[Insight-developers] GDCM empty cxx files

[Alexandre GOUAILLARD]

mark,

as indicated many times,
our duty goes to ITK and not to gdcm proper.
The priority is to fix the warning on the itk dashboard first.
Brad king mentionned you could modify in ITK and backport to gdcm later.
I explained to you how you could also have a separate git rep. to be
the buffer between gdcm and itk if need be.

now, our duty goes to ITK. The priority is to make the ITK dashboard green

alex.


On Sat, Mar 26, 2011 at 6:14 AM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
> Whatever is the best approach to fix the gdcm warnings. They are the major
> source of warnings are really getting annoying. They have been present since
> last summer.  I will continue to fix gdcm warnings until I see them dropping
> through another means.
>
> On Fri, Mar 25, 2011 at 6:10 PM, Mark Roden <mmroden at gmail.com> wrote:
>>
>> Ah, thanks for that.
>>
>> Since it took so long for my last gerrit patch to get a review, I've
>> switched to fixing tests and warnings in the gdcm main branch.  My
>> hope was to be able to drop that chunk of code into ITK, with the
>> warnings fixed there.  Many of them have already been fixed in ITK,
>> but moving over fixes from ITK to gdcm main produced some failing gdcm
>> tests.  These tests don't exist in ITK (hence removing the GDCM build
>> option from the ITK cmake options), and the ITK gdcm tests appear to
>> be building and running just fine (with the exception of IO test 5,
>> and that looks to be the same broken test it's always been, and should
>> be removed from the test suite).
>>
>> Is this approach a valid one?  I mean, it's taken around 2 weeks for
>> someone to review an ITK gerrit patch from me, but in the meantime,
>> I've been able to make much faster progress in the gdcm main branch
>> with this approach.  And, I think, a safer one, since there is more
>> complete test coverage.
>>
>> Mark
>>
>> On Fri, Mar 25, 2011 at 3:05 PM, Bill Lorensen <bill.lorensen at gmail.com>
>> wrote:
>> > I have the fix in my gerrit patch.
>> >
>> > On Fri, Mar 25, 2011 at 3:33 PM, Mark Roden <mmroden at gmail.com> wrote:
>> >>
>> >> I thought I submitted a patch for this as well, but it probably didn't
>> >> make it over the ITK modularization boundary.
>> >>
>> >> The last patch I submitted was done after the ITK changeover, but I
>> >> guess the changeover was ongoing when I submitted the patch, and so
>> >> now I have to redo it.  That's the
>> >> getting-rid-of-superfluous-cmake-options patch.
>> >>
>> >> On Fri, Mar 25, 2011 at 12:21 PM, Ryan, William J.
>> >> <ryan.william at mayo.edu> wrote:
>> >> > I think Mark Roden had a patch submitted that may fix this, but I'm
>> >> > not
>> >> > 100%
>> >> > sure about that.  Let's see what he has to say.
>> >> >
>> >> >
>> >> >
>> >> >
>> >> >
>> >> >   Bill Ryan
>> >> >
>> >> >
>> >> >
>> >> > From: insight-developers-bounces at itk.org
>> >> > [mailto:insight-developers-bounces at itk.org] On Behalf Of Bill
>> >> > Lorensen
>> >> > Sent: Friday, March 25, 2011 1:23 PM
>> >> > To: Mathieu Malaterre
>> >> > Cc: Insight Developers
>> >> > Subject: Re: [Insight-developers] GDCM empty cxx files
>> >> >
>> >> >
>> >> >
>> >> > Mathieu,
>> >> >
>> >> > We still support VS7.1. I would like to submit a patch that removes
>> >> > these
>> >> > files from the lib and see if it works.
>> >> >
>> >> > Unfortunately, the VS generator does not work with the modularized
>> >> > ITK.
>> >> > We
>> >> > may not be able to thoroughly test it until the patched cmake is
>> >> > released.
>> >> >
>> >> > Thanks,
>> >> >
>> >> > Bill
>> >> >
>> >> > On Fri, Mar 25, 2011 at 2:13 PM, Mathieu Malaterre
>> >> > <mathieu.malaterre at gmail.com> wrote:
>> >> >
>> >> > From the top of my head, this was V7.1 with /MT as compilation flag
>> >> > (instead of the default /MD used by Cmake). Can't remember if this
>> >> > was
>> >> > using shared or static lib.
>> >> >
>> >> > Does anyone still have this compiler ?
>> >> >
>> >> > HTH
>> >> >
>> >> > On Fri, Mar 25, 2011 at 7:06 PM, Bill Lorensen
>> >> > <bill.lorensen at gmail.com>
>> >> > wrote:
>> >> >> Mathieu,
>> >> >>
>> >> >> Several files including gdcmObject.cxx have no code in them. Your
>> >> >> comment
>> >> >> in
>> >> >> gdcmObject.cxx says:
>> >> >>   // Don't ask why, but this is EXTREMELY important on Win32
>> >> >>   // Apparently the compiler is doing something special the first
>> >> >> time
>> >> >> it
>> >> >> compiles
>> >> >>   // this instanciation unit
>> >> >>   // If this fake file is not present I get an unresolved symbol for
>> >> >> each
>> >> >> function
>> >> >>   // of the gdcm::Object class
>> >> >>
>> >> >> When linking these "empty" files, VS10 reports several warnings:
>> >> >>  warning LNK4221: This object file does not define any previously
>> >> >> undefined
>> >> >> public symbols, so it will not be used by any link operation that
>> >> >> consumes
>> >> >> this library
>> >> >>
>> >> >> I have removed this file from the lib as well as
>> >> >> gdcmProgressEvent.cxx, gdcmString.cxx, gdcmException.cxx
>> >> >> gdcmDeflateStream.cxx and gdcmByteSwap.cxx
>> >> >>
>> >> >> When I build the modified lib, I do not get the warnoings, nor do I
>> >> >> get
>> >> >> any
>> >> >> errors.
>> >> >>
>> >> >> What compiler was giving the errors you mention in the comments?
>> >> >>
>> >> >> Bill
>> >> >>
>> >> >
>> >> >
>> >> > --
>> >> > Mathieu
>> >> >
>> >> >
>> >
>> >
>
>
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://kitware.com/products/protraining.html
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-developers
>
>