when i do C++,

Amending: With QtC 4.9, i think it is .clang-format, not .clang-tidy.
There's a new entry for that in the contextual menu of the C++ editor,
something like "Show me where the rules come from" or something like
that, sorry i don't have a setup handy right now.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Dealing with python projects

2019-04-24 Thread Christian Gagneraud

On Wed, 24 Apr 2019 at 19:20, Christian Kandeler
 wrote:
>
> On Wed, 24 Apr 2019 12:49:19 +1200
> Christian Gagneraud  wrote:
>
> > So I recently switched to using Qbs. It currently allows me to
> > organise my python files, and QtCreator is happy with this "broken"
> > qbs project (broken=cannot be built).
> > One interesting feature is that i can configure QtC to run the test
> > webserver using a simple wrapper script. And with a project type of
> > "python" (or whatever is not recognised by qbs) QtC skip the build
> > process.
>
> I don't think you need to do that. If there's no cpp sources or libraries 
> around, the build process is a no-op for an Application as well.
> By the way, if your main python file has a shebang at the top, you can get a 
> run configuration for it by simply tagging it as an application. (It involves 
> a copy because source artifacts cannot be target artifacts in qbs atm, but 
> that's not a limitation in practice).
> For instance:
>
> -
> import qbs.File
>
> Application {
> Group {
> name: "main script"
> files: "pythontest.py"
> fileTags: "python"
> }
> Rule {
> inputs: "python"
> Artifact {
> filePath: input.baseName + "-exe.py"
> fileTags: "application"
> }
> prepare: {
> var cmd = new JavaScriptCommand();
> cmd.description = "creating " + output.fileName;
> cmd.sourceCode = function() { File.copy(input.filePath, 
> output.filePath); };
> return cmd;
> }
> }
> }
> -
>
> If you open this project in Qt Creator, Ctrl+R will run your python script.

Hi Christian,

Thanks a lot for the tip. I'll have to investigate further.
The python project i'm working on is django based [1], and it requires
me to run a particular script ("./manage.py runserver"), *in the
source directory*, within a virtual python env (hence my wrapper
script). I just started this (qtc+qbs) experiment today, and right now
i'm more interested by the "QtC experience" than the "Qbs experience".
By that i mean that i want my IDE to help me while developing:
add/remove/rename files (works better than CMake), intelligent code
completion (code model, see below), but ideally "run" and "test" steps
as well. "test" being way more important to me than "run" (by a huge
factor). I have 2 kind of tests: unit tests ("./manage.py test", light
&fast) and acceptance tests ("./manage.py behave", heavy & slow), this
is classic django stuff really.

As well i've noticed some sort of "static analyser" behaviour for
python within QtC, I know QtC picks up my .clang-tidy when i do C++,
but I wonder about python and my setup.cfg (pycodestyle+flake8), is it
recognised/supported by QtC? I definitely get insights related to my
pycodestyle+flake8 config... Looks good to me!
Downside is that code completion depends on my pip/virtualenv, so i
need to (re)start QtC within my venv shell each time i add a package
to it... :(

Chris

[1] https://www.djangoproject.com/
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] Dealing with python projects

2019-04-23 Thread Christian Gagneraud

Hi there,

I've started to use QtCreator for some of my python-based projects.
I've configured a python language server [1], and the result is quite
encouraging.

I've tried "Import Existing Project" that creates a
".creator" project file but is imho too basic and C/C++
oriented.
I saw the '.pyproject' option, but i couldn't find any documentation.
It seems to be tailored for python+Qt.
So I recently switched to using Qbs. It currently allows me to
organise my python files, and QtCreator is happy with this "broken"
qbs project (broken=cannot be built).
One interesting feature is that i can configure QtC to run the test
webserver using a simple wrapper script. And with a project type of
"python" (or whatever is not recognised by qbs) QtC skip the build
process.
What would be nice now is to have my unit tests integrated... :)
My test command can generate JUnit XML files.

I'm curious to know how other users deal with python projects, fancy
share your workflow?

Chris

[1] https://github.com/palantir/python-language-server
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] 4.9: gtest and cmake annoyance

2019-04-06 Thread Christian Gagneraud

On Sun, 7 Apr 2019 at 04:36, Cristian Adam  wrote:
>
> Hi,
>
> Please uncheck the "Autorun CMake" checkbox in settings.
> See https://doc.qt.io/qtcreator/creator-project-cmake.html

Thanks, it worked. But I guess it means that i now have to handle the
cmake re-run manually.
So it sounds like a bug to me. Do you want me to create a ticket, or
is it a known issue?

From the manual:
Note: When the Autorun CMake check box is enabled, Qt Creator
automatically runs CMake when you make changes to CMakeLists.txt
files. We recommend that you do not disable this check box, because
that would eventually make Qt Creator and the project disagree on what
the project information is.

Thanks,
Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] 4.9: gtest and cmake annoyance

2019-04-05 Thread Christian Gagneraud

Hi there,

Using the latest 4.9 beta, i've noticed that each time i edit and save
a unit test qtcreator re-run cmake, as my project is quite big, this
adds a (rather long) delay in my testing process. This makes QtC
almost unusable when working on unit tests.
Is this a known issue? Any plan to fix that with the 4.9 release?

Thanks,
Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] RFC: Dropping support for Python 2.x for GDB/LLDB

2019-02-26 Thread Christian Gagneraud

On Wed, 27 Feb 2019 at 01:37, Alexandru Croitor  wrote:
>
> I understand that the Python foundation will drop support for it, but it 
> doesn't mean that all of the community will.

Which community are you talking about? Apple community? There's no such thing.
At least not in the sense of "Open source" community.
There is a "capitalistic" Apple community, people who wants to make
money out of anything/nothing.
But none of this "community" is willing to contribute anything.
Python2 is dead and is being buried, for good.
Apple has no control over it. They can decide to maintain their own
python2 version if they wish, after all, that's what they've been
doing  with clang...

Chris.
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] RFC: Dropping support for Python 2.x for GDB/LLDB

2019-02-26 Thread Christian Gagneraud

On Wed, 27 Feb 2019 at 01:10, Alexandru Croitor  wrote:
>
> I know that people like to throw around the end of support date, but in my 
> opinion the reality will be just like with Windows XP.

Python is not Microsoft.
They'll drop it, they mean it. They even gave you a website for that.

Reminds me of Qt private headers:
//
//  W A R N I N G
//  -
//
// This file is not part of the Qt API.  It exists for the convenience
// of FooBarBaz. This header file may change from version to
// version without notice, or even be removed.
//
// We mean it.
//
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] RFC: Dropping support for Python 2.x for GDB/LLDB

2019-02-26 Thread Christian Gagneraud

On Tue, 26 Feb 2019 at 23:26, Alexandru Croitor  wrote:
>
> Hi,
>
> Afaik all macOS versions so far, link against and use the system Python 2.7.x 
> in lldb.
>
> In which case I assume that the dumper script and friends have to stay 2.7.x 
> compatible.

Python 2.7 will not be maintained past 2020.
https://pythonclock.org
You have 0 Years 10 Months 4 Days 22 Hours 57 Minutes and 53 Seconds,Huh?

Please update your computer/OS and stop holding us back.

Chris


>
> > On 21. Feb 2019, at 20:51, André Pönitz  wrote:
> >
> >
> > Hi all.
> >
> > As you may know, Qt Creator heavily uses Python in the Debugger
> > plugin to interact with the "real" debuggers (GDB, LLDB, CDB, PDB).
> >
> > Currently the scripts in the glue code works with both Python
> > 2.7 and 3.x. This is a workable setup, but makes it a bit
> > unpleasant to test and maintain and in some code paths less
> > efficient than a 3.x-only setup.
> >
> > I am therefore pondering the idea of dropping support for Python 2.x,
> > and I'd like to gauge the potential heat such a move will draw.
> >
> > What do people see as output of the following:
> >
> >python --version
> >python3 --version
> >gdb -version
> >gdb -batch -ex 'py import sys' -ex 'py print(sys.version)'
> >lldb -version
> >lldb -batch -o 'sc import sys' -o 'sc print(sys.version)
> >
> >
> > For the record, on my 18.04.1 LTS here this yields something like:
> >
> >Python 2.7.15rc1
> >Python 3.6.7
> >GNU gdb (Ubuntu 8.1-0ubuntu3) 8.1.0.20180409-git
> >3.6.7 (default, Oct 22 2018, 11:32:17)
> >lldb version 6.0.0
> >2.7.15rc1 (default, Nov 12 2018, 14:31:15)
> >
> >
> > Private reply would be just fine, I might have some follow-up
> > questions for certain setups.
> >
> > Andre'
> >
> > ___
> > Qt-creator mailing list
> > Qt-creator@qt-project.org
> > https://lists.qt-project.org/listinfo/qt-creator
>
> ___
> Qt-creator mailing list
> Qt-creator@qt-project.org
> https://lists.qt-project.org/listinfo/qt-creator
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] qtc/gdb wooes

2019-02-11 Thread Christian Gagneraud

On Tue, 12 Feb 2019 at 09:01, André Pönitz  wrote:
> On Mon, Feb 11, 2019 at 03:58:43PM +1300, Christian Gagneraud wrote:
> > On Sat, 9 Feb 2019 at 09:53, André Pönitz  wrote:
> > > On Fri, Feb 08, 2019 at 11:27:56AM +1300, Christian Gagneraud wrote:
> > > > Hi there,
> > > >
> > > > QtC is behaving very strangely when trying to debug a remote linux app.
> > > > My qmake project is a very simple demo, i just have a main.cpp,
> > > > application.cpp and mainwindow.cpp, cross compiled to arm-linux using
> > > > custom toolchains, custom gdb too.
> > > >
> > > > When I insert a breakpoint, gdb doesn't stop on it. and QtC is lost.
> > > > Clicking on the breakpoint in the debugger 'breakpoints' view gives me
> > > > a error popup:
> > > >
> > > > Could not open 
> > > > ""/home/chgans/Projects/bolt-demos/embedded/mainwindow.cpp":48"
> > > > for reading. Either the file does not exist or you do not have the
> > > > permissions to open it.
> > >
> > > Please open a bug report at bugreports.qt.io and attach the debugger
> > > log (Contents of right pane of Windows->Views->Debugger Log) taken
> > > at a time when this breakpoint is not hit.
> >
> > Thanks, see https://bugreports.qt.io/browse/QTCREATORBUG-21960
> >
> > > > I checked the 'Global debugger log' view, but didn't find clue.
> > > > Is there a way to activate some QtC internal logging?
> > >
> > > There is no extra debugger logging except for the debugger log.
> >
> > That's a pity, i find it very hard to troubleshoot QtC when having
> > issues related to gdb.
> > And going through the "Global Debugger Logs" is not straight forward.
> > Do you have a pretty-printer for these logs? There seems to be special
> > markers in the first few columns
>
> The special markers are for me to be able to grep such output more
> comfortably.
>
> Regarding the log content: There's usually enough in there to spot
> typical setup problems. I am open for suggestions what to change,
> but note that already now it can get huge, and parts are time
> sensitive, so there's a limit to what can be done.

Could you briefly explain the meaning of these prefixes, might be
useful if i (or others) want to investigate further, here is what I
found:
t: timestamp
w: warnings?
d: Debug?
=, >~, >&: gdb stdout/stderr, 2 column is ?

Thanks,
Chris

>
> Andre'
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] qtc/gdb wooes

2019-02-10 Thread Christian Gagneraud

On Sat, 9 Feb 2019 at 09:53, André Pönitz  wrote:
>
> On Fri, Feb 08, 2019 at 11:27:56AM +1300, Christian Gagneraud wrote:
> > Hi there,
> >
> > QtC is behaving very strangely when trying to debug a remote linux app.
> > My qmake project is a very simple demo, i just have a main.cpp,
> > application.cpp and mainwindow.cpp, cross compiled to arm-linux using
> > custom toolchains, custom gdb too.
> >
> > When I insert a breakpoint, gdb doesn't stop on it. and QtC is lost.
> > Clicking on the breakpoint in the debugger 'breakpoints' view gives me
> > a error popup:
> >
> > Could not open 
> > ""/home/chgans/Projects/bolt-demos/embedded/mainwindow.cpp":48"
> > for reading. Either the file does not exist or you do not have the
> > permissions to open it.
>
> Please open a bug report at bugreports.qt.io and attach the debugger
> log (Contents of right pane of Windows->Views->Debugger Log) taken
> at a time when this breakpoint is not hit.

Thanks, see https://bugreports.qt.io/browse/QTCREATORBUG-21960

> > I checked the 'Global debugger log' view, but didn't find clue.
> > Is there a way to activate some QtC internal logging?
>
> There is no extra debugger logging except for the debugger log.

That's a pity, i find it very hard to troubleshoot QtC when having
issues related to gdb.
And going through the "Global Debugger Logs" is not straight forward.
Do you have a pretty-printer for these logs? There seems to be special
markers in the first few columns
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

[Qt-creator] qtc/gdb wooes

2019-02-07 Thread Christian Gagneraud

Hi there,

QtC is behaving very strangely when trying to debug a remote linux app.
My qmake project is a very simple demo, i just have a main.cpp,
application.cpp and mainwindow.cpp, cross compiled to arm-linux using
custom toolchains, custom gdb too.

When I insert a breakpoint, gdb doesn't stop on it. and QtC is lost.
Clicking on the breakpoint in the debugger 'breakpoints' view gives me
a error popup:

Could not open ""/home/chgans/Projects/bolt-demos/embedded/mainwindow.cpp":48"
for reading. Either the file does not exist or you do not have the
permissions to open it.

and later on:

The gdb process has not responded to a command within 40 seconds. This
could mean it is stuck in an endless loop or taking longer than
expected to perform the operation.
You can choose between waiting longer or aborting debugging.

The 'mainwindow.cpp' is actually opened in an editor.
A variation of this, is break point on the first line of my 'main()'
function (which reads "Application app;"). For some reason, in the
breakpoint view, this BP is displayed as a tree node with 3 leaves:
all point on the first line of a main() of various Qt's evdev plugins.

That's really weird, and make me think that QtC/GDB might be confused
by my app command line (that need to contain the list of Qt plugin i
want to use). The command line contains things like '-platform
navicofb -plugin
evdevkeyboard:internalkeymap=/usr/share/NOS/resources/simrad.qmap:dir=/dev/input:mask=mouse.*:internalkey=/dev/input/event0:externalkey=/dev/input/event2
--expose-gc -plugin evdevtouch'

Would anyone have tip on how to troubleshoot the problem? I find it
very hard to troubleshoot QtC issues that are debugger related.

I checked the 'Global debugger log' view, but didn't find clue.
Is there a way to activate some QtC internal logging? For now QtC
stdout/err output is littered with 'qtc.ssh' message, but nothing
concerning the debugger.

Thanks,
Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
https://lists.qt-project.org/listinfo/qt-creator

Re: [Qt-creator] Target SDK 28 / QtC 4.7.2

2018-11-05 Thread Christian Gagneraud

On Tue, 6 Nov 2018 at 09:16, ekke  wrote:
>
> I have installed SDKs up to API 28
>
> it's no problem to select Build SDK 28, but from Android Manifest I can
> only select up to API 26 for Target SDK.

AFAIK, 28 requires clang, and only Qt-5.12 supports it, and it hasn't
been released yet.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] What are advantages to have Core plugin over MainWindow?!?

2018-08-15 Thread Christian Gagneraud

On 15 August 2018 at 20:43, Eike Ziller  wrote:
>
>
>> On 14. Aug 2018, at 21:40, Szyk Cech  wrote:
>>
>> Hi
>> I just read some docs "Extending Qt Creator Manual" and study 
>> qt-creator/src/app/main.cpp file. And I have question (maybe not very wise 
>> but important to me):
>>
>> What are advantages to have Core plugin over MainWindow?!?
>>
>> I mean MainWindow class which handle whole app skeleton (located in main app 
>> exe module).
>>
>> One reason for core plugin, from my point of view is that core plugin is 
>> shared object which can easily share its methods to other plugins - I think 
>> it is possible for exe (app exe module), but counter intuitive and 
>> cumbersome, and platform specific.
>>
>> There are some more reasons for create Core plugin?
>
> Handling it as just another plugin has the advantage that it uses the exact 
> same startup, shutdown, building, linking,  semantics as anything else. 
> No need to invent something else :)

Really? like QtCreator's main() doesn't have special handling for the
core plugin...
Last time I checked, the core plugin was loaded explicitly and require
(lot of) hand-crafted initialisation. If Core plugin is "just" another
plugin, why it is not loaded as a result of being the root dependency
of every other plugins?
I understand Eike's point of view, but I agree with Szyk, the 'Core'
plugin could very well be managed as a 'Core' Library.
QtCreator will refuse to start without it. This IMHO contradicts the
very definition of a plugin.

Having say that, QtCreator is awesome, and it's even more awesome to
be able to study its source code (and criticise it! ;)), I believe
that many of us have learned a lot of things by studying it's
architecture, at least I did.
Thanks again.

Chris

PS: I remember reading some (old) slides that started with "QtCreator
is actually just a plugin loader" or something along these lines.

>
> --
> Eike Ziller
> Principal Software Engineer
>
> The Qt Company GmbH
> Rudower Chaussee 13
> D-12489 Berlin
> eike.zil...@qt.io
> http://qt.io
> Geschäftsführer: Mika Pälsi,
> Juha Varelius, Mika Harjuaho
> Sitz der Gesellschaft: Berlin, Registergericht: Amtsgericht Charlottenburg, 
> HRB 144331 B
>
> ___
> Qt-creator mailing list
> Qt-creator@qt-project.org
> http://lists.qt-project.org/mailman/listinfo/qt-creator
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] QtC Runtime issue with 2 Qt version installed

2018-08-05 Thread Christian Gagneraud

On 6 August 2018 at 16:12, Henry Skoglund  wrote:
> On 2018-08-06 05:21, Christian Gagneraud wrote:
> Morning,
>
> is there any difference if launch Qt Creator using the qtcreator.sh script
> (it's in the same bin dir as qtcreator)?

It's not installed by default, I hoped i could do without.
I've never use it when i develop using Qt/QtC installed with the
online installer.


Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

[Qt-creator] QtC Runtime issue with 2 Qt version installed

2018-08-05 Thread Christian Gagneraud

Hi there,

I build Qt and QtC-4.7 myself, the first version of Qt (5.6) is on the
loader path (using ld.so.conf) and is part of my dev environement, the
second version of Qt (5.9) is installed beside QtC and is not on the
loader path (QtC 4.7 requires Qt>=5.9).

I originally had LD_LIBRARY_PATH pointing to Qt-5.6, but then QtC
refused to start because libqxcb.so (Qt 5.9) was picking up it's
dependencies from Qt-5.6.
I use Qt plugin loading debug and ldd, both pointed me to the same direction.
I then realised that using ld.so.conf worked around the problem for
Qt's own plugins, but i still have this problem with QtC's
libexec/clangbackend, the loader resolves it's dependencies to the
"other" Qt install (5.6, on the loader path).

If i'm not wrong, when i build Qt LD_LIBRARY_PATH is set and point to
Qt-5.6, but i use Qt-5.9's qmake to configure QtC.

Is there a way to force all these binaries to only rely on a reliable
rpath, where Qt-5.9 will have proiority?

Here is what I have (Qt-5.6 is installed in /opt/navico-sdk, Qt-5.9
and QtC are installed in /opt/navico-sdk/ide):

$ patchelf --print-rpath /opt/navico-sdk/ide/bin/qtcreator
$ORIGIN:$ORIGIN/..:$ORIGIN/../lib/qtcreator:/opt/navico-sdk/ide/lib
$ patchelf --print-rpath /opt/navico-sdk/ide/libexec/qtcreator/clangbackend
$ORIGIN/../../lib/qtcreator:$ORIGIN/../../lib/qtcreator/plugins:$ORIGIN:$ORIGIN/..:$ORIGIN/../lib/qtcreator:/opt/navico-sdk/lib:/opt/navico-sdk/ide/lib

QtCreator has no loader path issues, but clangbackend has.
That's weird.

I would like to find a proper solution before starting to do `patchelf
--set-rpath`... That would be cleaner IMHO. :)

Any point out?

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

[Qt-creator] 'Proper' Python support

2018-06-10 Thread Christian Gagneraud

Hi there,

With the upcoming "Python for Qt" release, is there any plan to have a
more advanced python editor?

There's a "Python Language Server" available[1], an implementation of
the "Language Server Protocol" [2] for python.

I personally hate companies like Palentir [3], and i'm not a Microsoft
fanboy (to say the least) , but these open source component could be
useful to QtCreator.

Visual Code is the new "Revolution" according to Microsoft fanboys
(trying to mimick Apple fanboys), where does QtCreator stand in this?

I'm working in a 98% MS environment, but every developer has a Linux
VM. Yet, what is the first thing they install in their Linux VM?
Visual Code! What drive me nuts, is that they all have a Qt License,
yet they don't give a sh*t about QtC, way better to require a MS
Visual Studio license as well... But i'm diverging.

The only reason i can think of, is that they are completely
brain-washed and addicted to MS technologies, to a point that they
won't even look at something else, even if it's better.

Now when it comes to python, Visual Code is actually way better than
QtCreator, thanks to their "Language Server Protocol". It's actually a
brilliant idea [4]

I do lot of C++, Qt and Python, often, but not always mixed together.
I can't wait for the new "Python for Qt" and I would love QtCreator to
have way better support for python, in the mean time emacs is still my
editor of choice, but i'm seriously thinking about using Visual Code
for editing python.

I use as well the Squish tool [5], a test framework, driven in python,
to test Qt/C++ applications. But I cannot use QtCreator, I have to use
an Eclipse based IDE... Some of my colleagues don't use the so called
"Squis IDE", instead they prefer a mix of command line and ... Visual
Code. Again!

Any thoughts?

Chris

[1] https://github.com/palantir/python-language-server
[2] https://github.com/Microsoft/language-server-protocol
[3] https://en.wikipedia.org/wiki/Palantir_Technologies#Controversies
[4] https://microsoft.github.io/language-server-protocol/
[5] https://www.froglogic.com/squish/
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] [Qbs] qbs project file parser

2018-05-27 Thread Christian Gagneraud

On 28 May 2018 at 07:09, sanjay chopra  wrote:

> Hi,
>
> I am thinking of writing a plugin for qtcreator which shows assembly
> output of ource at the time of editing(like godbolt.org), I am using qbs
> as build system. Qbs can have different build setting(compiler flags) for
> each source file, is there any library for parsing qbs file and read
> settings (compiler flags) for a source file in project ?
>

If it's for a QtC plugin, you shouldn't be Qbs specific then.Surely QtC
offers you all the information you need.

Chris


>
> *RegardsSanjay*
>
> ___
> Qbs mailing list
> q...@qt-project.org
> http://lists.qt-project.org/mailman/listinfo/qbs
>
>
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] [Qbs] qbs project file parser

2018-05-27 Thread Christian Gagneraud

On 28 May 2018 at 07:09, sanjay chopra  wrote:

> Hi,
>
> I am thinking of writing a plugin for qtcreator which shows assembly
> output of ource at the time of editing(like godbolt.org), I am using qbs
> as build system. Qbs can have different build setting(compiler flags) for
> each source file, is there any library for parsing qbs file and read
> settings (compiler flags) for a source file in project ?
>

You should maybe ask this question on QtCreator mailing list, if it's a QtC
plugin, you shouldn't be Qbs specific.
Surely QtC offers you all the information you need.
Now showing the assembler for an isolated, small piece of code is way
easier than "on the fly" in a C++ project.

Any way interesting.

>
> *RegardsSanjay*
>
> ___
> Qbs mailing list
> q...@qt-project.org
> http://lists.qt-project.org/mailman/listinfo/qbs
>
>
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] "Could not initialise GLX"

2018-01-06 Thread Christian Gagneraud

On 6 January 2018 at 22:59, René J.V. Bertin  wrote:
> Hi,
>
> What does Creator need GLX for and is there a way to disable the feature(s) 
> in question?
> I'm getting a GLX error when I try to use Creator on a remote X display and 
> don't get any further than that failure message.

I got that as well on a VM and/or on a remote docker container, AFAIR
this was not fatal.
Maybe you need to install mesa?

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] [Qbs] Qtc+Qbs+git wroktree = cannot lock build graph

2017-12-13 Thread Christian Gagneraud

On 13/12/2017 8:58 PM, Eike Ziller wrote:

On 12. Dec 2017, at 23:02, Christian Gagneraud
wrote:

On 12/12/2017 9:56 PM, Christian Kandeler wrote:
On Tue, 12 Dec 2017 10:03:58 +1300 Christian Gagneraud
wrote:

Not sure if it's a Qtc or a Qbs issue, but here is my problem:
I'm using git worktrees, which means i have the same project
checked-out on different branches in different sibling

directories, the main qbs files have the same name, and so by
default, Qtc try to use the same build directory, luckily qbs
refuse to go ahead (not sure if with qmake i'll get a warning
about that). Anyway, the solution is to edit the project's
build settings. It's OK, but it's a bit annoying over time.
What about if Qtc would hash the path to the project file and
use it (as a prefix) to name the default build folder?
Shouldn't you be able to set
Tools->Options->Build&Run->General->Default build directory to a
value that makes sense for your setup?

Yes, I would need to add a unique marker based on
CurrentProject::FilePath, is there a javascript function to
calculate a hash of the file path?

So my case is:

$ cd ~/Projects/$project $ ls master ...

Each directory contains a git worktree, so they all have
the same project name and so the same location and directory name
for their build:
~/Projects/$project/build---, i would like to
have
~/Projects/$project/build.

Maybe you could use %{CurrentProject:VcsTopic}

Interesting idea, in my case the topic contains '/', but this could
still work.

Otherwise

%{JS:Util.completeBaseName("%{CurrentProject:Path}”)}

will expand to the last path component of the project file, e.g. if
the project file is "/foo/bar/1.0/foo.pro" it will expand to “1.0”

This could work as well, thanks.

(The JS variable will evaluate the javascript following “JS:”; you
find the methods that are available in “Util” here:
http://code.qt.io/cgit/qt-creator/qt-creator.git/tree/src/plugins/coreplugin/corejsextensions.h
.)

Thanks for the link.
Wouldn't it be nice to have a javascript section int the QtC manual?
Maybe in the "Advanced use" main section.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] [Qbs] Qtc+Qbs+git wroktree = cannot lock build graph

2017-12-12 Thread Christian Gagneraud


On 12/12/2017 9:56 PM, Christian Kandeler wrote:

On Tue, 12 Dec 2017 10:03:58 +1300 Christian Gagneraud
 wrote:


Not sure if it's a Qtc or a Qbs issue, but here is my problem: I'm
using git worktrees, which means i have the same project
checked-out on different branches in different sibling directories,
the main qbs files have the same name, and so by default, Qtc try
to use the same build directory, luckily qbs refuse to go ahead
(not sure if with qmake i'll get a warning about that). Anyway, the
solution is to edit the project's build settings. It's OK, but it's
a bit annoying over time. What about if Qtc would hash the path to
the project file and use it (as a prefix) to name the default build
folder?


Shouldn't you be able to set
Tools->Options->Build&Run->General->Default build directory to a
value that makes sense for your setup?


Yes, I would need to add a unique marker based on 
CurrentProject::FilePath, is there a javascript function to calculate a 
hash of the file path?


So my case is:

$ cd ~/Projects/$project
$ ls
master


...

Each  directory contains a git worktree, so they all have the 
same project name and so the same location and directory name for their 
build: ~/Projects/$project/build---, i would like 
to have 
~/Projects/$project/build.


Thanks,
Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

[Qt-creator] Qtc+Qbs+git wroktree = cannot lock build graph

2017-12-11 Thread Christian Gagneraud


Hi there,

Not sure if it's a Qtc or a Qbs issue, but here is my problem:
I'm using git worktrees, which means i have the same project checked-out 
on different branches in different sibling directories, the main qbs 
files have the same name, and so by default, Qtc try to use the same 
build directory, luckily qbs refuse to go ahead (not sure if with qmake 
i'll get a warning about that).

Anyway, the solution is to edit the project's build settings.
It's OK, but it's a bit annoying over time. What about if Qtc would hash 
the path to the project file and use it (as a prefix) to name the 
default build folder?


Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] Disabling auto-reload of qbs files

2017-11-22 Thread Christian Gagneraud


On 22/11/2017 10:57 PM, Christian Kandeler wrote:

On Wed, 22 Nov 2017 17:27:11 +1300 Christian Gagneraud
 wrote:


When i work with QtC and qbs, and i need to tweak my qbs files,
Qtc auto-reload my project each time I save a qbs file, which in my
case means ca. 1 minute of un-usability.


First of all, that seems too long. Can you please file a task and
assign it to me? I'll provide you with instructions on how to see
where the time is spent. (Chances are it's somewhere else in Creator,
as project resolving happens in its own thread and should thus not
freeze the GUI).


OK, I just realised that i'm using a custom build of Qtc, "Qt Creator 
4.5.0-rc1 (4.4.83)" - built on the 18th of October.
The reason being that Qt doesn't support Linux-x86_32 anymore, so I have 
to build it myself (I cannot switch to Linux-x86_64).
So this might not help at having a stable QtC. I will first install a 
release build (4.4.1 tag) and see how it goes.


Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] Disabling auto-reload of qbs files

2017-11-22 Thread Christian Gagneraud


On 22/11/2017 10:57 PM, Christian Kandeler wrote:

On Wed, 22 Nov 2017 17:27:11 +1300 Christian Gagneraud
 wrote:


When i work with QtC and qbs, and i need to tweak my qbs files,
Qtc auto-reload my project each time I save a qbs file, which in my
case means ca. 1 minute of un-usability.


First of all, that seems too long. Can you please file a task and
assign it to me? I'll provide you with instructions on how to see
where the time is spent. (Chances are it's somewhere else in Creator,
as project resolving happens in its own thread and should thus not
freeze the GUI).


I have to say it's hard to reproduce, it's morning here, and my machine 
restarted overnight, so I've just run QtC and opened my project, and i 
cannot notice any slow down yet.


As the day goes i will try to report the exact condition(s), but i think 
it has to do with:

- Edit qbs (a build might be running)
- kill build and save qbs
- The qbs process kicks in
- The "Scanning for tests" kicks in
- The code model process kicks in
- Hit build  again (maybe while the test scan and code model are still 
running)


Thinking further as i type this message, this could be actually a 
resource problem with my machine: "only" 4 cores, it's a VM (gnome/mir 
thingy) that has poor graphical performance (I had to disable some gnome 
settings).
I'm constantly monitoring the CPU load, and I've notice that QtC is 
resource intensive (that might be due to the code model and test scanning).




I would prefer to edit and save my qbs files as I wish, and then 
explicitly tell QtC to "re-run qbs".


I think the auto-reload is very useful on small project, but is
not desirable on big projects. It would be nice to have an option
somewhere to control this behaviour.

Any comment on this feature request? I'm sure I'm not the only one
to be "hit" by this.


Independent of the build system, there are certain cases where the
"auto-reload" functionality is annoying, e.g. when you type in a new
build directory (there is a delay, but it only helps so much). On the
other hand, suppose you could switch that off via some hypothetical
checkbox. If you then forget to manually reload, Creator will have an
outdated view of the project, which may result in surprising behavior
e.g. regarding the code model, in turn leading to bogus bug reports.
That's not to be underestimated.


I fully agree with you, but i feel comfortable with that.
The reason being that, on this project, right now, i mainly do "qbs 
stuff" within QtC, my goal is to build this project with QtC and Qbs. At 
this stage, I do not really care about up-to-date code model (which 
takes quite some time) and the auto-tests (which takes quite some time 
too, but I've now disabled this plugin).
I would say that the size of my project is very roughly similar to QtC's 
one. My Qbs files are actually not complicated, there's no crazy 
dependency, ...


Basically my work flow is:
- try to build
- analyse build failure
- tweak Qbs
- rince and repeat

Will see later in the day, if I can capture the scenario and i have the 
feeling that it's not due to my VM, I will create a ticket with all the 
necessary details.


Thanks for your help,
Chris

___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

[Qt-creator] Disabling auto-reload of qbs files

2017-11-21 Thread Christian Gagneraud


Hi there,

When i work with QtC and qbs, and i need to tweak my qbs files, Qtc 
auto-reload my project each time I save a qbs file, which in my case 
means ca. 1 minute of un-usability.
I would prefer to edit and save my qbs files as I wish, and then 
explicitly tell QtC to "re-run qbs".


I think the auto-reload is very useful on small project, but is not 
desirable on big projects. It would be nice to have an option somewhere 
to control this behaviour.


Any comment on this feature request? I'm sure I'm not the only one to be 
"hit" by this.


Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] Clang Static Analyzer does not find STL headers

2017-11-05 Thread Christian Gagneraud

On 5 November 2017 at 21:06, Konstantin Tokarev  wrote:
>
>
> 05.11.2017, 02:23, "Christian Gagneraud" :
>> On 5 November 2017 at 03:03, Christian Gagneraud  wrote:
>>>  On 5 November 2017 at 01:40, Benjamin TERRIER  wrote:
>>>>  2017-11-03 18:07 GMT+01:00 Nikolai Kosjar :
>>>>>  There is not much configuration for the clang static analyzer.
>>>>>
>>>>>  Please create a bug report with:
>>>>>   * used versions of Qt Creator and configured clang version for the 
>>>>> analyzer
>>>>>   * whether this already happens with a "normal" toolchain. If not, please
>>>>>  provide some information how to set up your used toolchain.
>>>
>>>  Hi Nikolai, Benjamin
>>>
>>>  What is a "normal toolchain", Boot2Qt, Yocto, Qbs or own QtC sauce?
>>
>> Or maybe you meant a non-bare-metal toolchain? Or a toolchain that
>> does not involved Qt/qmake? Or...
>
> When people say "toolchain" without further clarification, they usually mean 
> compiler
> toolchain which consists of preprocessor, compiler, assembler, and linker

Hi Konstantin,

My question was not what a "toolchain" is, my question was what a
"normal toolchain" is.
I don't think "normal" brings clarification here.

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] Clang Static Analyzer does not find STL headers

2017-11-04 Thread Christian Gagneraud

On 5 November 2017 at 03:03, Christian Gagneraud  wrote:
> On 5 November 2017 at 01:40, Benjamin TERRIER  wrote:
>> 2017-11-03 18:07 GMT+01:00 Nikolai Kosjar :
>>>
>>> There is not much configuration for the clang static analyzer.
>>>
>>> Please create a bug report with:
>>>  * used versions of Qt Creator and configured clang version for the analyzer
>>>  * whether this already happens with a "normal" toolchain. If not, please
>>> provide some information how to set up your used toolchain.
>
> Hi Nikolai, Benjamin
>
> What is a "normal toolchain", Boot2Qt, Yocto, Qbs or own QtC sauce?

Or maybe you meant a non-bare-metal toolchain? Or a toolchain that
does not involved Qt/qmake? Or...

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] Clang Static Analyzer does not find STL headers

2017-11-04 Thread Christian Gagneraud

On 5 November 2017 at 01:40, Benjamin TERRIER  wrote:
> 2017-11-03 18:07 GMT+01:00 Nikolai Kosjar :
>>
>> There is not much configuration for the clang static analyzer.
>>
>> Please create a bug report with:
>>  * used versions of Qt Creator and configured clang version for the analyzer
>>  * whether this already happens with a "normal" toolchain. If not, please
>> provide some information how to set up your used toolchain.

Hi Nikolai, Benjamin

What is a "normal toolchain", Boot2Qt, Yocto, Qbs or own QtC sauce?

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] Clang Static Analyzer does not find STL headers

2017-11-03 Thread Christian Gagneraud

On 4 November 2017 at 06:07, Nikolai Kosjar  wrote:
> On 11/03/2017 11:06 AM, Benjamin TERRIER wrote:
>>
>> Hi,
>>
>> I have an arm-none-eabi gcc toolchain. I have successfully configured Qt
>> Creator
>> to use it in a Kit and was able to use it to build QBS projects.
>>
>> However, when I try to run the static analyzer, clang does not find
>> the standard headers:
>>
>>  main.cpp:9:10: fatal error: 'cstring' file not found
>>
>> A solution is to add the include paths of the standard headers in the
>> QBS project:
>>
>>  cpp.includePaths: [
>>
>> ".../toolchain/arm/arm-gnu-toolchain/arm-none-eabi/include/c++/6.3.1",
>>  ".../toolchain/arm/arm-gnu-toolchain/arm-none-eabi/include",
>>  ]
>>
>> But this is a bit ugly.
>>
>> Is there some configuration I missed (I guess not since the build
>> works), or is it a known bug ?

Are you trying to substitute a gcc cross-compiler with a native clang compiler?

Have you tried to use gcc's -isysroot that point to the tooolchain
installation prefix?

There's as well a few clang-static-analyser and/or scan-build options
such as --analyzer-target --triplet, ...
I don't have the recipe at hand, but AFAIR you need to tell clang that
this is cross compiled.

Chris


>
>
> There is not much configuration for the clang static analyzer.
>
> Please create a bug report with:
>  * used versions of Qt Creator and configured clang version for the analyzer
>  * whether this already happens with a "normal" toolchain. If not, please
> provide some information how to set up your used toolchain.
>
> Thanks,
>Nikolai
>
> ___
> Qt-creator mailing list
> Qt-creator@qt-project.org
> http://lists.qt-project.org/mailman/listinfo/qt-creator
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] A new syntax highlighting color scheme


On 27/10/2017 2:42 PM, Yasser Sobhy wrote:


On Oct 27, 2017 3:38 AM, "Christian Gagneraud" <mailto:chg...@gmail.com>> wrote:


On 27/10/2017 2:05 PM, Yasser Sobhy wrote:

Both of them look beautiful, but none of them is shipped with my
current version of Qt Creator, that's why I made a new one.


Which version of Qt Creator are you using?


4.0.2


According to the comment in QTCREATORBUG-14803, the solarized themes 
made it into QtC 4.1.


Any reason you don't update QtC?

Chris
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] A new syntax highlighting color scheme


On 27/10/2017 2:05 PM, Yasser Sobhy wrote:
Both of them look beautiful, but none of them is shipped with my current 
version of Qt Creator, that's why I made a new one.


Which version of Qt Creator are you using?

PS: Please keep the mailing list posted when you answer an email.
___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] A new syntax highlighting color scheme


On 27/10/2017 1:46 PM, Yasser Sobhy wrote:

Which color scheme is based on Solarized?


it's call "Solarized Dark" and "Solarized Light", they have been added 
some time ago, i did a request to integrate the work of someone else i 
found on github, and it turned out there was already a ticket.


See http://lists.qt-project.org/pipermail/qt-creator/2016-June/005795.html
and
https://bugreports.qt.io/browse/QTCREATORBUG-14803

At the end, i don't want which of the 2 themes have been used, but 
solarized is now  in QtC.


Chris



For git/diff views, yes, I think I need to review them

On Oct 27, 2017 2:41 AM, "Christian Gagneraud" <mailto:chg...@gmail.com>> wrote:


On 27/10/2017 1:12 PM, Yasser Sobhy wrote:

I have just made a new color scheme for Qt Creator and wanted to
share it with you

https://github.com/yasser-sobhy/qt-creator-solarized
<https://github.com/yasser-sobhy/qt-creator-solarized>


QtC ships with a Solarized already. I think a few things could be
improved, colors are not very good IMHO in git/diff views.

Chris


Regards


___
Qt-creator mailing list
Qt-creator@qt-project.org <mailto:Qt-creator@qt-project.org>
http://lists.qt-project.org/mailman/listinfo/qt-creator
<http://lists.qt-project.org/mailman/listinfo/qt-creator>





___
Qt-creator mailing list
Qt-creator@qt-project.org
http://lists.qt-project.org/mailman/listinfo/qt-creator

Re: [Qt-creator] A new syntax highlighting color scheme