[with patch, positive review] Weird printing issues with gcc, caused by LANG environment variable

[craigcitro]

So I've noticed for a little while now that I have the following strange problem:

sage/rings/number_field/totallyreal_data.c: In function â:
sage/rings/number_field/totallyreal_data.c:1920: error: â undeclared (first use in this function)

But then, if you comment out the fix for trac ticket #276 (which sets the environment variable LANG), that same error becomes:

sage/rings/number_field/totallyreal_data.c: In function '__pyx_f_4sage_5rings_12number_field_16totallyreal_data_lagrange_degree_3_c':
sage/rings/number_field/totallyreal_data.c:1920: error: 'intp' undeclared (first use in this function)

Obviously we'd prefer the second. What I think we should do is move setting the LANG environment variable somewhere closer in the build & run process to where Maxima gets initialized. I haven't had time to sit down and look at this at all -- I'm assigning it to myself, but if someone else wants to jump in and fix it before me, I definitely won't mind.

I'm also adding robertwb to the cc, because I think I recall him saying he ran into this, too, and he's (1) probably happy to have a workaround, and (2) likely to come up with a fix himself, saving me work. :)

[craigcitro]

Typo in the title.

[craigcitro]

So I just happened to glance back at this ticket tonight because I hit this same bug; looking at #276, that was a fix for clisp -- which is now gone! So I'm attaching a patch that just un-does the work of the patch at #276, which will fix this weird printing issue.