Re: gcc-5 transition of GDCM

[Aron Xu]

Hi,

I orphaned it today, so whoever would like to take care it can step up
directly. The package is simple and straghtforwrad, shouldn't be a lot
of trouble.

Regards,
Aron

On Fri, Aug 14, 2015 at 5:13 AM,   wrote:
> Hi
>
> thanks for your insights. I guess python-pygccxml is not directly needed
> here in Debian Med. On ITK’s side, pygccxml is present in ITK’s repo
> as a subtree. So it is not needed as a dependency, and there is no plan
> to change this.
>
> I’ll leave the decision up to you. I personally recommend people to install
> it through pip, which is the fastest way to get it. Having a debian package of
> course better; maybe as there is already one it would only need an update
> to the newest version.
>
> I am not ready to maintain the package myself for the moment, as I have
> other ongoing projects. But I can help if there are any questions.
> Tell me if you want to update the package; I’ll send a mail to Aron
> as soon as the new version is out.
>
> Regards
>
> Michka
>
>> On 13 Aug 2015, at 13:30, Aron Xu  wrote:
>>
>> Actually I'm seeking to orphan the package as I don't use it myself
>> anymore, but it's not a burden so far so that I haven't done it yet.
>> If there are people interested in adopting or co-maintaining it'll be
>> appreciated. There's no any form of repository at the moment and I can
>> create one if no one else is more interested, :)
>>
>> Thanks,
>> Aron
>>
>> On Thu, Aug 13, 2015 at 3:29 PM, Andreas Tille  wrote:
>>> Hi Michka,
>>>
>>> I keep the Debian maintainer Aron Xu in CC.  Gccxml was maintained as a
>>> predependency of some Debian Med package by the according team.  I'm not
>>> aware that python-pygccxml has any reverse dependencies in Debian Med
>>> but surely ITK is of interest.  Could the ITK people please raise their
>>> opinion whether it would be good to team maintain this package here as
>>> well.  Aron, would you consider using some common Git repository either
>>> in Debian Med or collab-maint?
>>>
>>> Kind regards
>>>
>>> Andreas.
>>>
>>> On Thu, Aug 13, 2015 at 08:37:13AM +0200, michkapop...@gmail.com wrote:
> On August 5, 2015 03:29:30 PM Sébastien Jodogne wrote:
>
>>> I will try hard to get the script into VCS tonight and then I will 
>>> welcome
>>> help in testing the builds on dependent packages.
>>
>> Great! Let me know when things are ready.
>
> OK.  The gccxml package is updated with a wrapper around castxml (the real
> binary is renamed to gccxml.real).
>
> Still working on testing build-deps with it.  Sadly, pygccxml fails :-(
>
> -Steve
> 

 Hi

 I was told someone mentioned pygccxml here :)

 I am the current maintainer of the project 
 (https://github.com/gccxml/pygccxml ).
 During the last months I worked on the integration of CastXML in pygccxml 
 and I am
 almost done. Basically, there were around 80 unit tests failing, and I am 
 down
 to 10 failing tests. The core unit tests are running, and ITK already uses 
 pygccxml
 internally. So as long as one does not do fancy stuff, it should be ok.

 I delayed the new release in the hope to be able to fix the remaining 
 tests; but this
 takes a lot of time (as I did not write pygccxml in the first place, I am 
 discovering
 the code). But I think there is some pressure to release a new version 
 with basic
 support so people can start using it. I even heard from 1 or 2 people 
 which are using
 the current unstable version in combination with pyplusplus to build their 
 project
 using CastXML instead of GccXML.

 I will publish a new pygccxml version (1.7.0) in the next weeks. The 
 package name
 will not change, and is backwards compatible with gccxml. I will 
 temporarily disable
 the failing tests when running with CastXML.

 Please tell me if I need to notify someone when the new version is out or 
 if there
 is something I can to to help the packaging. pygccxml will also be 
 available
 through pip.

 Regards

 Michka




>>>
>>> --
>>> http://fam-tille.de
>>
>>
>>
>> --
>> Regards,
>> Aron Xu
>



-- 
Regards,
Aron Xu

Re: gcc-5 transition of GDCM

[michkapopoff]

Hi

thanks for your insights. I guess python-pygccxml is not directly needed
here in Debian Med. On ITK’s side, pygccxml is present in ITK’s repo
as a subtree. So it is not needed as a dependency, and there is no plan
to change this.

I’ll leave the decision up to you. I personally recommend people to install
it through pip, which is the fastest way to get it. Having a debian package of
course better; maybe as there is already one it would only need an update
to the newest version.

I am not ready to maintain the package myself for the moment, as I have
other ongoing projects. But I can help if there are any questions.
Tell me if you want to update the package; I’ll send a mail to Aron
as soon as the new version is out.

Regards

Michka

> On 13 Aug 2015, at 13:30, Aron Xu  wrote:
> 
> Actually I'm seeking to orphan the package as I don't use it myself
> anymore, but it's not a burden so far so that I haven't done it yet.
> If there are people interested in adopting or co-maintaining it'll be
> appreciated. There's no any form of repository at the moment and I can
> create one if no one else is more interested, :)
> 
> Thanks,
> Aron
> 
> On Thu, Aug 13, 2015 at 3:29 PM, Andreas Tille  wrote:
>> Hi Michka,
>> 
>> I keep the Debian maintainer Aron Xu in CC.  Gccxml was maintained as a
>> predependency of some Debian Med package by the according team.  I'm not
>> aware that python-pygccxml has any reverse dependencies in Debian Med
>> but surely ITK is of interest.  Could the ITK people please raise their
>> opinion whether it would be good to team maintain this package here as
>> well.  Aron, would you consider using some common Git repository either
>> in Debian Med or collab-maint?
>> 
>> Kind regards
>> 
>> Andreas.
>> 
>> On Thu, Aug 13, 2015 at 08:37:13AM +0200, michkapop...@gmail.com wrote:
 On August 5, 2015 03:29:30 PM Sébastien Jodogne wrote:
 
>> I will try hard to get the script into VCS tonight and then I will 
>> welcome
>> help in testing the builds on dependent packages.
> 
> Great! Let me know when things are ready.
 
 OK.  The gccxml package is updated with a wrapper around castxml (the real
 binary is renamed to gccxml.real).
 
 Still working on testing build-deps with it.  Sadly, pygccxml fails :-(
 
 -Steve
 
>>> 
>>> Hi
>>> 
>>> I was told someone mentioned pygccxml here :)
>>> 
>>> I am the current maintainer of the project 
>>> (https://github.com/gccxml/pygccxml ).
>>> During the last months I worked on the integration of CastXML in pygccxml 
>>> and I am
>>> almost done. Basically, there were around 80 unit tests failing, and I am 
>>> down
>>> to 10 failing tests. The core unit tests are running, and ITK already uses 
>>> pygccxml
>>> internally. So as long as one does not do fancy stuff, it should be ok.
>>> 
>>> I delayed the new release in the hope to be able to fix the remaining 
>>> tests; but this
>>> takes a lot of time (as I did not write pygccxml in the first place, I am 
>>> discovering
>>> the code). But I think there is some pressure to release a new version with 
>>> basic
>>> support so people can start using it. I even heard from 1 or 2 people which 
>>> are using
>>> the current unstable version in combination with pyplusplus to build their 
>>> project
>>> using CastXML instead of GccXML.
>>> 
>>> I will publish a new pygccxml version (1.7.0) in the next weeks. The 
>>> package name
>>> will not change, and is backwards compatible with gccxml. I will 
>>> temporarily disable
>>> the failing tests when running with CastXML.
>>> 
>>> Please tell me if I need to notify someone when the new version is out or 
>>> if there
>>> is something I can to to help the packaging. pygccxml will also be available
>>> through pip.
>>> 
>>> Regards
>>> 
>>> Michka
>>> 
>>> 
>>> 
>>> 
>> 
>> --
>> http://fam-tille.de
> 
> 
> 
> -- 
> Regards,
> Aron Xu

Re: gcc-5 transition of GDCM

[Aron Xu]

Actually I'm seeking to orphan the package as I don't use it myself
anymore, but it's not a burden so far so that I haven't done it yet.
If there are people interested in adopting or co-maintaining it'll be
appreciated. There's no any form of repository at the moment and I can
create one if no one else is more interested, :)

Thanks,
Aron

On Thu, Aug 13, 2015 at 3:29 PM, Andreas Tille  wrote:
> Hi Michka,
>
> I keep the Debian maintainer Aron Xu in CC.  Gccxml was maintained as a
> predependency of some Debian Med package by the according team.  I'm not
> aware that python-pygccxml has any reverse dependencies in Debian Med
> but surely ITK is of interest.  Could the ITK people please raise their
> opinion whether it would be good to team maintain this package here as
> well.  Aron, would you consider using some common Git repository either
> in Debian Med or collab-maint?
>
> Kind regards
>
>  Andreas.
>
> On Thu, Aug 13, 2015 at 08:37:13AM +0200, michkapop...@gmail.com wrote:
>> > On August 5, 2015 03:29:30 PM Sébastien Jodogne wrote:
>> >
>> >>> I will try hard to get the script into VCS tonight and then I will 
>> >>> welcome
>> >>> help in testing the builds on dependent packages.
>> >>
>> >> Great! Let me know when things are ready.
>> >
>> > OK.  The gccxml package is updated with a wrapper around castxml (the real
>> > binary is renamed to gccxml.real).
>> >
>> > Still working on testing build-deps with it.  Sadly, pygccxml fails :-(
>> >
>> > -Steve
>> > 
>>
>> Hi
>>
>> I was told someone mentioned pygccxml here :)
>>
>> I am the current maintainer of the project 
>> (https://github.com/gccxml/pygccxml ).
>> During the last months I worked on the integration of CastXML in pygccxml 
>> and I am
>> almost done. Basically, there were around 80 unit tests failing, and I am 
>> down
>> to 10 failing tests. The core unit tests are running, and ITK already uses 
>> pygccxml
>> internally. So as long as one does not do fancy stuff, it should be ok.
>>
>> I delayed the new release in the hope to be able to fix the remaining tests; 
>> but this
>> takes a lot of time (as I did not write pygccxml in the first place, I am 
>> discovering
>> the code). But I think there is some pressure to release a new version with 
>> basic
>> support so people can start using it. I even heard from 1 or 2 people which 
>> are using
>> the current unstable version in combination with pyplusplus to build their 
>> project
>> using CastXML instead of GccXML.
>>
>> I will publish a new pygccxml version (1.7.0) in the next weeks. The package 
>> name
>> will not change, and is backwards compatible with gccxml. I will temporarily 
>> disable
>> the failing tests when running with CastXML.
>>
>> Please tell me if I need to notify someone when the new version is out or if 
>> there
>> is something I can to to help the packaging. pygccxml will also be available
>> through pip.
>>
>> Regards
>>
>> Michka
>>
>>
>>
>>
>
> --
> http://fam-tille.de



-- 
Regards,
Aron Xu

Re: gcc-5 transition of GDCM

[Andreas Tille]

Hi Michka,

I keep the Debian maintainer Aron Xu in CC.  Gccxml was maintained as a
predependency of some Debian Med package by the according team.  I'm not
aware that python-pygccxml has any reverse dependencies in Debian Med
but surely ITK is of interest.  Could the ITK people please raise their
opinion whether it would be good to team maintain this package here as
well.  Aron, would you consider using some common Git repository either
in Debian Med or collab-maint?

Kind regards

 Andreas.

On Thu, Aug 13, 2015 at 08:37:13AM +0200, michkapop...@gmail.com wrote:
> > On August 5, 2015 03:29:30 PM Sébastien Jodogne wrote:
> > 
> >>> I will try hard to get the script into VCS tonight and then I will welcome
> >>> help in testing the builds on dependent packages.
> >> 
> >> Great! Let me know when things are ready.
> > 
> > OK.  The gccxml package is updated with a wrapper around castxml (the real
> > binary is renamed to gccxml.real).
> > 
> > Still working on testing build-deps with it.  Sadly, pygccxml fails :-(
> > 
> > -Steve
> > 
> 
> Hi
> 
> I was told someone mentioned pygccxml here :)
> 
> I am the current maintainer of the project 
> (https://github.com/gccxml/pygccxml ).
> During the last months I worked on the integration of CastXML in pygccxml and 
> I am
> almost done. Basically, there were around 80 unit tests failing, and I am down
> to 10 failing tests. The core unit tests are running, and ITK already uses 
> pygccxml
> internally. So as long as one does not do fancy stuff, it should be ok.
> 
> I delayed the new release in the hope to be able to fix the remaining tests; 
> but this
> takes a lot of time (as I did not write pygccxml in the first place, I am 
> discovering
> the code). But I think there is some pressure to release a new version with 
> basic
> support so people can start using it. I even heard from 1 or 2 people which 
> are using
> the current unstable version in combination with pyplusplus to build their 
> project
> using CastXML instead of GccXML.
> 
> I will publish a new pygccxml version (1.7.0) in the next weeks. The package 
> name
> will not change, and is backwards compatible with gccxml. I will temporarily 
> disable
> the failing tests when running with CastXML.
> 
> Please tell me if I need to notify someone when the new version is out or if 
> there
> is something I can to to help the packaging. pygccxml will also be available
> through pip.
> 
> Regards
> 
> Michka
> 
> 
> 
> 

-- 
http://fam-tille.de

Re: gcc-5 transition of GDCM

[michkapopoff]

> On August 5, 2015 03:29:30 PM Sébastien Jodogne wrote:
> 
>>> I will try hard to get the script into VCS tonight and then I will welcome
>>> help in testing the builds on dependent packages.
>> 
>> Great! Let me know when things are ready.
> 
> OK.  The gccxml package is updated with a wrapper around castxml (the real
> binary is renamed to gccxml.real).
> 
> Still working on testing build-deps with it.  Sadly, pygccxml fails :-(
> 
> -Steve
> 

Hi

I was told someone mentioned pygccxml here :)

I am the current maintainer of the project (https://github.com/gccxml/pygccxml 
).
During the last months I worked on the integration of CastXML in pygccxml and I 
am
almost done. Basically, there were around 80 unit tests failing, and I am down
to 10 failing tests. The core unit tests are running, and ITK already uses 
pygccxml
internally. So as long as one does not do fancy stuff, it should be ok.

I delayed the new release in the hope to be able to fix the remaining tests; 
but this
takes a lot of time (as I did not write pygccxml in the first place, I am 
discovering
the code). But I think there is some pressure to release a new version with 
basic
support so people can start using it. I even heard from 1 or 2 people which are 
using
the current unstable version in combination with pyplusplus to build their 
project
using CastXML instead of GccXML.

I will publish a new pygccxml version (1.7.0) in the next weeks. The package 
name
will not change, and is backwards compatible with gccxml. I will temporarily 
disable
the failing tests when running with CastXML.

Please tell me if I need to notify someone when the new version is out or if 
there
is something I can to to help the packaging. pygccxml will also be available
through pip.

Regards

Michka

Re: gcc-5 transition of GDCM

[Andreas Tille]

Hi Sébastien,

On Fri, Aug 07, 2015 at 11:48:06AM +0200, Sébastien Jodogne wrote:
> 
> Thanks for this proposition! I would of course be very happy to be able to 
> upload the orthanc-related packages by myself. Orthanc 0.9.3 will soon be 
> released and updates to orthanc-webviewer/orthanc-postgresql are pending, so 
> it could be a nice timing to experience the upload process.
> 
> Is there something I should do to apply?

   https://wiki.debian.org/DebianMaintainer

This page might be relevant for some other people with recent activities
as well. 

Kind regards

  Andreas.

-- 
http://fam-tille.de


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20150807102254.gi22...@an3as.eu

Re: gcc-5 transition of GDCM

[Sébastien Jodogne]

> > BTW, thanks for uploading the package!
> 
> The thank goes fully to you.  Just rebuilding and calling dput is not a
> big deal.
> 
> BTW, do you want to get the upload feeling yourself?  You are now
> maintaining several packages and have dived into the bug squashing
> business.  So applying to become a DM seems a natural step in my
> opinion.

Thanks for this proposition! I would of course be very happy to be able to 
upload the orthanc-related packages by myself. Orthanc 0.9.3 will soon be 
released and updates to orthanc-webviewer/orthanc-postgresql are pending, so it 
could be a nice timing to experience the upload process.

Is there something I should do to apply?

Regards,
Sébastien-


--
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/1922141197.2967295.1438940886984.javamail.zim...@chu.ulg.ac.be

Re: gcc-5 transition of GDCM

[Andreas Tille]

On Fri, Aug 07, 2015 at 10:53:13AM +0200, Sébastien Jodogne wrote:
> 
> No problem with this: I just try and centralize all the Debian-related 
> automated mailing in a single mailbox, but this can certainly wait the next 
> version of the package :)

Uploaders do not get any mail anyway.  It might just be that spammers
fetch your other address - but it was visible for me anyway and so it is
for spammers as well.
 
> BTW, thanks for uploading the package!

The thank goes fully to you.  Just rebuilding and calling dput is not a
big deal. 

BTW, do you want to get the upload feeling yourself?  You are now
maintaining several packages and have dived into the bug squashing
business.  So applying to become a DM seems a natural step in my
opinion.

Kind regards

 Andreas.

-- 
http://fam-tille.de


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20150807091813.gh22...@an3as.eu

Re: gcc-5 transition of GDCM

[Sébastien Jodogne]

Hi Andreas,

> I can confirm that I uploaded gdcm now to unstable.  Unfortunately I
> missed the last change of your e-mail address as Uploader.  I hope this
> will be no problem for you.

No problem with this: I just try and centralize all the Debian-related 
automated mailing in a single mailbox, but this can certainly wait the next 
version of the package :)

BTW, thanks for uploading the package!

Sébastien-


--
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/1535640774.2948017.1438937593140.javamail.zim...@chu.ulg.ac.be

Re: gcc-5 transition of GDCM

[Andreas Tille]

Hi,

On Fri, Aug 07, 2015 at 09:41:39AM +0200, Andreas Tille wrote:
> On Fri, Aug 07, 2015 at 09:38:03AM +0200, Sébastien Jodogne wrote:
> > Hi,
> > 
> > > But it also points out the way to simply replace gccxml by castxml.  So I
> > > updated your patch to do that.
> > 
> > Yes, this is much better now!
> > 
> > > Unfortunately, I get a build failure with gdcm: [...]
> > > How did you avoid the "__int128" problem?  I'm building on adm64 in an 
> > > up-to-
> > > date sid chroot.
> > 
> > This is simply because the argument "--castxml-start=_cable_" should be 
> > split as "--castxml-start _cable_".
> > 
> > I have just committed the fix:
> > http://anonscm.debian.org/viewvc/debian-med?view=revision&revision=19866

I can confirm that I uploaded gdcm now to unstable.  Unfortunately I
missed the last change of your e-mail address as Uploader.  I hope this
will be no problem for you.  Otherwise I'd upload another revision
(while cleaning up some lintian errors) - just tell me.  In any case I
added a changelog entry mentioning it for the next revision.

Thanks a lot for your really helpful work

  Andreas.

-- 
http://fam-tille.de


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20150807084827.gd22...@an3as.eu

Re: gcc-5 transition of GDCM

[Andreas Tille]

Ahh, I'll try again - our mails have crossed

On Fri, Aug 07, 2015 at 09:38:03AM +0200, Sébastien Jodogne wrote:
> Hi,
> 
> > But it also points out the way to simply replace gccxml by castxml.  So I
> > updated your patch to do that.
> 
> Yes, this is much better now!
> 
> > Unfortunately, I get a build failure with gdcm: [...]
> > How did you avoid the "__int128" problem?  I'm building on adm64 in an 
> > up-to-
> > date sid chroot.
> 
> This is simply because the argument "--castxml-start=_cable_" should be split 
> as "--castxml-start _cable_".
> 
> I have just committed the fix:
> http://anonscm.debian.org/viewvc/debian-med?view=revision&revision=19866
> 
> Sébastien-
> 

-- 
http://fam-tille.de


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20150807074139.gc22...@an3as.eu

Re: gcc-5 transition of GDCM

[Andreas Tille]

Hi,

On Fri, Aug 07, 2015 at 01:40:16AM -0500, Steve M. Robbins wrote:
> [ 72%] Generating csharp/vtkGDCMImageReader-sentinel.txt
> cd /home/steve/Packages/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-
> gnu/Utilities/VTK && /usr/bin/mummy --suppress-warnings 6005 6006 6009 6010 
> 6012 6013 6015 6016 6017 6018 6019 --settings-file 
> /home/steve/Packages/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-
> gnu/Utilities/VTK/xml/MummySettings.xml --gccxml-file 
> /home/steve/Packages/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-
> gnu/Utilities/VTK/xml/vtkGDCMImageReader.xml --csharp-file 
> /home/steve/Packages/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-
> gnu/Utilities/VTK/csharp/vtkGDCMImageReader.cs --export-layer-file 
> /home/steve/Packages/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-
> gnu/Utilities/VTK/export-layer/vtkGDCMImageReaderEL.cxx
> ERROR: In 
> /tmp/buildd/cableswig-0.1.0+cvs20111009/Cable/Parsers/cableFundamentalType.cxx,
>  
> line 99
> FundamentalType (0x302a3c0): Invalid type name "__int128" for FundamentalType.
> 
> ERROR: In 
> /tmp/buildd/cableswig-0.1.0+cvs20111009/Cable/Parsers/cableXMLSourceParser.cxx,
>  
> line 138
> XMLSourceParser (0x1f43b30): Error creating source representation.
> 
> # (0): error: m5008: MummyApplication::ProcessSource
> Utilities/VTK/CMakeFiles/vtkgdcmGenerateWrappers.dir/build.make:70: recipe 
> for 
> target 'Utilities/VTK/csharp/vtkGDCMImageReader-sentinel.txt' failed

I can confirm

[ 72%] Generating csharp/vtkGDCMImageReader-sentinel.txt
cd /tmp/buildd/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK && /usr/bin/mummy 
--suppress-warnings 6005 6006 6009 6010 6012 6013 6015 6016 6017 6018 6019 
--settings-file 
/tmp/buildd/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/xml/MummySettings.xml 
--gccxml-file 
/tmp/buildd/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/xml/vtkGDCMImageReader.xml
 --csharp-file 
/tmp/buildd/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/csharp/vtkGDCMImageReader.cs
 --export-layer-file 
/tmp/buildd/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/export-layer/vtkGDCMImageReaderEL.cxx
ERROR: In 
/tmp/buildd/cableswig-0.1.0+cvs20111009/Cable/Parsers/cableFundamentalType.cxx, 
line 99
FundamentalType (0x286b1a0): Invalid type name "__int128" for FundamentalType.

ERROR: In 
/tmp/buildd/cableswig-0.1.0+cvs20111009/Cable/Parsers/cableXMLSourceParser.cxx, 
line 138
XMLSourceParser (0x1784a80): Error creating source representation.

# (0): error: m5008: MummyApplication::ProcessSource
Utilities/VTK/CMakeFiles/vtkgdcmGenerateWrappers.dir/build.make:70: recipe for 
target 'Utilities/VTK/csharp/vtkGDCMImageReader-sentinel.txt' failed
make[3]: *** [Utilities/VTK/csharp/vtkGDCMImageReader-sentinel.txt] Error 144
make[3]: Leaving directory '/tmp/buildd/gdcm-2.4.4/obj-x86_64-linux-gnu'
CMakeFiles/Makefile2:815: recipe for target 
'Utilities/VTK/CMakeFiles/vtkgdcmGenerateWrappers.dir/all' failed
make[2]: *** [Utilities/VTK/CMakeFiles/vtkgdcmGenerateWrappers.dir/all] Error 2
make[2]: Leaving directory '/tmp/buildd/gdcm-2.4.4/obj-x86_64-linux-gnu'
Makefile:139: recipe for target 'all' failed
make[1]: *** [all] Error 2
make[1]: Leaving directory '/tmp/buildd/gdcm-2.4.4/obj-x86_64-linux-gnu'

 
> 
> How did you avoid the "__int128" problem?  I'm building on adm64 in an up-to-
> date sid chroot.

Same updated sid chroot used for building.
 
> > Andreas, would it be possible for you to review this change and, possibly,
> > close #777868 and #791508 if the requirements are met?
> 
> The patch looks good to me.  Just need to resolve the build problem.

I added Sébastien Jodogne as uploaders.

Kind regards

Andreas.

-- 
http://fam-tille.de


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20150807073950.gb22...@an3as.eu

Re: gcc-5 transition of GDCM

[Sébastien Jodogne]

Hi,

> But it also points out the way to simply replace gccxml by castxml.  So I
> updated your patch to do that.

Yes, this is much better now!

> Unfortunately, I get a build failure with gdcm: [...]
> How did you avoid the "__int128" problem?  I'm building on adm64 in an up-to-
> date sid chroot.

This is simply because the argument "--castxml-start=_cable_" should be split 
as "--castxml-start _cable_".

I have just committed the fix:
http://anonscm.debian.org/viewvc/debian-med?view=revision&revision=19866

Sébastien-


--
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/540431967.2916710.1438933083815.javamail.zim...@chu.ulg.ac.be

Re: gcc-5 transition of GDCM

[Steve M. Robbins]

Hi Sébastien,

Great work!


On August 6, 2015 04:28:37 PM Sébastien Jodogne wrote:

> Sorry for the multiple posts, but thanks to Steve's upload of gccxml last
> evening, I have finally managed to make the transition of the gdcm package
> to gcc-5. A single patch to the CMake scripts is necessary [1].

Hmm..  The patch points out a couple of deficiencies with my script (now 
fixed).  

But it also points out the way to simply replace gccxml by castxml.  So I 
updated your patch to do that.

Unfortunately, I get a build failure with gdcm:

[ 72%] Generating csharp/vtkGDCMImageReader-sentinel.txt
cd /home/steve/Packages/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-
gnu/Utilities/VTK && /usr/bin/mummy --suppress-warnings 6005 6006 6009 6010 
6012 6013 6015 6016 6017 6018 6019 --settings-file 
/home/steve/Packages/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-
gnu/Utilities/VTK/xml/MummySettings.xml --gccxml-file 
/home/steve/Packages/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-
gnu/Utilities/VTK/xml/vtkGDCMImageReader.xml --csharp-file 
/home/steve/Packages/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-
gnu/Utilities/VTK/csharp/vtkGDCMImageReader.cs --export-layer-file 
/home/steve/Packages/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-
gnu/Utilities/VTK/export-layer/vtkGDCMImageReaderEL.cxx
ERROR: In 
/tmp/buildd/cableswig-0.1.0+cvs20111009/Cable/Parsers/cableFundamentalType.cxx, 
line 99
FundamentalType (0x302a3c0): Invalid type name "__int128" for FundamentalType.

ERROR: In 
/tmp/buildd/cableswig-0.1.0+cvs20111009/Cable/Parsers/cableXMLSourceParser.cxx, 
line 138
XMLSourceParser (0x1f43b30): Error creating source representation.

# (0): error: m5008: MummyApplication::ProcessSource
Utilities/VTK/CMakeFiles/vtkgdcmGenerateWrappers.dir/build.make:70: recipe for 
target 'Utilities/VTK/csharp/vtkGDCMImageReader-sentinel.txt' failed


How did you avoid the "__int128" problem?  I'm building on adm64 in an up-to-
date sid chroot.


> Andreas, would it be possible for you to review this change and, possibly,
> close #777868 and #791508 if the requirements are met?

The patch looks good to me.  Just need to resolve the build problem.

Thanks,
-Steve


signature.asc
Description: This is a digitally signed message part.

Re: gcc-5 transition of GDCM

[Andreas Tille]

Hi Sébastien,

On Thu, Aug 06, 2015 at 04:28:37PM +0200, Sébastien Jodogne wrote:
> 
> Sorry for the multiple posts, but thanks to Steve's upload of gccxml last 
> evening, I have finally managed to make the transition of the gdcm package to 
> gcc-5. A single patch to the CMake scripts is necessary [1].
> 
> My packages "orthanc-webviewer" and "orthanc-dicomweb" now also compile 
> correctly, as the gdcm shared libraries uses the same C++11 ABI as the other 
> C++ dependencies (such as libjsoncpp).

I'm *really* happy about your work on this!
 
> Andreas, would it be possible for you to review this change and, possibly, 
> close #777868 and #791508 if the requirements are met?

I'll wait for Steve's opinion since he should have a say.  Steve, either
upload yourself or ping me if you think the upload would be OK. 

Sébastien, thanks a lot

  Andreas.

-- 
http://fam-tille.de


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20150806145448.gz7...@an3as.eu

Re: gcc-5 transition of GDCM

[Sébastien Jodogne]

Hi Andreas,
Hi Steve,

> > As a consequence, I would like to know when updated versions of the
> > libgdcm2.4 and libgdcm2-dev packages built with gcc-5 will be available in
> > the repositories. Is this information available somewhere?
> 
> If you read the Debian Med mailing list you might have noticed Steve's
> status update[1].  The short answer is:  If you will help fixing #777868
> /  #791508 it might happen faster.

Sorry for the multiple posts, but thanks to Steve's upload of gccxml last 
evening, I have finally managed to make the transition of the gdcm package to 
gcc-5. A single patch to the CMake scripts is necessary [1].

My packages "orthanc-webviewer" and "orthanc-dicomweb" now also compile 
correctly, as the gdcm shared libraries uses the same C++11 ABI as the other 
C++ dependencies (such as libjsoncpp).

Andreas, would it be possible for you to review this change and, possibly, 
close #777868 and #791508 if the requirements are met?

Regards,
Sébastien-


[1] 
http://anonscm.debian.org/viewvc/debian-med/trunk/packages/gdcm/trunk/debian/patches/castxml.patch?view=markup


--
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/81918553.2791379.1438871317537.javamail.zim...@chu.ulg.ac.be

Re: gcc-5 transition of GDCM

[Sébastien Jodogne]

> Fine! With your updated version of gccxml, I was able to fix several build
> errors in the gdcm package. I had to apply one patch [1].
> 
> Unfortunately, there are still build errors related to mummy that read as
> follow: [...]

Pursuing my investigations, I have just noticed that the argument 
"--castxml-start=" seems to be non properly understood by castxml. 

To spot this problem, I have included inside the "/usr/bin/gccxml" wrapper 
around castxml, a "print castxml_cmd" command to display the command that is 
invoked when running CMake. Here is the output:

=
[  0%] Generating xml/vtkGDCMImageReader.xml
cd /tmp/build-area/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK && 
/usr/bin/gccxml -fxml-start=_cable_ -I/usr/include/mummy-1.0.3 
-I/usr/include/vtk-5.8 -DCABLE_CONFIGURATION 
/tmp/build-area/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/xml/vtkGDCMImageReader_gccxml.cxx
 -Wno-deprecated 
-fxml=/tmp/build-area/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/xml/vtkGDCMImageReader.xml
TRANSLATION from gccxml to castxml:
['/usr/bin/castxml', '--castxml-start', '_cable_', 
'-I/usr/include/mummy-1.0.3', '-I/usr/include/vtk-5.8', 
'-DCABLE_CONFIGURATION', 
'/tmp/build-area/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/xml/vtkGDCMImageReader_gccxml.cxx',
 '-Wno-deprecated', '--castxml-gccxml', '-o', 
'/tmp/build-area/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/xml/vtkGDCMImageReader.xml']
error: no such file or directory: '_cable_'
=

As you notice, the "-fxml-start=_cable_" argument is replaced by two 
"--castxml-start _cable_" arguments, which is the proper behavior according to 
the manual of castxml:

=
--castxml-start [,]...

  Start AST traversal at declaration(s) with the given (qualified) name(s).  
Multiple names may be specified as a comma-separated  list  or  by
  repeating the option.
=

The error "no such file or directory: '_cable_'" is however totally unexpected, 
as the expected argument to "--castxml-start" is a comma-separated list of 
qualified names, and not a file. This looks as a bug inside castxml. Steve, do 
you have any idea?

Sébastien-


--
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/1739653810.2714223.1438857532989.javamail.zim...@chu.ulg.ac.be

Re: gcc-5 transition of GDCM

[Sébastien Jodogne]

Hi,

> > > I will try hard to get the script into VCS tonight and then I will
> > > welcome
> > > help in testing the builds on dependent packages.
> > 
> > Great! Let me know when things are ready.
> 
> OK.  The gccxml package is updated with a wrapper around castxml (the real
> binary is renamed to gccxml.real).

Fine! With your updated version of gccxml, I was able to fix several build 
errors in the gdcm package. I had to apply one patch [1].

Unfortunately, there are still build errors related to mummy that read as 
follow:

=
[ 40%] Generating csharp/vtkGDCMThreadedImageReader-sentinel.txt
cd 
/home/jodogne/Subversion/Debian/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK
 && /usr/bin/mummy --suppress-warnings 6005 6006 6009 6010 6012 6013 6015 6016 
6017 6018 6019 --settings-file 
/home/jodogne/Subversion/Debian/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/xml/MummySettings.xml
 --gccxml-file 
/home/jodogne/Subversion/Debian/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/xml/vtkGDCMThreadedImageReader.xml
 --csharp-file 
/home/jodogne/Subversion/Debian/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/csharp/vtkGDCMThreadedImageReader.cs
 --export-layer-file 
/home/jodogne/Subversion/Debian/gdcm/build-area/gdcm-2.4.4/obj-x86_64-linux-gnu/Utilities/VTK/export-layer/vtkGDCMThreadedImageReaderEL.cxx
ERROR: In 
/tmp/buildd/cableswig-0.1.0+cvs20111009/Cable/Parsers/cableXMLSourceParser.cxx, 
line 584
XMLSourceParser (0x2563bf0): No name on Named object Destructor

ERROR: In 
/tmp/buildd/cableswig-0.1.0+cvs20111009/Cable/Parsers/cableXMLSourceParser.cxx, 
line 479
XMLSourceParser (0x2563bf0): Error adding element Destructor _2075

ERROR: In 
/tmp/buildd/cableswig-0.1.0+cvs20111009/Cable/Parsers/cableXMLSourceParser.cxx, 
line 138
XMLSourceParser (0x2563bf0): Error creating source representation.

# (0): error: m5008: MummyApplication::ProcessSource
Utilities/VTK/CMakeFiles/vtkgdcmGenerateWrappers.dir/build.make:118: recipe for 
target 'Utilities/VTK/csharp/vtkGDCMThreadedImageReader-sentinel.txt' failed
make[3]: *** [Utilities/VTK/csharp/vtkGDCMThreadedImageReader-sentinel.txt] 
Error 144
=

Any idea is welcome.

Cheers,
Sébastien-


[1] 
http://anonscm.debian.org/viewvc/debian-med/trunk/packages/gdcm/trunk/debian/patches/castxml.patch?view=markup


--
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/1945820479.2673947.1438851649445.javamail.zim...@chu.ulg.ac.be

Re: gcc-5 transition of GDCM

[Steve M. Robbins]

On August 5, 2015 03:29:30 PM Sébastien Jodogne wrote:

> > I will try hard to get the script into VCS tonight and then I will welcome
> > help in testing the builds on dependent packages.
> 
> Great! Let me know when things are ready.

OK.  The gccxml package is updated with a wrapper around castxml (the real 
binary is renamed to gccxml.real).  

Still working on testing build-deps with it.  Sadly, pygccxml fails :-(

-Steve


signature.asc
Description: This is a digitally signed message part.

Re: gcc-5 transition of GDCM

[Sébastien Jodogne]

> It was only incidentally related, in that GDCM build-depends on gccxml, which
> I'm trying to replace by a wrapper around castxml.  I was doing so because of
> things like the above-mentioned bugs.

OK, thanks for clarifying this :)

> I didn't try, but I was under the impression that gdcm does not build with
> gcc
> 5 -- because of gccxml issues.  But you say you rebuilt gdcm by hand?  Did
> you
> build the entire Debian package or just the C++ libraries?

No, the full package does not currently compile. I have only compiled the 
static and shared libraries from the upstream package, as they are sufficient 
for my use.

> > Steve, I have no experience with ITK, but is there some help I can provide
> > you on the "gccxml" and "gdcm" side of things?
> 
> I will try hard to get the script into VCS tonight and then I will welcome
> help in testing the builds on dependent packages.

Great! Let me know when things are ready.

Sébastien-


--
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/1063925370.2503241.1438781369994.javamail.zim...@chu.ulg.ac.be

Re: gcc-5 transition of GDCM

[Steve M. Robbins]

On August 4, 2015 07:19:14 PM Sébastien Jodogne wrote:
> > > As a consequence, I would like to know when updated versions of the
> > > libgdcm2.4 and libgdcm2-dev packages built with gcc-5 will be available
> > > in
> > > the repositories. Is this information available somewhere?
> > 
> > If you read the Debian Med mailing list you might have noticed Steve's
> > status update[1].  The short answer is:  If you will help fixing #777868
> > /  #791508 it might happen faster.
> 
> Sorry, I continuously monitor Debian Med mailing list, but I had not
> understood that Steve's post was also related to GDCM.

It was only incidentally related, in that GDCM build-depends on gccxml, which 
I'm trying to replace by a wrapper around castxml.  I was doing so because of 
things like the above-mentioned bugs.  

I didn't try, but I was under the impression that gdcm does not build with gcc 
5 -- because of gccxml issues.  But you say you rebuilt gdcm by hand?  Did you 
build the entire Debian package or just the C++ libraries?  

> Steve, I have no experience with ITK, but is there some help I can provide
> you on the "gccxml" and "gdcm" side of things?

I will try hard to get the script into VCS tonight and then I will welcome 
help in testing the builds on dependent packages. 

-Steve



signature.asc
Description: This is a digitally signed message part.

Re: gcc-5 transition of GDCM

[Sébastien Jodogne]

> > As a consequence, I would like to know when updated versions of the
> > libgdcm2.4 and libgdcm2-dev packages built with gcc-5 will be available in
> > the repositories. Is this information available somewhere?
> 
> If you read the Debian Med mailing list you might have noticed Steve's
> status update[1].  The short answer is:  If you will help fixing #777868
> /  #791508 it might happen faster.

Sorry, I continuously monitor Debian Med mailing list, but I had not understood 
that Steve's post was also related to GDCM.

Steve, I have no experience with ITK, but is there some help I can provide you 
on the "gccxml" and "gdcm" side of things?

Sébastien-


--
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/542104116.2317956.1438708754566.javamail.zim...@chu.ulg.ac.be

Re: gcc-5 transition of GDCM

[Andreas Tille]

Hi

On Tue, Aug 04, 2015 at 05:53:35PM +0200, Sébastien Jodogne wrote:
> 
> As a consequence, I would like to know when updated versions of the 
> libgdcm2.4 and libgdcm2-dev packages built with gcc-5 will be available in 
> the repositories. Is this information available somewhere?

If you read the Debian Med mailing list you might have noticed Steve's
status update[1].  The short answer is:  If you will help fixing #777868
/  #791508 it might happen faster.

Kind regards

Andreas.

[1] https://lists.debian.org/debian-med/2015/08/msg00010.html

-- 
http://fam-tille.de


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20150804164413.gb23...@an3as.eu

gcc-5 transition of GDCM

[Sébastien Jodogne]

Dear all,

I am currently packaging a new plugin to bring DICOMweb support to Orthanc [1]. 
When I build the plugin from upstream inside Debian sid, I get the following 
error:

=
Linking CXX shared library libOrthancDicomWeb.so
CMakeFiles/OrthancDicomWeb.dir/Plugin/Dicom.cpp.o: In function 
`OrthancPlugins::ConvertDicomStringToUf8(std::__cxx11::basic_string, std::allocator >&, gdcm::Dict const&, gdcm::File 
const*, gdcm::DataElement const&, Orthanc::Encoding)':
Dicom.cpp:(.text+0xfb8): undefined reference to 
`gdcm::StringFilter::ToString[abi:cxx11](gdcm::Tag const&) const'
collect2: error: ld returned 1 exit status
=

This linker error is related to the fact that the GDCM packages have not been 
recompiled with gcc-5 yet. As a consequence, there are ABI problems (hence the 
"[abi:cxx11]" tag). If I manually build GDCM from source, this error vanishes 
and the plugin works fine. But, of course, "dpkg-checkbuilddeps" and "pbuilder" 
complain about the missing packages, which is a showstopper for my packaging 
work.

As a consequence, I would like to know when updated versions of the libgdcm2.4 
and libgdcm2-dev packages built with gcc-5 will be available in the 
repositories. Is this information available somewhere?

TIA,
Sébastien-


[1] http://www.orthanc-server.com/static.php?page=dicomweb


--
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/2047918828.2294334.1438703615110.javamail.zim...@chu.ulg.ac.be