Bug #108378 for Math-GSL: Math-GSL-0.35 fails to build against GSL-2.0

Mon Nov 02 04:09:21 2015 sisyphus [...] cpan.org - Ticket created

Subject:

Math-GSL-0.35 fails to build against GSL-2.0

Attempting to build Math-GSL-0.35 against GSL-2.0 results in the following error: Process XS files version 1.16 (GSL version 2.0) xs\BSpline_wrap.1.16.c: In function '_wrap_gsl_bspline_deriv_eval': xs\BSpline_wrap.1.16.c:3282:19: error: too many arguments to function 'gsl_bspline_deriv_eval' result = (int)gsl_bspline_deriv_eval(arg1,arg2,arg3,arg4,arg5); Will swig automatically handle the various API changes that gsl-2.0 presents ? Cheers, Rob

Mon Nov 02 06:34:48 2015 ambs [...] perl-hackers.net - Correspondence added

Subject:	Re: [rt.cpan.org #108378] Math-GSL-0.35 fails to build against GSL-2.0
Date:	Mon, 02 Nov 2015 11:34:55 +0000
To:	bug-Math-Gsl [...] rt.cpan.org
From:	Alberto Simoes <ambs [...] perl-hackers.net>

Ah, great. More work :-) Thanks for pointing GSL 2.0 release. I hope we can make the module work with 1.x and 2.x, but I did not look into it yet, so, not sure. Also, not sure if I will be able to look it as fast as I would like. Best, Alberto On 02/11/2015 09:09, Sisyphus via RT wrote: Show quoted text

> Mon Nov 02 04:09:21 2015: Request 108378 was acted upon. > Transaction: Ticket created by SISYPHUS > Queue: Math-Gsl > Subject: Math-GSL-0.35 fails to build against GSL-2.0 > Broken in: 0.35 > Severity: Normal > Owner: Nobody > Requestors: sisyphus@cpan.org > Status: new > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=108378 > > > > Attempting to build Math-GSL-0.35 against GSL-2.0 results in the following error: > > Process XS files version 1.16 (GSL version 2.0) > xs\BSpline_wrap.1.16.c: In function '_wrap_gsl_bspline_deriv_eval': > xs\BSpline_wrap.1.16.c:3282:19: error: too many arguments to function 'gsl_bspline_deriv_eval' > result = (int)gsl_bspline_deriv_eval(arg1,arg2,arg3,arg4,arg5); > > Will swig automatically handle the various API changes that gsl-2.0 presents ? > > Cheers, > Rob >

Mon Nov 02 06:34:48 2015 The RT System itself - Status changed from 'new' to 'open'

Mon Nov 02 07:37:06 2015 sisyphus1 [...] optusnet.com.au - Correspondence added

Subject:	Re: [rt.cpan.org #108378] Math-GSL-0.35 fails to build against GSL-2.0
Date:	Mon, 2 Nov 2015 23:36:24 +1100
To:	<bug-Math-Gsl [...] rt.cpan.org>
From:	<sisyphus1 [...] optusnet.com.au>

Show quoted text

-----Original Message----- From: Alberto Simoes via RT

> <URL: https://rt.cpan.org/Ticket/Display.html?id=108378 > > > Ah, great. More work :-) > > Thanks for pointing GSL 2.0 release. > I hope we can make the module work with 1.x and 2.x, but I did not look > into it yet, so, not sure. > > Also, not sure if I will be able to look it as fast as I would like.

No great rush from my POV. I'm not sure how swig will go about accommodating the various functions whose prototypes have changed from 1.16 to 2.0 - or detecting that the gsl-2.0 libraries don't resolve the deprecated functions. I guess another option is to leave Math::GSL as the module that one uses with gsl-1.x, and create a separate module named (eg) Math::GSL2 for use with versions 2.x of gsl. But you'll probably have a better idea of how to deal with this than I do. (Well .... hopefully, anyway :-) Cheers, Rob

Mon Nov 02 07:39:38 2015 ambs [...] perl-hackers.net - Correspondence added

Subject:	Re: [rt.cpan.org #108378] Math-GSL-0.35 fails to build against GSL-2.0
Date:	Mon, 02 Nov 2015 12:39:45 +0000
To:	bug-Math-Gsl [...] rt.cpan.org
From:	Alberto Simoes <ambs [...] perl-hackers.net>

On 02/11/2015 12:37, sisyphus1@optusnet.com.au via RT wrote: Show quoted text

> Queue: Math-Gsl > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=108378 > > > -----Original Message----- > From: Alberto Simoes via RT

>> <URL: https://rt.cpan.org/Ticket/Display.html?id=108378 > >> >> Ah, great. More work :-) >> >> Thanks for pointing GSL 2.0 release. >> I hope we can make the module work with 1.x and 2.x, but I did not look >> into it yet, so, not sure. >> >> Also, not sure if I will be able to look it as fast as I would like.

> No great rush from my POV. > I'm not sure how swig will go about accommodating the various functions > whose prototypes have changed from 1.16 to 2.0 - or detecting that the > gsl-2.0 libraries don't resolve the deprecated functions. > > I guess another option is to leave Math::GSL as the module that one uses > with gsl-1.x, and create a separate module named (eg) Math::GSL2 for use > with versions 2.x of gsl. > > But you'll probably have a better idea of how to deal with this than I do. > (Well .... hopefully, anyway :-) >

For another module, where the author of the C library changes the API in minor version changes, I am shipping two swig files, and use one or another, when needed. But that is a much smaller api than gsl. Maybe gsl2 isn't that bad idea. But would love to give it a try first. Leto? thoughts?

Fri Sep 30 16:58:37 2016 jonathan [...] leto.net - Correspondence added

GSL 2.x is now supported. Note that GSL 1.15/1.16 are currently broke in Math::GSL 0.38 but I am hoping to fix that soon. We have to do some special things for GSL API changes but I have it mostly figured out. Turns out their Bspline API change is the most annoying to deal with: https://github.com/leto/math--gsl/issues/97 Duke On Mon Nov 02 04:09:21 2015, SISYPHUS wrote: Show quoted text

> Attempting to build Math-GSL-0.35 against GSL-2.0 results in the > following error: > > Process XS files version 1.16 (GSL version 2.0) > xs\BSpline_wrap.1.16.c: In function '_wrap_gsl_bspline_deriv_eval': > xs\BSpline_wrap.1.16.c:3282:19: error: too many arguments to function > 'gsl_bspline_deriv_eval' > result = (int)gsl_bspline_deriv_eval(arg1,arg2,arg3,arg4,arg5); > > Will swig automatically handle the various API changes that gsl-2.0 > presents ? > > Cheers, > Rob

Fri Sep 30 16:58:43 2016 jonathan [...] leto.net - Status changed from 'open' to 'resolved'

Fri Sep 30 16:58:43 2016 jonathan [...] leto.net - Fixed in 0.38 added