djvulibre: don't mangle locales #1565
Conversation
It's generally a bad idea, because this rusn in our main thread, and locales in POSIX are a mess. May help w/ koreader/koreader#9918
|
Picked a few upstream fixes while I was in there ;o). |
| # define DJVU_LOCALE do { \ | ||
| - setlocale(LC_ALL,""); \ | ||
| - setlocale(LC_NUMERIC,"C"); \ | ||
| djvu_programname(argv[0]); } while (0) | ||
| #endif |
There was a problem hiding this comment.
This one does seem a bit curious and in need of patching.
There was a problem hiding this comment.
It should only affect tools, though, which we probably don't use ;o).
The culprit is likely the context creation one, I would imagine.
(Again, I did a very wide sweep just to be safe).
There was a problem hiding this comment.
I didn't look at the context in detail, common.h does sound quite, ahem, common though xp
| -#ifndef DO_CHANGELOCALE | ||
| -# define DO_CHANGELOCALE 0 |
There was a problem hiding this comment.
Might be better to just -DDO_CHANGELOCALE=0 or whatever instead of doing a patch that needs to be maintained? (Although this particular dependency is fairly static.)
There was a problem hiding this comment.
Yeah, I went with the nuclear option just to be safe (and because I already had a patch to generate anyway).
It's generally a bad idea, because this runs in our main thread, and locales in POSIX are a mess.
May help w/ koreader/koreader#9918
This change is