SG++ 2.1.0 fails to build on linux machine about sgpp HOT 10 CLOSED

In GitLab by @valentjn on Mar 28, 2017, 13:03

Most likely he's having some umlauts in his environment variables. (Congrats on being the first user with this :first_place:) He can check with export | grep -P "[\x80-\xFF]", which displays all variables with non-ASCII characters in it.

I can reproduce this if I do export BLA="täst" before the scons.

In the original snippet, the SCons people seem to convert unicode strings to standard str strings. I seem to have replaced that with that ASCII re-encoding, probably because it didn't work on my system. (See, that's how I work: Just do random stuff until it's fixed.) I just checked and all the variables of spawnEnv are already strs before doing anything, at least for my Ubuntu.

Maybe someone can confirm that (print type(spawnEnv[var]) before converting)? We might not need the conversion for loop at all in this case. It seems to work just fine on my system without it.

Florian might try removing that for loop; if it works, I'll check on Windows and if that works, I'll push a fix, probably to codingdays.

from sgpp.

In GitLab by @krsgpp on Mar 31, 2017, 12:37

Florian told me, that he is able to compile when there are no env-vars with umlauts.

from sgpp.

In GitLab by @valentjn on Mar 31, 2017, 14:06

Excellent! But since we want that SG++ runs even on platforms with umlauts in their environment (and we want to solve bugs here and not just post workarounds), it'd be nice if someone can check if we need that conversion for loop at all.

from sgpp.

In GitLab by @valentjn on Aug 29, 2017, 15:16

This bug seems to be current in master. @krsgpp had problems pushing a MR since Jenkins's GitLab plugin injected his name in the environment.

from sgpp.

In GitLab by @valentjn on Aug 31, 2017, 07:49

mentioned in merge request !25

from sgpp.

In GitLab by @krsgpp on Sep 6, 2017, 12:12

Ok, I tried around and your proposed solution (removing the for loop that does the conversion) works at my system.
I suggest that I create a separate MR for this one and push it, so that the ENV is again "bad" and we can see how jenkins handles this. Ok?

from sgpp.

In GitLab by @krsgpp on Sep 6, 2017, 12:18

mentioned in merge request !26

from sgpp.

In GitLab by @krsgpp on Sep 6, 2017, 12:22

closed via commit 45b122e

from sgpp.

In GitLab by @krsgpp on Sep 6, 2017, 12:22

closed via merge request !26

from sgpp.

In GitLab by @krsgpp on Sep 15, 2017, 15:57

mentioned in commit 45b122e

from sgpp.