[Gmsh] huge files in version control

[Nicolas Marsic]

Dear all,

I'm the one to blame for this huge commit.
I committed a working directory, and not the clean intended one.
A new revision without those meshes is now available.
However, my mistake will unfortunately live forever in the svn tree.

Sorry for the inconvenience,
Nicolas.


On 04/11/15 00:16, Nico Schlömer wrote:
> Hi everyone,
>
> I noticed that today, some HUGE files were committed to the SVN repo,
> for example
> ```
> ./projects/small_fem/tests/ddm/hp-convergence/waveguide_X0.50_Y0.25_Z0.25_structured/guide3d_64.msh
> ./projects/small_fem/tests/ddm/hp-convergence/waveguide_X0.50_Y0.25_Z0.25_structured/unstr/guide3d.msh
> ```
> This is basically the end of gmsh's Git mirror [1] since Git preserves
> history for every clone, so everyone who ever checks out gmsh through
> git will have to fetch 400MB+ for today alone.
>
> Generally, committing data into version control repos isn't a good idea
> but there are several ways around it. Since you're using CMake, it might
> be worth checking out CMake's ExternalData feature [1]. I've been using
> it for some years now and am really happy with it.
>
> Cheers,
> Nico
>
>
> [1] https://cmake.org/cmake/help/v3.0/module/ExternalData.html
>
>
> [1] https://github.com/live-clones/gmsh
>
>
> _______________________________________________
> gmsh mailing list
> gmsh at geuz.org
> http://www.geuz.org/mailman/listinfo/gmsh
>