[frameworks-kpackage] [Bug 400798] New: install might cause a segfault

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=400798

Bug ID: 400798
   Summary: install might cause a segfault
   Product: frameworks-kpackage
   Version: unspecified
  Platform: Other
OS: Linux
Status: REPORTED
  Severity: normal
  Priority: NOR
 Component: default
  Assignee: notm...@gmail.com
  Reporter: m...@debian.org
  Target Milestone: ---

Created attachment 116150
  --> https://bugs.kde.org/attachment.cgi?id=116150=edit
Package-might-be-gone.patch

SUMMARY

With the introduction of "Emit signals when a package is installed/uninstalled"
(b42973d964df159119467030689f36cc005715b5) the autotest
PlasmoidPackageTest::noCrashOnAsyncInstall started failing since the test is
explicitly making sure the package is destroyed before the install job is
executed.

The new connect code in the install procedure uses the package which, in this
case is a nullptr. This is supposedly how it's in use by plasmashell (according
to the test description).

The attached patch is a workaround that assumes that the package won't become
nullptr while handling the connect, thus it's not perfect.

This issue seems to be easily reproducible running the autotests, it would be
nice if the autotests are actually tested, at least on release.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[print-manager] [Bug 397125] print-manager sends password to remote cups server

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=397125

Maximiliano Curia  changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #2 from Maximiliano Curia  ---
Hi,

I think that the issue here is not that the password needs to be sent, but that
the dialog is not clear about what's the password for and if the password is
going to be sent encrypted or not. I think that the dialog could do with that
additional information, making it clear that you are about to expose the
password to the server and/or the whole network (if sent non encrypted).

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[kde-cli-tools] [Bug 396018] Circular dependency after "Use platform detection from KWorkSpace in KCMShell" (896dd55)

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=396018

--- Comment #2 from Maximiliano Curia  ---
I guess this is fine as an eventual fix. From the Debian side, we downgraded
the plasma-workspace dependency on kde-cli-tools to a recommends to avoid the
circular dependencies at build time till this is fixed. I assume this isn't the
only reason to bump the qt dependency to 5.11.

Is kde frameworks also planning to add a qt 5.11 dependency?

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[kde-cli-tools] [Bug 396018] New: Circular dependency after "Use platform detection from KWorkSpace in KCMShell" (896dd55)

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=396018

Bug ID: 396018
   Summary: Circular dependency after "Use platform detection from
KWorkSpace in KCMShell" (896dd55)
   Product: kde-cli-tools
   Version: unspecified
  Platform: Other
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: general
  Assignee: aleix...@kde.org
  Reporter: m...@gnuservers.com.ar
  Target Milestone: ---

Hi,

Starting with "Use platform detection from KWorkSpace in KCMShell" (896dd55),
kde-cli-tools builds depends on libkworkspace (which is part of
plasma-workspace). On the other hand, plasma-workspace uses kde-cli-tools as a
runtime dependency.

This generates a circular dependency that when using strict version
dependencies makes the package impossible to build. These strict version
dependencies are necessary because components inside plasma tend to change
without indicating the versions that they need.

One possibility would be to copy the detectPlatform function into kde-cli-tools
(it's only 22 lines, no interaction with the rest of the library). This has
already been done in kded.

Another possibility would be to move the detectPlatform function to
libkf5plasma in plasma-frameworks, and update plasma-workspace, kded,
kde-cli-tools, systemsettings, and powerdevil to use that exported function.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwallet-pam] [Bug 393856] patches for CVE-2018-10380 "Access to privileged files" break access to existing wallets

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=393856

Maximiliano Curia <m...@gnuservers.com.ar> changed:

   What|Removed |Added

 CC||m...@gnuservers.com.ar
  Latest Commit||https://commits.kde.org/kwa
   ||llet-pam/8da1a47035fc92bc14
   ||96059583772bc4bd6e8ba6

--- Comment #6 from Maximiliano Curia <m...@gnuservers.com.ar> ---
Git commit 8da1a47035fc92bc1496059583772bc4bd6e8ba6 by Maximiliano Curia.
Committed on 05/05/2018 at 10:00.
Pushed by maximilianocuria into branch 'Plasma/5.12'.

Avoid giving an stderr to kwallet

Summary:
The fixes for CVE-2018-10380 introduced a regression for most users not
using kde, and some for kde sessions. In particular the reorder of the
close calls and creating a new socket caused that the socket is always
assigned the file descriptor 2, aka stderr.

Test Plan: It works

Reviewers: #plasma, aacid

Reviewed By: aacid

Subscribers: asturmlechner, rdieter, davidedmundson, plasma-devel

Tags: #plasma

Differential Revision: https://phabricator.kde.org/D12702

M  +4-1pam_kwallet.c

https://commits.kde.org/kwallet-pam/8da1a47035fc92bc1496059583772bc4bd6e8ba6

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwallet-pam] [Bug 393856] patches for CVE-2018-10380 "Access to privileged files" break access to existing wallets

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=393856

--- Comment #3 from Maximiliano Curia <m...@debian.org> ---
Created attachment 112421
  --> https://bugs.kde.org/attachment.cgi?id=112421=edit
Use any fd, but not stderr

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwallet-pam] [Bug 393856] patches for CVE-2018-10380 "Access to privileged files" break existing wallets

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=393856

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #2 from Maximiliano Curia <m...@debian.org> ---
Also reported in Debian:

https://bugs.debian.org/897687

-- 
You are receiving this mail because:
You are watching all bug changes.

[frameworks-ktexteditor] [Bug 393493] CVE-2018-10361: ktexteditor / Kate local privilege escalation

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=393493

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

-- 
You are receiving this mail because:
You are watching all bug changes.

[frameworks-ktexteditor] [Bug 393493] New: CVE-2018-10361: ktexteditor / Kate local privilege escalation

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=393493

Bug ID: 393493
   Summary: CVE-2018-10361: ktexteditor / Kate local privilege
escalation
   Product: frameworks-ktexteditor
   Version: unspecified
  Platform: unspecified
OS: All
Status: UNCONFIRMED
  Severity: grave
  Priority: NOR
 Component: general
  Assignee: kwrite-bugs-n...@kde.org
  Reporter: m...@debian.org
  Target Milestone: ---

Created attachment 112232
  --> https://bugs.kde.org/attachment.cgi?id=112232=edit
ktexteditor_tmpfile_fix.patch

This bug was originally reported in the debian bugtracker by Salvatore
Bonaccorso :

Hi

See http://www.openwall.com/lists/oss-security/2018/04/24/1 for
details (and proposed patch).

Regards,
Salvatore

-- 
You are receiving this mail because:
You are watching all bug changes.

[frameworks-kdelibs4support] [Bug 393255] Port from libnm-glib/libnm-util to libnm

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=393255

--- Comment #12 from Maximiliano Curia <m...@debian.org> ---
(In reply to mbiebl from comment #11)
> (In reply to Maximiliano Curia from comment #10)
> > Anyway, you might want to forward this patch or the one that drops the use
> > of network manager from kdelibs4support to the codereview system:
> > https://phabricator.kde.org

> This means creating yet another user account (it seems the user account I
> created for bugs.kde.org is not accepted for phabricator.kde.org).

> Maxy, if you think this should be forwarded to phabricator.kde.org, would
> you mind doing that?

If I forward the patch, then I'll be playing ball for every change requested
upstream. And I think that the change should drop the code that was not being
used. Anyway, it's way better if you can be the one responding the review.

The documentation on how to deal with phabricator is here:
 - https://community.kde.org/Infrastructure/Phabricator

-- 
You are receiving this mail because:
You are watching all bug changes.

[frameworks-kdelibs4support] [Bug 393255] Port from libnm-glib/libnm-util to libnm

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=393255

--- Comment #10 from Maximiliano Curia <m...@debian.org> ---
(In reply to mbiebl from comment #9)
> (In reply to Maximiliano Curia from comment #8)
> > Well kdelibs4support is a porting aid, a library that provides some
> > interfaces that were available in kde4 but are not part of kde frameworks 5
> > (they have a suitable replacement in the kf5 family but it might not always
> > be trivial to migrate to), as such it's highly disrecommended to use it. So,
> > I wouldn't bet on upstream fixing, or even accepting a patch to fix,
> > something that was always broken.

> To clarify: The build was broken in kdelibs4support, it is not broken in
> kde-runtime.

Yeah, the NM_0_7 define is present in kde-runtime.

> > It might make more sense to drop the dead code (as long as it doesn't break
> > any api).

> I don't know KDE too well, so I don't know which KF5 programs still rely on
> the functionality of the networkstatus kded module or if it's safe to drop
> it. If you think dropping the functionality is the way to go, do you think
> the same should be applied to kde-runtime?

kde-runtime is no longer being maintained by kde.org, the repository is still
available, but that's it, so the patch won't be applied upstream. On the Debian
side, we don't really know if somebody uses the networkstatus kded module, a
quick search in codesearch showed no results, but the idea is to drop
kde-runtime (and qt4 for that matter) for buster, so disabling an optional
feature should be fine.

Anyway, you might want to forward this patch or the one that drops the use of
network manager from kdelibs4support to the codereview system:
https://phabricator.kde.org

Upstream usually has a lower latency when a code review is requested.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[frameworks-kdelibs4support] [Bug 393255] Port from libnm-glib/libnm-util to libnm

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=393255

--- Comment #8 from Maximiliano Curia <m...@debian.org> ---
(In reply to mbiebl from comment #5)
> (In reply to Maximiliano Curia from comment #4)
> > (In reply to mbiebl from comment #3)
> > > Perhaps just dropping the Build-Depends, disabling the futile attempt
> > > to use NetworkManager would be a better option than bothering with a
> > > patch?

> > That's my current plan. As soon as kdesignplugin leaves NEW, I'll upload
> > 5.45, which drops the networkmanager build dependency.

> I guess it would still make sense to fix this code upstream. If that is not
> going to happen, maybe it 
> (In reply to Maximiliano Curia from comment #4)
> > (In reply to mbiebl from comment #3)
> > > Perhaps just dropping the Build-Depends, disabling the futile attempt
> > > to use NetworkManager would be a better option than bothering with a
> > > patch?

> > That's my current plan. As soon as kdesignplugin leaves NEW, I'll upload
> > 5.45, which drops the networkmanager build dependency.
> 
> As lubo correctly noted, due to the bug in build system, the
> networkmanagerstatus.cpp code was never compiled, as NM_0_7 was never set,
> even if NetworkManager.pc was found.

> So the question is, whether to rip out the NetworkManager code completely
> (upstream) or fix it. 

Well kdelibs4support is a porting aid, a library that provides some interfaces
that were available in kde4 but are not part of kde frameworks 5 (they have a
suitable replacement in the kf5 family but it might not always be trivial to
migrate to), as such it's highly disrecommended to use it. So, I wouldn't bet
on upstream fixing, or even accepting a patch to fix, something that was always
broken.

It might make more sense to drop the dead code (as long as it doesn't break any
api).


Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[frameworks-kdelibs4support] [Bug 393255] Port from libnm-glib/libnm-util to libnm

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=393255

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #4 from Maximiliano Curia <m...@debian.org> ---
(In reply to mbiebl from comment #3)
> Perhaps just dropping the Build-Depends, disabling the futile attempt
> to use NetworkManager would be a better option than bothering with a
> patch?

That's my current plan. As soon as kdesignplugin leaves NEW, I'll upload 5.45,
which drops the networkmanager build dependency.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[Breeze] [Bug 373088] Current tab highlighting counter-intuitive when using dark themes

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=373088

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #15 from Maximiliano Curia <m...@debian.org> ---
(In reply to Hugo Pereira Da Costa from comment #13)
> (In reply to Christoph Feck from comment #12)
> > According to the bug's history, it was you who removed you as the assignee 
> > :)
> > 
> > https://bugs.kde.org/show_activity.cgi?id=373088
> 
> Well, I guess that must be the other me, taking action when I turn my back.
> More seriously, about to push.
> - to master
> - to Plasma/5.9 (since this is a bug right ?)
> Should it also be pushed to Plasma/5.8 ? (LTS) ?

Yes, please, the original report was reported against 5.8.4, and the
https://bugs.debian.org/866996 follow ups indicate that users of 5.8 are still
affected by this.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[plasmashell] [Bug 379312] gtk systray icons aren't receiving the forwarded mouse events

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=379312

--- Comment #4 from Maximiliano Curia <m...@debian.org> ---
Created attachment 106135
  --> https://bugs.kde.org/attachment.cgi?id=106135=edit
Patch for 5.8.7

-- 
You are receiving this mail because:
You are watching all bug changes.

[plasmashell] [Bug 379312] gtk systray icons aren't receiving the forwarded mouse events

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=379312

--- Comment #3 from Maximiliano Curia <m...@debian.org> ---
(In reply to Wolfgang Bauer from comment #2)
> Should be fixed meanwhile:
> https://cgit.kde.org/plasma-workspace.git/commit/
> ?id=7df184afa19f148c1cd09ae9588645bb2b4556fc

I'm currently taking xembed-sni-proxy/* from the mentioned patch on top of
5.8.7, and the status icons now work as expected.
(I guess that's 9df815e843a4385465fff0cb9a76ddc94ed35b38,
f0e45cd42ba7e67ec444a4c14bac855b0ac195df and
7df184afa19f148c1cd09ae9588645bb2b4556fc)

I'm attaching the patch as used in the Debian packages.

-- 
You are receiving this mail because:
You are watching all bug changes.

[plasmashell] [Bug 379312] New: gtk systray icons aren't receiving the forwarded mouse events

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=379312

Bug ID: 379312
   Summary: gtk systray icons aren't receiving the forwarded mouse
events
   Product: plasmashell
   Version: 5.8.6
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: XembedSNIProxy
  Assignee: plasma-b...@kde.org
  Reporter: m...@debian.org
  Target Milestone: 1.0

Created attachment 105241
  --> https://bugs.kde.org/attachment.cgi?id=105241=edit
Minimal

Hi,

A GtkStatusIcon under plasma doesn't receive the "activate" "popup-menu" or
"button-press-event" signals, but the xembedsniproxy shows the event in it's
debug output:
 kde.xembedsniproxy: Received click 1 with passed x*y 1537 1052
 kde.xembedsniproxy: Received click 3 with passed x*y 1537 1052
 kde.xembedsniproxy: Received click 2 with passed x*y 1537 1052

Also, the tooltip text isn't correctly updated. I'm attaching a minimal program
that starts a GtkStatusIcon which works correctly in other desktops that still
support the systray (like cinnamon).

I've also tried applying 9df815e843a4385465fff0cb9a76ddc94ed35b38, and other
than a missing unused header (xcb/xinput.h) that I had to comment it made no
difference.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[Breeze] [Bug 355540] Tooltips color wrong in gtk applications

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=355540

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #35 from Maximiliano Curia <m...@debian.org> ---
(In reply to bartek from comment #1)
> I can confirm that on update Arch Linux, however there is something more. I
> fo you do:
> find ~ -name "*gtkrc*"

> remove all found content and reconfigure gtk theme in kde systemsettings
> (choosing breeze of course), then things are getting back to normal for a
> session time. Both in inkscape and gimp tooltips consist of white letters on
> black background. After relogin tooltips are not readable again, as shown is
> previous post.

This is consistent with the behavior I'm seeing (Debian plasma 5.8.6). It seems
that when you are restoring a session with inkscape running, inkscape starts
before the "right" color scheme for the tooltips is loaded, and this ends up
with white on white tooltips on inkscape. This is probably an issue in
plasma-workspace.

-- 
You are receiving this mail because:
You are watching all bug changes.

[frameworks-sonnet] [Bug 376974] sonnet should not just add regions to languages

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=376974

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 Status|UNCONFIRMED |CONFIRMED
 Ever confirmed|0   |1

--- Comment #1 from Maximiliano Curia <m...@debian.org> ---
This is also reported in the Debian bug tracker as
https://bugs.debian.org/801728
I think that a two or even three level selection of the dictionary (ala
keyboard layout), would be easier to deal with.

-- 
You are receiving this mail because:
You are watching all bug changes.

[frameworks-sonnet] [Bug 376974] sonnet should not just add regions to languages

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=376974

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

-- 
You are receiving this mail because:
You are watching all bug changes.

[kio] [Bug 103331] Group should be preserved when making backup

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=103331

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #16 from Maximiliano Curia <m...@debian.org> ---
(In reply to Martin Koller from comment #15)
> Git commit d252f7958633f3374e77e1cc52f11c534ca55219 by Martin Koller.
> Committed on 22/01/2017 at 19:49.
> Pushed by mkoller into branch 'master'.
> 
> preserve group/owner on file copy
> REVIEW: 129864
> 
> M  +28   -12   src/ioslaves/file/file_unix.cpp
> 
> https://commits.kde.org/kio/d252f7958633f3374e77e1cc52f11c534ca55219

This patch breaks the setgid directories and default acl semantics. It might be
better to avoid setting the group if the dest dir is setgid, and avoid setting
the file permissions if the destdir has a default acl value.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[kio] [Bug 329983] Possibly wrong locale selected for testtrash

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=329983

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #2 from Maximiliano Curia <m...@debian.org> ---
(In reply to David Faure from comment #1)
> Git commit 459411055bb1236caf5d25d83fd4a62a58279fce by David Faure.
> Committed on 14/01/2017 at 12:59.
> Pushed by dfaure into branch 'master'.
> 
> Use C rather than en_US so that the locale doesn't have to be installed
> 
> M  +1-1src/ioslaves/trash/tests/testtrash.cpp
> 
> https://commits.kde.org/kio/459411055bb1236caf5d25d83fd4a62a58279fce

Please be aware that C.UTF-8 is a Debian (and derivatives) specific locale name
(generated at build time by the debian/rules.d/build.mk file), and afaik, it
hasn't been added to the vanilla glib's locale. So you might need to use plain
"C" if C.UTF-8 is not available.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[i18n] [Bug 378117] New: Invalid use of html markup in polkit action file

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=378117

Bug ID: 378117
   Summary: Invalid use of html markup in polkit action file
   Product: i18n
   Version: unspecified
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: hr
  Assignee: zarko.pin...@gmail.com
  Reporter: m...@debian.org
  Target Milestone: ---

This is a forward of the bug reported to the Debian bug tracker: 
 https://bugs.debian.org/696905

The file processcore/actions.actions includes an hr translation for the string 
 "Change the niceness of a given process" that says
 "Promijeni procesorski udio (nice) za dani proces"
and this text ends up in a policykit action file that doesn't allow the use of 
html markups. Please drop the "" and "".

The original report reads:
---
From: Michael Biebl 
Subject: HTML markup not allowed in PolicyKit action file 

Hi,

when running polkitd in debug mode I get
** (polkitd:2714): WARNING **: skipping unknown tag  at line 335

That warning messages comes from
Promijeni procesorski udio (nice) za dani
proces

It looks like HTML markup is not allowed in policy files.

Michael
---

Thanks,

-- 
You are receiving this mail because:
You are watching all bug changes.

[Breeze] [Bug 377572] New: Please install all the gtk-3.* themes

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=377572

Bug ID: 377572
   Summary: Please install all the gtk-3.* themes
   Product: Breeze
   Version: 5.9.3
  Platform: Other
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: gtk theme
  Assignee: scionicspec...@gmail.com
  Reporter: m...@debian.org
  Target Milestone: ---

Hi

Currently breeze-gtk only installs the gtk theme associated with the gtk
version found at build time. But gtk3 is smart enough to look for the right
theme folder when they are all installed (gtk-3.0 would be used if nothing
better, gtk-3.18 would be used if we are running gtk-3.18 or better, and so
on).

This would improve the upgrade path for our users, as it avoids having to
rebuild breeze-gtk for the "right" gtk.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[Breeze] [Bug 377568] New: Please install all the gtk-3.* themes

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=377568

Bug ID: 377568
   Summary: Please install all the gtk-3.* themes
   Product: Breeze
   Version: unspecified
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: gtk theme
  Assignee: scionicspec...@gmail.com
  Reporter: m...@debian.org
  Target Milestone: ---

Hi

Currently breeze-gtk only installs the gtk theme associated with the gtk
version found at build time. But gtk3 is smart enough to look for the right
theme folder when they are all installed (gtk-3.0 would be used if nothing
better, gtk-3.18 would be used if we are running gtk-3.18 or better, and so
on).

This would improve the upgrade path for our users, as it avoids having to
rebuild breeze-gtk for the "right" gtk.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[Breeze] [Bug 377569] New: Please install all the gtk-3.* themes

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=377569

Bug ID: 377569
   Summary: Please install all the gtk-3.* themes
   Product: Breeze
   Version: unspecified
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: gtk theme
  Assignee: scionicspec...@gmail.com
  Reporter: m...@debian.org
  Target Milestone: ---

Hi

Currently breeze-gtk only installs the gtk theme associated with the gtk
version found at build time. But gtk3 is smart enough to look for the right
theme folder when they are all installed (gtk-3.0 would be used if nothing
better, gtk-3.18 would be used if we are running gtk-3.18 or better, and so
on).

This would improve the upgrade path for our users, as it avoids having to
rebuild breeze-gtk for the "right" gtk.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[frameworks-kxmlgui] [Bug 374083] New: setCaption and setPlainCaption always append the application name

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=374083

Bug ID: 374083
   Summary: setCaption and setPlainCaption always append the
application name
   Product: frameworks-kxmlgui
   Version: 5.0.0
  Platform: Other
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: general
  Assignee: kdelibs-b...@kde.org
  Reporter: m...@debian.org
  Target Milestone: ---

This is a forward of the bug reported in the Debian bug tracker
https://bugs.debian.org/848180 :

konsole calls setCaption or setPlainCaption depending on the value of the user
setting "Show application name on the titlebar", this used to do the right
thing in qt4, but in qt5 both calls do the same and qt's qpa plugin is
responsible for appending the application name in the title.

Please consider either avoiding the append of the application name in the title
when calling setPlainCaption, or at least update the documentation to make it
clear that setCaption and setPlainCaption will do the same.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[kalgebra] [Bug 373384] New: Only a single 3D graph plotted

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=373384

Bug ID: 373384
   Summary: Only a single 3D graph plotted
   Product: kalgebra
   Version: unspecified
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: general
  Assignee: aleix...@kde.org
  Reporter: m...@debian.org
  Target Milestone: ---

This is a forwarded report from the Debian issue:
https://bugs.debian.org/845612


Package: kalgebra
Version: 4:16.08.3-1
Severity: normal

Start kalgebra, choose "3D Graph". Either use one of the examples, or enter
a formula like "x*y", press . Graph gets plotted.

Now change the formula to eg. "x+y", press . Graph doesn't show any
surface any more, only the axes.

Even reverting to the previous formula doesn't help - no new data will be
plotted


Additionally, using with the example: t->vector{t,t**2,t}, somehow kalgebra
does manage to redraw this particular plot, somehow I think that this is
managed by analitza in a different way.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[Akonadi] [Bug 373363] New: Kmail folder properties aren't shown after closing them

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=373363

Bug ID: 373363
   Summary: Kmail folder properties aren't shown after closing
them
   Product: Akonadi
   Version: 5.2.3
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: libakonadi
  Assignee: kdepim-b...@kde.org
  Reporter: m...@debian.org
  Target Milestone: ---

Created attachment 102652
  --> https://bugs.kde.org/attachment.cgi?id=102652=edit
Delete the dialog on close

Hi,

When closing the folder properties window with the X button it's no longer
possible to show it again. This is due to the fact that the window is not
deleted on close (as done in with the cancel button pressed signal). I'm
attaching a patch that I'm planning to apply in the Debian packages.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[systemsettings] [Bug 373115] New: Incorrect locale values in ~/.config/plasma* and environment variables

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=373115

Bug ID: 373115
   Summary: Incorrect locale values in ~/.config/plasma* and
environment variables
   Product: systemsettings
   Version: 5.8.4
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: kcm_formats
  Assignee: se...@kde.org
  Reporter: m...@debian.org
  Target Milestone: ---

This is a forward of the bug reported in the Debian bug tracker:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=845788

Currently the values used to fill the combos in the formats kcm come from a:
 QLocale::matchingLocales(QLocale::AnyLanguage, QLocale::AnyScript,
QLocale::AnyCountry);

Sadly this includes values that are not supported by the system locales
libraries, as qt uses it's own locales database. Even more, this disregards the
fact that most users have only a subset of the locales generated (see the
output of locales -a to obtain a list of the currently available locales).

The Debian report has been reassigned to the qt maintainers, and it might be
that the issue can be addressed on the qt side. As per suggestion of Kai Uwe
Broulik via irc I'm forwarding the issue here to raise the issue awareness.

The original report from "Jeroen N. Witmond" reads:

I live in the Netherlands but I prefer to communicate with my computer systems
in English.  I have answered the related questions 
during the installation of Stretch Alpha 8 accordingly.

Now a number of applications complain about the locale. An example is:
jeroen@zandbak:~$ man ssh-keygen
man: can't set the locale; make sure $LC_* and $LANG are correct

The $LC_* variables appear to not be set.

jeroen@zandbak:~$ echo $LANG
en_NL.UTF-8
jeroen@zandbak:~$ echo $LANGUAGE
en_US:en

Some digging showed that en_NL.UTF-8 is an incorrect value, and that it is 
set in files:
jeroen@zandbak:~/.config$ grep en_NL *
plasma-locale-settings.sh:export LANG=en_NL.UTF-8
plasma-localerc:LANG=en_NL.UTF-8

Questions:
1. Can I overwrite the values in these files manually, or do I need to use some
dialog?
2. How to keep other users from stubling on this problem?

Regards, jeroen.


Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[konsole] [Bug 372500] Crash when creating and closing splitted views

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=372500

--- Comment #1 from Maximiliano Curia <m...@debian.org> ---
The backtrace shown in this bug shown in the Debian bug and the one that I
produced (and the ones I've seen before submitting the report) are quite
different from one another. My guess is that on closing the splitted view some
pointer is not being updated correctly, and then the invalid reference might
cause the segfault in different parts of the code.

-- 
You are receiving this mail because:
You are watching all bug changes.

[konsole] [Bug 372500] New: Crash when creating and closing splitted views

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=372500

Bug ID: 372500
   Summary: Crash when creating and closing splitted views
   Product: konsole
   Version: 16.08.2
  Platform: Debian testing
OS: Linux
Status: UNCONFIRMED
  Keywords: drkonqi
  Severity: crash
  Priority: NOR
 Component: general
  Assignee: konsole-de...@kde.org
  Reporter: m...@debian.org
  Target Milestone: ---

Application: konsole (16.08.2)

Qt Version: 5.7.1
Frameworks Version: 5.27.0
Operating System: Linux 4.8.0-1-amd64 x86_64
Distribution: Debian GNU/Linux testing (stretch)

-- Information about the crash:
- What I was doing when the application crashed:

I was testing the bug reported to the Debian bug tracker:
https://bugs.debian.org/844380 and I could reproduce it by creating some
vertical splits, closing some of them, and creating some more.

The crash can be reproduced sometimes.

-- Backtrace:
Application: Konsole (konsole), signal: Segmentation fault
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[Current thread is 1 (Thread 0x7f0d2e9101c0 (LWP 23439))]

Thread 3 (Thread 0x7f0d2a726700 (LWP 23441)):
#0  __libc_enable_asynccancel () at
../sysdeps/unix/sysv/linux/x86_64/cancellation.S:67
#1  0x7f0d3fcc0562 in poll () at ../sysdeps/unix/syscall-template.S:84
#2  0x7f0d3722f9d6 in g_main_context_poll (priority=,
n_fds=1, fds=0x7f0d1c003020, timeout=, context=0x7f0d1c000990)
at ././glib/gmain.c:4228
#3  g_main_context_iterate (context=context@entry=0x7f0d1c000990,
block=block@entry=1, dispatch=dispatch@entry=1, self=) at
././glib/gmain.c:3924
#4  0x7f0d3722faec in g_main_context_iteration (context=0x7f0d1c000990,
may_block=may_block@entry=1) at ././glib/gmain.c:3990
#5  0x7f0d3c8bd73b in QEventDispatcherGlib::processEvents
(this=0x7f0d1c0008c0, flags=...) at kernel/qeventdispatcher_glib.cpp:425
#6  0x7f0d3c8670ca in QEventLoop::exec (this=this@entry=0x7f0d2a725c90,
flags=..., flags@entry=...) at kernel/qeventloop.cpp:212
#7  0x7f0d3c6890e3 in QThread::exec (this=this@entry=0x7f0d4032ed60
<(anonymous namespace)::Q_QGS__q_manager::innerFunction()::holder>) at
thread/qthread.cpp:507
#8  0x7f0d402b96d5 in QDBusConnectionManager::run (this=0x7f0d4032ed60
<(anonymous namespace)::Q_QGS__q_manager::innerFunction()::holder>) at
qdbusconnection.cpp:178
#9  0x7f0d3c68dd98 in QThreadPrivate::start (arg=0x7f0d4032ed60 <(anonymous
namespace)::Q_QGS__q_manager::innerFunction()::holder>) at
thread/qthread_unix.cpp:368
#10 0x7f0d38ef0464 in start_thread (arg=0x7f0d2a726700) at
pthread_create.c:333
#11 0x7f0d3fcc99df in clone () at
../sysdeps/unix/sysv/linux/x86_64/clone.S:105

Thread 2 (Thread 0x7f0d2c69b700 (LWP 23440)):
#0  0x7f0d3fcc056d in poll () at ../sysdeps/unix/syscall-template.S:84
#1  0x7f0d39315150 in poll (__timeout=-1, __nfds=1, __fds=0x7f0d2c69abc0)
at /usr/include/x86_64-linux-gnu/bits/poll2.h:46
#2  _xcb_conn_wait (c=c@entry=0x564f53b24dc0, cond=cond@entry=0x564f53b24e00,
vector=vector@entry=0x0, count=count@entry=0x0) at ../../src/xcb_conn.c:479
#3  0x7f0d39316ee9 in xcb_wait_for_event (c=0x564f53b24dc0) at
../../src/xcb_in.c:693
#4  0x7f0d2e3c0b69 in QXcbEventReader::run (this=0x564f53b2f1d0) at
qxcbconnection.cpp:1343
#5  0x7f0d3c68dd98 in QThreadPrivate::start (arg=0x564f53b2f1d0) at
thread/qthread_unix.cpp:368
#6  0x7f0d38ef0464 in start_thread (arg=0x7f0d2c69b700) at
pthread_create.c:333
#7  0x7f0d3fcc99df in clone () at
../sysdeps/unix/sysv/linux/x86_64/clone.S:105

Thread 1 (Thread 0x7f0d2e9101c0 (LWP 23439)):
[KCrash Handler]
#6  std::__atomic_base::operator++ (this=) at
/usr/include/c++/6/bits/atomic_base.h:296
#7  QAtomicOps::ref (_q_value=...) at
../../include/QtCore/../../src/corelib/arch/qatomic_cxx11.h:265
#8  QBasicAtomicInteger::ref (this=) at
../../include/QtCore/../../src/corelib/thread/qbasicatomic.h:110
#9  QtPrivate::RefCount::ref (this=) at
../../include/QtCore/../../src/corelib/tools/qrefcount.h:61
#10 QString::operator= (this=this@entry=0x564f541e4970, other=...) at
tools/qstring.cpp:1814
#11 0x7f0d402795e9 in QDomNodePrivate::QDomNodePrivate
(this=this@entry=0x564f541e4920, n=n@entry=0x564f540ae790,
deep=deep@entry=true) at dom/qdom.cpp:1462
#12 0x7f0d40279785 in QDomCharacterDataPrivate::QDomCharacterDataPrivate
(deep=true, n=0x564f540ae790, this=0x564f541e4920) at dom/qdom.cpp:3879
#13 QDomTextPrivate::QDomTextPrivate (deep=true, n=0x564f540ae790,
this=0x564f541e4920) at dom/qdom.cpp:5167
#14 QDomTextPrivate::cloneNode (this=0x564f540ae790, deep=) at
dom/qdom.cpp:5173
#15 0x7f0d4027962e in QDomNodePrivate::QDomNodePrivate
(this=this@entry=0x564f541e48b0, n=n@entry=0x564f540ae720,
deep=deep@entry=true) at dom/qdom.cpp:1471
#16 0x7f0d40279735 in QDomAttrPrivate::QDomAttrPrivate (deep=true,
n=0x564f540ae720, this=0x564f541e48b0) at dom/qdom.cpp:4106
#17 QDomAttrPrivate::cloneNode 

[krusader] [Bug 372259] New: The tarball is missing the manpage generation rules

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=372259

Bug ID: 372259
   Summary: The tarball is missing the manpage generation rules
   Product: krusader
   Version: 2.5.0
  Platform: Other
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: general
  Assignee: m...@fork.pl
  Reporter: m...@debian.org
  Target Milestone: ---

Created attachment 102143
  --> https://bugs.kde.org/attachment.cgi?id=102143=edit
Build and install the manpages

It seems that in the release process the CMakeLists.txt in the translation dirs
was removed. In irc I was told that if you use create_tarballs_kf5 you might
avoid this kind of issues.

Anyway, I'm attaching a patch that I'm shipping with the Debian package.

-- 
You are receiving this mail because:
You are watching all bug changes.

[konsole] [Bug 371939] New: Dragging tabs lag

[Maximiliano Curia]

https://bugs.kde.org/show_bug.cgi?id=371939

Bug ID: 371939
   Summary: Dragging tabs lag
   Product: konsole
   Version: 16.08.2
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: tabbar
  Assignee: konsole-de...@kde.org
  Reporter: m...@debian.org
  Target Milestone: ---

This is a forward of the bug reported in the Debian bugtracker:
https://bugs.debian.org/842730

The original submitter reports:

When one tries to drag a tab there is a lot of lag at the start of the
drag gesture. As a result the tab is likely to be trailing the mouse
pointer by inches so that figuring out where the tab will end up is
quite confusing.

For a bug such as this a video is worth a million words so I recorded a
drag session and uploaded it there:

http://fgouget.free.fr/tmp/konsole-drag-16.08.2.ogv

Note that this is not a new bug: it has been there for at least a year.

The workaround is to mark a pause between pressing the left mouse button
and dragging the mouse around.

---
In my experience, the tab doesn't start floating immediately but the ^ markers
are shown where the tab will be moved.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[kmix] [Bug 339913] KMix tries to reconnect to pulseaudio daemon every 50ms

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=339913

--- Comment #3 from Maximiliano Curia <m...@debian.org> ---
Created attachment 101774
  --> https://bugs.kde.org/attachment.cgi?id=101774=edit
Patch as posted in the Debian bugtracker

-- 
You are receiving this mail because:
You are watching all bug changes.

[kmix] [Bug 339913] KMix tries to reconnect to pulseaudio daemon every 50ms

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=339913

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #2 from Maximiliano Curia <m...@debian.org> ---
We have received a patch for this issue in the Debian bugtracker in the bug:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=789887

which still seems to apply.

Index: kmix-4.14.0/backends/mixer_pulse.cpp
===
--- kmix-4.14.0.orig/backends/mixer_pulse.cpp
+++ kmix-4.14.0/backends/mixer_pulse.cpp
@@ -37,6 +37,8 @@
 #  include 
 #endif

+#include 
+
 // PA_VOLUME_UI_MAX landed in pulseaudio-0.9.23, so this can be removed
when/if
 // minimum requirement is ever bumped up (from 0.9.12 currently)
 #ifndef PA_VOLUME_UI_MAX
@@ -716,7 +718,14 @@ static void context_state_callback(pa_co
 if (s_mixers.contains(KMIXPA_PLAYBACK)) {
 kWarning(67100) << "Connection to PulseAudio daemon closed.
Attempting reconnection.";
 s_pulseActive = UNKNOWN;
-QTimer::singleShot(50, s_mixers[KMIXPA_PLAYBACK],
SLOT(reinit()));
+static time_t last_timestamp = 0;
+time_t current_time = time(NULL);
+if ((current_time - last_timestamp) > 3) {
+  QTimer::singleShot(50, s_mixers[KMIXPA_PLAYBACK], SLOT(reinit()));
+} else {
+  QTimer::singleShot(6, s_mixers[KMIXPA_PLAYBACK],
SLOT(reinit()));
+}
+last_timestamp = current_time;
 }
 }
 }

-- 
You are receiving this mail because:
You are watching all bug changes.

[kio] [Bug 371657] New: Facebook web shortcut no longer works

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=371657

Bug ID: 371657
   Summary: Facebook web shortcut no longer works
   Product: kio
   Version: unspecified
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: minor
  Priority: NOR
 Component: general
  Assignee: fa...@kde.org
  Reporter: m...@debian.org

Forwarding Debian bug: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=820925
"""

The file 
/usr/share/kservices5/searchproviders/facebook.desktop

provides a web shortcut to search facebook.

However the query used seems to be out of date and no longer works.

It should be replaced with this line

https://www.facebook.com/search/top/?init=quick=\\{@}

"""
DISCLAIMER: I'm not able to test this, as I don't use facebook, and the search
link requires to be logged in.

Reproducible: Always

-- 
You are receiving this mail because:
You are watching all bug changes.

[kalgebra] [Bug 352526] 3D graph ->views not working

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=352526

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #9 from Maximiliano Curia <m...@debian.org> ---
This issue is still reproducible. It seems that the analitza 3d backend is
missing an update after setting the plotstyle.
Maybe this report needs to be reassigned to analitza.

-- 
You are receiving this mail because:
You are watching all bug changes.

[konsole] [Bug 371414] When open a new konsole, it uses the last profile instead of the default

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=371414

--- Comment #2 from Maximiliano Curia <m...@debian.org> ---
(In reply to stecsa from comment #0)
> Right mouse -> Switch Profile doesn't work.

In a follow up of the Debian bug it's mentioned that the window doesn't follow
the dimensions change. I think this is a completely different issue.

-- 
You are receiving this mail because:
You are watching all bug changes.

[konsole] [Bug 371414] When open a new konsole, it uses the last profile instead of the default

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=371414

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #1 from Maximiliano Curia <m...@debian.org> ---
This was originally reported in the Debian bug tracker:
https://bugs.debian.org/841502

>From what I could reproduce, launching a second konsole from krunner or the kde
menu calls the new window dbus action. Which seems to be using the original
command arguments to select the profile, (instead of profile configured as the
default).

-- 
You are receiving this mail because:
You are watching all bug changes.

[kdeplasma-addons] [Bug 370625] New: comic applet segfaults while trying to show http://phdcomics.com/comics/archive.php?comicid=1889

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=370625

Bug ID: 370625
   Summary: comic applet segfaults while trying to show
http://phdcomics.com/comics/archive.php?comicid=1889
   Product: kdeplasma-addons
   Version: 5.8.0
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: Comic Applet
  Assignee: notm...@gmail.com
  Reporter: m...@debian.org

This was originally reported to the Debian bugtracker in
https://bugs.debian.org/839710

When the comic applet segfaults it causes plasmashell to crash, leaving the
user with a blank desktop.

Reproducible: Always

Steps to Reproduce:
1.  Add the comic plasmoid
2. Install the phdcomics in the comic plasmoid
3. Go back in the comic history till you hit
http://phdcomics.com/comics/archive.php?comicid=1889

Actual Results:  
This causes a segfault.

Expected Results:  
Even if the page can't be parsed it shouldn't cause a segfault.

The plasmashell issue (i. e. that an applet can cause it to crash) was
originally reported here:
https://bugs.kde.org/show_bug.cgi?id=370194

But plasma maintainers tagged the bug as invalid.

-- 
You are receiving this mail because:
You are watching all bug changes.

[plasmashell] [Bug 368660] Wireless and aeroplane mode Oxygen icons missing in plasma-nm

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=368660

--- Comment #36 from Maximiliano Curia <m...@debian.org> ---
I see the commit that adds the missing icons to Air. What about Oxygen?

Also, the icon-themes (at least the ones in /usr/share/icons/*) are
inheritable, wouldn't it make sense to inherit from default for air and oxygen
desktopthemes?

-- 
You are receiving this mail because:
You are watching all bug changes.

[plasma-nm] [Bug 370541] New: Missing icons using the oxygen icon theme

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=370541

Bug ID: 370541
   Summary: Missing icons using the oxygen icon theme
   Product: plasma-nm
   Version: 5.8.0
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: applet
  Assignee: jgrul...@redhat.com
  Reporter: m...@debian.org
CC: lu...@kde.org

This bug was originally reported in the Debian bugtracker:
https://bugs.debian.org/839715

When using the oxygen icon theme some icons are missing. It might be worth
adding some text/tooltip if the icon is not present.


Reproducible: Always

Steps to Reproduce:
1. Using systemsettings set the icon theme to oxygen
2. Restart the plasma session
3. Open the plasma-nm dialog

Actual Results:  
No icon is shown for the airplane mode/flightmode checkbox

Expected Results:  
An icon/text or at least a tooltip in the "anonymous" checkbox is shown.

The code seems to refer to network-flightmode-on and network-flightmode-off
icons, which I couldn't find in breeze or hicolor, but it works fine using
breeze at least. Where are these icons located? Are these stored with an offset
in a larger file?

-- 
You are receiving this mail because:
You are watching all bug changes.

[ksmserver] [Bug 370528] New: Autostart, stderr is closed if daemonized (conky)

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=370528

Bug ID: 370528
   Summary: Autostart, stderr is closed if daemonized (conky)
   Product: ksmserver
   Version: 5.8.0
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: minor
  Priority: NOR
 Component: general
  Assignee: l.lu...@kde.org
  Reporter: m...@debian.org

This seems to be related to the move of the autostart code from klauncher to
ksmserver.

Some programs expect a valid stderr (to send their errors to
~/.xsession-errors) even after daemonizing themselves.

Reports about this can be found:
 https://bbs.archlinux.org/viewtopic.php?id=217920
 https://bugs.debian.org/840478

Avoiding daemonization, or redirecting stderr solves the problem, but still
plasma 5.8 breaks backwards compatibility with these programs.


Reproducible: Always

Steps to Reproduce:
1. Install conky
2. Add an autostart desktop file ~/.config/autostart/conky.desktop with:
 [Desktop Entry]
 Exec=/usr/bin/conky -d
 Icon=system-run
 Path=
 Terminal=false
 Type=application
3. Restart your session

Actual Results:  
conky is not running

Expected Results:  
conky is running

This worked fine with previous plasma versions.

-- 
You are receiving this mail because:
You are watching all bug changes.

[plasmashell] [Bug 370194] Random plasmoids can make plasmashell to crash

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=370194

--- Comment #3 from Maximiliano Curia <m...@debian.org> ---
(In reply to Maximiliano Curia from comment #2)
>  - Go back in the comic history till you hit
> http://phdcomics.com/comics/archive.php?comicid=188

Typo, it's actually: http://phdcomics.com/comics/archive.php?comicid=1889

-- 
You are receiving this mail because:
You are watching all bug changes.

[plasmashell] [Bug 370194] Random plasmoids can make plasmashell to crash

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=370194

--- Comment #2 from Maximiliano Curia <m...@debian.org> ---
(In reply to David Edmundson from comment #1)
> If you have a specific crash we'll fix it.

> A bug report saying an application should not crash is not a useful report.

There is specific crash in the Debian bug report:
 - Add the comic plasmoid
 - Install the phdcomics in the comic plasmoid
 - Go back in the comic history till you hit
http://phdcomics.com/comics/archive.php?comicid=188

This causes a segfault in plasmashell, probably caused in
plasma_applet_comic.so

The problem here is that while it might be straight forward to fix the
plasma_applet to avoid the crash, the underlying issue is that the
plasma_applet_comic is running in the plasmashell process, and as such it's
impossible to avoid a crashing applet to crash plasmashell.

I don't expect this issue to be solved easily, maybe only considered for the
design of plasma 6. But I wasn't expect the issue to be closed as invalid, when
it's clearly reproducible.

-- 
You are receiving this mail because:
You are watching all bug changes.

[plasmashell] [Bug 370194] New: Random plasmoids can make plasmashell to crash

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=370194

Bug ID: 370194
   Summary: Random plasmoids can make plasmashell to crash
   Product: plasmashell
   Version: 5.8.0
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: crash
  Priority: NOR
 Component: general
  Assignee: k...@davidedmundson.co.uk
  Reporter: m...@debian.org
CC: bhus...@gmail.com, plasma-b...@kde.org

plasmashell currently provides most of what the users consider their desktop, a
plasmashell crash leaves the user with blank screen and little or no way to
recover it. Sadly, crashing plasmashell is easier than it should, for example:

https://bugs.debian.org/839710


Reproducible: Always



Expected Results:  
plasmashell should be resilient to crashing plasmoids.

-- 
You are receiving this mail because:
You are watching all bug changes.

[konsole] [Bug 359669] If the configured font is not present, konsole show a black window

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=359669

--- Comment #2 from Maximiliano Curia <m...@debian.org> ---
Sorry, it was a long time ago, and I don't remember which font I've used. But
before I forget, I'm currently using:

For system fonts the fonts-3270 and fonts-anonymous-pro packages, and for user
installed fonts the ones provided in https://github.com/ryanoasis/nerd-fonts/

In particular:
https://github.com/ryanoasis/nerd-fonts/blob/master/patched-fonts/Monofur/Regular/complete/monofur%20for%20Powerline%20Nerd%20Font%20Complete%20Mono.ttf

I'm testing konsole 16.08.0 and the issue seems almost completely gone, for the
fonts I've tested at least. When working with a system font, after removal
konsole fallsback to Monospace. But, when the font is installed in the user
home then after removal the fallback can be any font, and the font metrics for
the fallback seem to be  wrong (the glyphs are cut).

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwalletmanager] [Bug 362805] KF5Wallet can't open previuosly created wallet with error code -9

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=362805

Maximiliano Curia <m...@debian.org> changed:

   What|Removed |Added

 CC||m...@debian.org

--- Comment #17 from Maximiliano Curia <m...@debian.org> ---
With the addtion of the test case in:
https://quickgit.kde.org/?p=kwallet.git=commitdiff=7e7644b608e22a13e110284ed6c5426c1b493b43

It's clear now that the blowfish backend doesn't work in big endian, as shown
in the buildds logs for mips and powerpc in Debian:
https://buildd.debian.org/status/package.php?p=kwallet-kf5

Adding back the include:
#include 

and changing the checks for #if Q_BYTE_ORDER == Q_LITTLE_ENDIAN
(effectively reverting the blowfish backend to the state in 
https://quickgit.kde.org/?p=kwallet.git=blob=e09d5422dfb404f2f15e74d693b24bb6cd8e1689=87e774825b779ba846315a8b2ffe6479dd9f9814=src%2Fruntime%2Fkwalletd%2Fbackend%2Fblowfish.cc)
makes the test pass in little and big endian machines.

This changes nothing for little endian machines. And in big endian machines it
should fix the blowfish backend.

Happy hacking,

-- 
You are receiving this mail because:
You are watching all bug changes.

[phonon-backend-vlc] [Bug 363240] New: List of mime types do not match the ones in the vlc desktop file

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=363240

Bug ID: 363240
   Summary: List of mime types do not match the ones in the vlc
desktop file
   Product: phonon-backend-vlc
   Version: 0.9-git
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: general
  Assignee: sit...@kde.org
  Reporter: m...@debian.org
CC: bcooks...@kde.org, f...@kde.org, j...@videolan.org,
myr...@kde.org

Forwarding the Debian bug #824695 originally reported by Petter Reinholdsten
(http://bugs.debian.org/824695)

When comparing the list of MIME types in
phonon-backend-vlc-0.8.2/MimeTypes.cmake with the file
share/vlc.desktop.mimetypes in the vlc upstream git, the list of MIME
types do mot match each other.  These ones are missing in the phonon
backend build rules:

application/mpeg4-iod
application/mpeg4-muxcodetable
application/mxf
application/ram
application/sdp
application/vnd.apple.mpegurl
application/vnd.ms-asf
application/vnd.ms-wpl
application/vnd.rn-realmedia-vbr
application/x-extension-m4a
application/x-flac
application/x-matroska
application/x-quicktime-media-link
application/x-shockwave-flash
application/xspf+xml
audio/aac
audio/ac3
audio/AMR
audio/AMR-WB
audio/dv
audio/eac3
audio/flac
audio/m4a
audio/mp1
audio/mp2
audio/mpegurl
audio/mpg
audio/ogg
audio/opus
audio/scpls
audio/vnd.dolby.heaac.1
audio/vnd.dolby.heaac.2
audio/vnd.dolby.mlp
audio/vnd.dts
audio/vnd.dts.hd
audio/x-aac
audio/x-gsm
audio/x-mp1
audio/x-mp2
audio/x-mpg
audio/x-ms-asf
audio/x-ms-asx
audio/x-ms-wax
audio/x-musepack
audio/x-pn-realaudio
audio/x-scpls
audio/x-shorten
audio/x-speex
audio/x-tta
audio/x-vorbis
audio/x-wavpack
image/vnd.rn-realpix
misc/ultravox
text/google-video-pointer
text/x-google-video-pointer
video/3gp
video/dv
video/fli
video/mp2t
video/mp4v-es
video/vnd.divx
video/vnd.mpegurl
video/vnd.rn-realvideo
video/x-avi
video/x-flc
video/x-fli
video/x-mpeg2
video/x-ms-asf-plugin
video/x-ms-asx
video/x-ms-wm
video/x-ms-wmx
video/x-nsv
video/x-ogm+ogg
video/x-theora
video/x-theora+ogg
x-content/audio-cdda
x-content/audio-player
x-content/video-dvd
x-content/video-svcd
x-content/video-vcd
x-scheme-handler/icy
x-scheme-handler/icyx
x-scheme-handler/mms
x-scheme-handler/mmsh
x-scheme-handler/rtmp
x-scheme-handler/rtp
x-scheme-handler/rtsp

While these are missing in the vlc upstream git reposiory:

application/x-annodex
audio/168sv
audio/8svx
audio/aiff
audio/mpeg2
audio/mpeg3
audio/prs.sid
audio/vnd.rn-realmedia
audio/x-16sv
audio/x-8svx
audio/x-basic
audio/x-mpeg2
audio/x-mpeg3
audio/x-ogg
audio/x-speex+ogg
image/ilbm
image/png
image/x-ilbm
image/x-png
video/anim
video/avi
video/mkv
video/mng
video/mpg
video/x-flic
video/x-mng
video/x-ms-wma
video/x-quicktime

Should these two list be synced?

The list of MIME types in the upstream git repository was recently
extended, but I was not aware of the list of MIME types in the phonon
backend when doing so.

--
Happy hacking
Petter Reinholdtsen

Reproducible: Always



Expected Results:  
The list of supported mime types should be the same as vlc.

-- 
You are receiving this mail because:
You are watching all bug changes.

[cantor] [Bug 359918] New: Licensing issues

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=359918

Bug ID: 359918
   Summary: Licensing issues
   Product: cantor
   Version: unspecified
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: general
  Assignee: fil...@kde.org
  Reporter: m...@debian.org
CC: alexanderrie...@gmail.com

The files src/backends/qalculate/settingswidget.cpp and
src/backends/qalculate/settingswidget.h are licensed under GPL-3+ while most of
the code is licensed GPL-2+, this make the complete backend effectively GPL-3
(for ~10 lines of code), at the same time the GPL-3+ license requires the
distribution of the full license together with the source code and the included
COPYING file is a GPL-2 license.

To fix this issue we could either re-license the mentioned files (either with
the permission of Matteo Agostinelli or rewriting them), to match the license
of the rest of the code, or we could add the missing COPYING.gpl3 file.

Note: This is considered a serious bug in Debian (as can be seen in
http://bugs.debian.org/792078) as it violates the Debian Policy.

Happy hacking,

Reproducible: Always

-- 
You are receiving this mail because:
You are watching all bug changes.

[konsole] [Bug 359669] New: If the configured font is not present, konsole show a black window

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=359669

Bug ID: 359669
   Summary: If the configured font is not present, konsole show a
black window
   Product: konsole
   Version: 15.12.1
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: font
  Assignee: konsole-de...@kde.org
  Reporter: m...@debian.org

I'm forwarding the Debian bug reported: https://bugs.debian.org/815559

If the font configured is missing konsole shows a $BACKGROUND colored window,
and stays in an almost non responsive mode till you change the font (by non
responsive I mean, it takes a long time for the menus to appear, or the
setttings window to change tabs).

It would be better if konsole could fallback to some other fonts, either using
the one configured as the applications wide monospace font, or simply trying
the "Monospace" font, or using any of the available fonts should be enough.

Happy hacking,

Reproducible: Always

Steps to Reproduce:
1. Install a new monospace font
2. Start konsole and configure it to use the newly installed font
3. Stop konsole
4. Deinstall the font
5. Start konsole again

Actual Results:  
A black screen is shown, konsole is almost non responsive

Expected Results:  
konsole starts with the default font, or fallbacks to any other available font.

-- 
You are receiving this mail because:
You are watching all bug changes.

[kdesu] [Bug 359621] New: Unneeded setgid requirement

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=359621

Bug ID: 359621
   Summary: Unneeded setgid requirement
   Product: kdesu
   Version: unspecified
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: kdesud
  Assignee: kdesu-bugs-n...@kde.org
  Reporter: m...@debian.org

Hi,

The kdesu framework currently requires the kdesud to be setgid, the
documentation about this requirement says (client.h):
 The daemon should be installed setgid nogroup, in order to be able to act as
an inaccessible,
 trusted 3rd party.

Even the check for the daemon file to be setgid is part of the public API of
the kdesu framework.:
class KDESU_EXPORT KDEsuClient
{
public:
 ...
 bool isServerSGID();
...

But, AFAICS, this provides no additional "security". In fact, it would be
better if the check were "make sure the daemon is not setuid", or if it denies
ptrace.

Afaik, having the setgid in place only serves as a way to change the effective
primary group, which will be used for files created by this process.

Unless there is a real reason for this requirement, please drop it.

Happy hacking,

Reproducible: Always

-- 
You are receiving this mail because:
You are watching all bug changes.

[frameworks-kinit] [Bug 359169] New: man page lists $ in front of namess of environmental variables used

[Maximiliano Curia via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=359169

Bug ID: 359169
   Summary: man page lists $ in front of namess of environmental
variables used
   Product: frameworks-kinit
   Version: unspecified
  Platform: Debian unstable
OS: Linux
Status: UNCONFIRMED
  Severity: wishlist
  Priority: NOR
 Component: general
  Assignee: fa...@kde.org
  Reporter: m...@debian.org
CC: kdelibs-b...@kde.org

(Forwarding debian https://bugs.debian.org/811165)
I read on the manual page for kdeinit5:

ENVIRONMENT VARIABLES
   $HOME
   Specifies the home directory of the current user

   $KDE_HOME_READONLY
   Specifies if the home directory of the current user is read only

   $KDE_IS_PRELINKED
   If set, tells kdeinit5 that the KDE programs are pre-linked.

   (Prelinking is a process that allows you to speed up the process of
   dynamic linking.)

However they should be listed without the evaluation operator $ at the start.

Reproducible: Always

-- 
You are receiving this mail because:
You are watching all bug changes.