| <<<Back 1 day (to 2018/05/20) | 20180521 |
inflex | If I wanted to add a menu of sorts to muPDF, keeping with FreeGLUT portability, what's the preferred method? More FreeGLUT? | 07:30.39 |
tor8 | df (for the logs): build from source, and complain to your distribution package maintainer -- they're probably the ones who decided to *not* use our enhanced freeglut which supports copy & paste. | 09:03.22 |
| inflex: have you looked at the tor/ui branch? | 09:10.44 |
| inflex: git remote add tor git://git.ghostscript.com/user/tor/mupdf.git && git remote update tor | 09:11.32 |
moolc | tor8: seen my message (here on the channel) regarding c11 warning splat from fitz? | 09:17.53 |
inflex | Thanks tor8 . | 09:18.36 |
tor8 | inflex: if you use that ui framework, there's support for popup menus in there | 09:22.00 |
| moolc: the one about _MSC_VER? | 09:22.42 |
moolc | tor8: aye | 09:23.10 |
inflex | I don't have a strong need for much in the way of UI, was thinking of using a menu or status bar for information updates, current context-menu/popups might suffice for the user's needs. Early days of testing I suppose. | 09:24.45 |
tor8 | inflex: I'm not sure how well freeglut support the GLUT menu api. | 09:25.29 |
| moolc: I have a possible fix on tor/master (in two parts) if you want to have a look | 09:25.38 |
| moolc: just how did you manage to trigger it? | 09:25.49 |
inflex | tor8, well, might be something for me to work on too :D | 09:25.50 |
tor8 | inflex: the ui branch should get merged to master in a month or two | 09:26.20 |
moolc | tor8: superior build system makes that easy | 09:28.42 |
inflex | How long has mupdf-gl been in development for? | 09:29.01 |
moolc | tor8: works like a charm | 09:31.43 |
| tor8: (your fix that is) | 09:31.53 |
tor8 | inflex: the mupdf-gl viewer, I think work on that viewer started in 2015 or so | 09:37.38 |
| the code for the mupdf-x11/-w32 viewer is a bit ... complicated | 09:37.54 |
inflex | ja, saw it was a bit more so. I like the -gl version. | 09:38.20 |
moolc | tor8: and i guess i misunderstood your question earlier, it's trivially triggered by doing 'clang -Wundef -std=c11 somethingthatincludesfitz.c' | 09:41.29 |
| tor8: or even simpler: gcc -x c -Wundef -I ~/x/rcs/git/mupdf/include -include mupdf/fitz.h /dev/null | 09:44.07 |
tor8 | moolc: how do you get around sigsetjmp not being defined with -std=c11? that's another build error I get there. | 09:57.07 |
moolc | tor8: - gcc -c -Werror -x c -Wno-undef -I ~/x/rcs/git/mupdf/include -include mupdf/fitz.h /dev/null && echo success | 09:58.36 |
| success | 09:58.36 |
| | 09:58.36 |
| iow i don't do anything special | 09:58.42 |
tor8 | but if you use -std=c11 when building an actual source file (and not just checking that the header goes through) you should find it blows up whenever fz_try is used | 10:01.16 |
inflex | tor8, okay, fairly young project. | 10:05.58 |
moolc | tor8: well, it doesn't | 10:06.06 |
inflex | (esp when you consider how far Ghostscript goes back) | 10:06.14 |
moolc | i've built llpp with both gcc and clang + -std=c11 | 10:06.17 |
tor8 | inflex: mupdf itself is *much* older ... got started back in 2003 or 2004 | 10:06.47 |
moolc | -std=c11 + -Weverything on clang (some stratgeic diagnostics are disabled in the source though) | 10:07.09 |
tor8 | moolc: include/mupdf/fitz/system.h:79:20: error: unknown type name 'sigjmp_buf' | 10:11.29 |
| #define fz_jmp_buf sigjmp_buf | 10:11.29 |
moolc | tor8: if you are still on debian all bets are off | 10:14.18 |
tor8 | moolc: I guess so. I just need to guard the use of sigsetjmp with ifndef _STRICT_ANSI_ | 10:14.42 |
| or do you set _POSIX_C_SOURCE? | 10:15.01 |
| or _XOPEN_SOURCE? | 10:15.07 |
moolc | _GNU_SOURCE | 10:15.08 |
| not in this example obviously | 10:15.15 |
| but i (sadly) needed it in the real deal | 10:15.22 |
| and _GNU_SOURCE implies a lot of stuff | 10:15.35 |
tor8 | right. that would probably explain why you can build with -std=c11 ... defining _*_SOURCE turns 'off' a lot of the strictness | 10:15.55 |
moolc | i've removed it.. let me try enabling c11 with gcc | 10:16.46 |
| yep works. without _GNU_SOURCES after adding extern char **environ; at the top | 10:17.29 |
| so there | 10:17.31 |
tor8 | moolc: tor/master builds, with a handful of warnings in freeglut (only to be expected I guess) | 10:20.13 |
moolc | tor8: i only do make libs anyway ;) | 10:23.44 |
| mupdf-[x11|gl] are to baroque for my tastes | 10:23.58 |
| s;to;too | 10:24.02 |
| not to mention being slower and lacking features.. so libs for me | 10:25.53 |
pihug12 | Hello tor8 | 15:20.38 |
| I asked F-Droid to update the MuPDF builds | 15:20.54 |
| https://gitlab.com/fdroid/fdroiddata/issues/1211 | 15:21.01 |
| And they can't because the version code (20) is the same as 1.12.0 | 15:21.23 |
| http://git.ghostscript.com/?p=mupdf-android-viewer.git;a=blob;f=app/build.gradle;hb=HEAD | 15:21.45 |
| http://git.ghostscript.com/?p=mupdf-android-viewer-mini.git;a=blob;f=app/build.gradle;hb=HEAD | 15:21.54 |
| Can you, please, increase it and put the tag on this new commit? | 15:22.37 |
tor8 | pihug12: I've bumped the versionCode | 15:25.52 |
pihug12 | Perfect! Thanks :) | 15:28.27 |
| Forward 1 day (to 2018/05/22)>>> | |