2.3

[djsutherland]

Patch is modified to only include the cmake-relevant bits from https://github.com/ampl/gsl/tree/v2.3.0. :crossed_fingers: that this patch works for this and future versions as well; it changes almost no code compared to the previous patch.

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[djsutherland]

So for the record, I'm slightly worried about this huge change to the Windows patch even if it passes, because there's only existence tests on Windows – Unix runs make check but we don't run the cmake-defined tests on Windows. Should add that.

[jakirkham]

So the patch for CMake was added when there was no CMakeLists.txt before. However upstream now has one and does appear to do some special handling for MSVC in there. Maybe it is better to just use what they have?

EDIT: NVM appears that is what you are doing.

[djsutherland]

ampl/gsl is the Windows fork that the patch is based on, not upstream. The patch I added here is just the diff between that and the upstream download (or at least, only the relevant bits – skipped appveyor config, NuGet files, etc).

[jakirkham]

Yep, sorry, realized my mistake and edited my comment above. Sorry for the noise.

[ocefpaf]

@conda-forge-admin, please rerender

[djsutherland]

Blegh:

LINK: command "C:\PROGRA~2\MICROS~1.0\VC\bin\link.exe /nologo @CMakeFiles\gsl.dir\objects1.rsp /out:bin\gsl.dll /implib:gsl.lib /pdb:C:\bld\gsl_1520170835403\build_dir\bin\gsl.pdb /dll /version:0.0 /machine:X86 /INCREMENTAL:NO /DEF:C:\bld\gsl_1520170835403\work\gsl-2.3\gsl.def gslcblas.lib kernel32.lib user32.lib gdi32.lib winspool.lib shell32.lib ole32.lib oleaut32.lib uuid.lib comdlg32.lib advapi32.lib /MANIFEST /MANIFESTFILE:bin\gsl.dll.manifest" failed (exit code 1120) with the following output:
2627gsl.def : error LNK2001: unresolved external symbol gsl_multilarge_nlinear_df
2628gsl.def : error LNK2001: unresolved external symbol gsl_multilarge_nlinear_fdfvv
2629gsl.lib : fatal error LNK1120: 2 unresolved externals

Don't know what's causing this.

[jakirkham]

Yep, as you have already noted elsewhere, looks like issue ( https://github.com/ampl/gsl/issues/10 ) that @jeromekelleher raised, which should have been fixed with PR ( https://github.com/ampl/gsl/pull/14 ).

[djsutherland]

Closing in favor of #24, unless anyone strongly feels the need for a 2.3 release.