[Insight-developers] removing the dash in the module names

Bill Lorensen bill.lorensen at gmail.com
Fri Jun 24 08:03:48 EDT 2011 

I think it should not be done now because our meeting is starting on Monday.

I am also concerned that until the wrapping is compete and tested that
you may find other issues.

Bill

2011/6/24 Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr>:
>
> Bill,
>
> Would it be less disruptive in two weeks?
> I don't think so.
>
> Also please note that we are actually working on the beta.
> I'm not yet aware of any commit freeze or commit restriction - please
> correct me if I'm wrong.
>
> ----
>
> The change is here
>
>  http://review.source.kitware.com/#change,1973
>
> and the (green) experimental build is here
>
>  http://www.cdash.org/CDash/buildSummary.php?buildid=1264533
>
> I've catched a typo in a module name, so at least that won't have been
> completely useless, even if the change is rejected.
>
> Gaëtan
>
>
>
> Le 23 juin 11 à 20:56, Bill Lorensen a écrit :
>
>> I suggest the following:
>>
>> 1) Clone the itk repo.
>> 2) Apply your suggested changes to the module names (uses -'s to delimit)
>> 3) Do all of the wrapping changes you need to accommodate itk modular.
>> I assume there is some work to be done here and you may uncover other
>> issues.
>>
>> Once this is complete, propose a solution to the development team.
>>
>> We cannot afford to delay the beta. It has already slipped several months.
>>
>> Comments from the team?
>>
>> Bill
>>
>> 2011/6/23 Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr>:
>>>
>>> Le 23 juin 11 à 19:23, Bill Lorensen a écrit :
>>>
>>>> This really should have been discussed last March.
>>>
>>> I must agree on that point.
>>> Unfortunately things have not moved as expected and we are dealing with
>>> that
>>> issue now.
>>>
>>>> With all of the
>>>> other things we need to do before the beta this seems disruptive at
>>>> this time. Almost 1800 files will need changes and over 4800 lines.
>>>> Even if an automated script catches 90% of the changes, we can expect
>>>> some turmoil. All if the A2D2 projects will be affected.
>>>
>>> 90% seems quite underestimated.
>>> I won't guarantee that it won't cause any problem though, but I expect
>>> them
>>> to be quite limited.
>>>
>>> We could also do it in the same series of commits that Hans was planning
>>> to
>>> rename the .txx in .hxx.
>>> That would lower the annoyance of merging conflicts when most of the
>>> files
>>> are touched.
>>>
>>>>
>>>> Also, are there other issues that may come up regarding wrapping?
>>>
>>> Probably a few other things - mostly fixing broken code which hasn't been
>>> catched by our tests before.
>>>
>>>
>>> I proposed this change because it requires a lot less effort than dealing
>>> with non usable names for wrapping - in addition to avoid complicating
>>> the
>>> wrapping code.
>>> Unfortunately I've already spent way more time to discuss that issue than
>>> the initial edit to actually fix it - less than 5 minutes:
>>>
>>>  modules=`find * -name itk-module.cmake | xargs grep itk_module | cut
>>> -d'('
>>> -f2 | cut -d')' -f1 | cut -d" " -f1`
>>>  for m in $modules; do
>>>   M=${m//-/_}
>>>   echo $M
>>>   find * -type f | xargs perl -e "s/$m/$M/g" -pi
>>>  done
>>>
>>> Should be not too long to fix the remaining problems by hand
>>>
>>>  [glehmann at marvin ITK]$ git grep ITK- | wc -l
>>>  36
>>>  [glehmann at marvin ITK]$ git diff --stat | tail -1
>>>  1750 files changed, 4475 insertions(+), 4482 deletions(-)
>>>
>>> Gaëtan
>>>
>>>>
>>>> Bill
>>>>
>>>> 2011/6/23 Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr>:
>>>>>
>>>>> Hi Xiaxio,
>>>>> Hi David,
>>>>>
>>>>> Le 23 juin 11 à 17:35, David Cole a écrit :
>>>>>
>>>>>> Why couldn't there be a name-mapping (reversible, 2-way) in just the
>>>>>> Wrapping libraries? Are the "real" module names referenced directly in
>>>>>> many
>>>>>> places other than target_link_libraries calls?
>>>>>
>>>>> They are reused in many places in wrapping.
>>>>> Moving to the new modules defined in ITK is not that easy
>>>>> unfortunately,
>>>>> and
>>>>> dealing with non compatible module names make it even harder.
>>>>>
>>>>>>
>>>>>> How many languages can't handle the "-" in names? Is it just python,
>>>>>> or
>>>>>> are dashes also bad in some other language?
>>>>>>
>>>>>
>>>>> At least python and java.
>>>>> I haven't checked for Tcl.
>>>>>
>>>>>>
>>>>>> On Thu, Jun 23, 2011 at 11:15 AM, Xiaoxiao Liu
>>>>>> <xiaoxiao.liu at kitware.com>
>>>>>> wrote:
>>>>>> Yup, not only most of the cmake files need some string replacements,
>>>>>> all
>>>>>> the header files (for doxygen) need to be touched as well.
>>>>>
>>>>> Yes - nothing some regular expressions can't do.
>>>>>
>>>>>> It might affect all other under-development external modules (e.g.
>>>>>> Lesion
>>>>>> sizing toolkit, maybe a couple of other out there).
>>>>>>
>>>>>
>>>>> We are doing many changes which are affecting the development of other
>>>>> modules.
>>>>> I think it shouldn't affect a design decision at that point.
>>>>> In that case, using underscores instead of dashes doesn't make the
>>>>> names
>>>>> less good, but it does make the wrapping easier.
>>>>>
>>>>>> If there is no easy work-around, I prefer Bra'd's  "underscore"
>>>>>> naming.
>>>>>
>>>>> Ok!
>>>>>
>>>>> Gaëtan
>>>>>
>>>>>
>>>>>
>>>>>>
>>>>>> 2011/6/23 Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr>
>>>>>>
>>>>>> Le 23 juin 11 à 15:19, Bill Lorensen a écrit :
>>>>>>
>>>>>>
>>>>>> Virtually every .h file in the toolkit will need editing... Is there
>>>>>> some other, perhaps wrapping-specific solution?
>>>>>>
>>>>>> That would force us to do a lot of conversions with cmake.
>>>>>> In addition to make the code more complicated, I'm concerned that it
>>>>>> may
>>>>>> make the already quite slow configuration even slower.
>>>>>>
>>>>>> Is it really a big problem to touch all the headers?
>>>>>>
>>>>>> Gaëtan
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Thu, Jun 23, 2011 at 8:53 AM, Brad King <brad.king at kitware.com>
>>>>>> wrote:
>>>>>> On 06/23/2011 04:07 AM, Gaëtan Lehmann wrote:
>>>>>> most of the languages won't support a dash in the module name.
>>>>>>
>>>>>> Would it be ok to remove those dash in the module names?
>>>>>>
>>>>>>  ITK-Common
>>>>>>  ITK-IO-Base
>>>>>>
>>>>>> would become
>>>>>>
>>>>>>  ITKCommon
>>>>>>  ITKIOBase
>>>>>>
>>>>>> Perhaps underscores?
>>>>>>
>>>>>>  ITK_Common
>>>>>>  ITK_IO_Base
>>>>>>
>>>>>> IMO having a separator breaks with our C++ coding guideline naming
>>>>>> style enough to clearly distinguish the names as modules.  It also
>>>>>> makes the name components clear, especially when the module names
>>>>>> are all uppercase.  Consider
>>>>>>
>>>>>>  ITK-IO-PNG   =>  ITKIOPNG  v. ITK_IO_PNG
>>>>>>  ITK-IO-TIFF  =>  ITKIOTIFF v. ITK_IO_TIFF
>>>>>>
>>>>>> -Brad
>>>>>> _______________________________________________
>>>>>> 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
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Gaëtan Lehmann
>>>>>> Biologie du Développement et de la Reproduction
>>>>>> INRA de Jouy-en-Josas (France)
>>>>>> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
>>>>>> http://mima2.jouy.inra.fr  http://www.itk.org
>>>>>> http://www.bepo.fr
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> 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
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>>
>>>>>>
>>>>>> ---------------------------------------------
>>>>>> Xiaoxiao Liu, Ph.D.
>>>>>> R & D Engineer
>>>>>> Kitware Inc.
>>>>>> Clifton Park, NY
>>>>>> Phone: (518) 881-4924  or  (518) 371-3971 x124
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> 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
>>>>>>
>>>>>>
>>>>>
>>>>> --
>>>>> Gaëtan Lehmann
>>>>> Biologie du Développement et de la Reproduction
>>>>> INRA de Jouy-en-Josas (France)
>>>>> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
>>>>> http://mima2.jouy.inra.fr  http://www.itk.org
>>>>> http://www.bepo.fr
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> 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
>>>>>
>>>>>
>>>
>>> --
>>> Gaëtan Lehmann
>>> Biologie du Développement et de la Reproduction
>>> INRA de Jouy-en-Josas (France)
>>> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
>>> http://mima2.jouy.inra.fr  http://www.itk.org
>>> http://www.bepo.fr
>>>
>>>
>
> --
> Gaëtan Lehmann
> Biologie du Développement et de la Reproduction
> INRA de Jouy-en-Josas (France)
> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
> http://mima2.jouy.inra.fr  http://www.itk.org
> http://www.bepo.fr
>
>

More information about the Insight-developers mailing list