| <<<Back 1 day (to 2018/05/08) | 20180509 |
tor8 | Robin_Watts: sebras: some commits on tor/master could do with a review. | 13:14.58 |
Robin_Watts | looking | 13:22.10 |
sebras | tor8: I don't immediately understand "Fix 699319: Apply clip path after drawing, not before." enough to have an opinion, but it looks like it would compile of course. | 13:23.23 |
tor8 | sebras: pdfref17 page 234, last paragraph on the page. | 13:26.27 |
sebras | tor8: in "Fix java build." it looks to me like you accidentally rebased in the removal of Text_cloneNative from another branch, the fucntion is still there on origin/master... | 13:26.53 |
Robin_Watts | tor: All look good to me, with some minor qualms about the unused function purge. | 13:27.20 |
tor8 | sebras: oops. yes, you're right. that hunk should be in the 'unused function purge' commit. | 13:27.38 |
| Robin_Watts: avoid_white_fix_flag one is the only one that gave me some real pause | 13:28.00 |
Robin_Watts | Removing the gz_gen_id stuff scares me a bit in that it makes me wonder if we should be using that somewhere :) | 13:28.00 |
tor8 | the gen_id stuff ... was never used. you added it but I couldn't find where you actually used it in the same commit. | 13:28.28 |
sebras | is it good to disable optinizations in the windows build of libresource.vsproj? | 13:28.33 |
| or rather, why is it? | 13:28.43 |
Robin_Watts | sebras: optimizations gain us nothing. | 13:29.33 |
tor8 | and we've made further changes in how ids are handled in tiling devices after that point too, which I think removes the need/motivation for gen_id | 13:29.33 |
Robin_Watts | and if enabled, it makes the compiler search for stuff that isn't there that takes lots of memory etc. | 13:30.10 |
tor8 | sebras: asking the compiler to do optimization passes on big chunks of static data is (a) pointless, and (b) eats up memory and time | 13:30.11 |
| I got one report about VS2017 running out of memory compiling the big CJK font file | 13:30.42 |
sebras | tor8: oh, right, libresources are the fonts and cmaps and whatnot... | 13:30.46 |
tor8 | sebras: yes. the big fat font files. | 13:30.53 |
| I expect VS2017 to be more wasteful of RAM than VS2005... but not to the point where it can't even build :( | 13:31.21 |
sebras | tor continuously updating CHANGES seems like a good idea. :) | 13:31.31 |
Robin_Watts | seems like an idea that we'll never keep up, but... | 13:32.11 |
tor8 | I figured I could do it while I remember :) | 13:32.31 |
| but don't count on me keeping up | 13:32.39 |
sebras | Robin_Watts: no, but doing any update of CHANGES befor the release is better than not doing it. :) | 13:32.49 |
Robin_Watts | yeah. | 13:32.58 |
sebras | tor8: the new changes to the mujs interface.. aare any of those supposed to be in the java code too? | 13:33.25 |
tor8 | sebras: the Shade.bound one? probably. | 13:34.01 |
| we ought to expose more stuff there, so you could hold off until I add the mesh processor callbacks to mujs | 13:34.19 |
sebras | ok. | 13:34.44 |
| I was looking at the convenince objects/functions. | 13:34.52 |
| like color.equal() and color.convert(). | 13:35.00 |
tor8 | sebras: those are not for 'mutool run' they are for the acrobat js | 13:35.09 |
sebras | oh, util.js! I'm so silly today. | 13:35.24 |
| s/today/always/ | 13:35.32 |
| tor8: btw, I noticed the library libbabl yesterday. GIMP appears to think they can replace lcms2 with that (and it is supposedly faster too). I mentioned it to a certain michael http://gegl.org/babl/ | 13:36.46 |
moolc | sebras: this babl thing is choke-full of glibism abominations.. (not that this is an indication of quality of the code, but still - leaves a bad aftertaste) | 14:25.49 |
sebras | moolc: I guess that's not unexpected coming from gegl. | 14:33.07 |
| moolc: oh well, maybe something useful can be learned from it anyway. | 14:33.24 |
moolc | sebras: just sayin'.. there's probably some decent code in glib/gtk standard breakage notwithstanding | 14:33.58 |
| tor8: I would have changed "Fix windows: strsep is missing on MSVC, use fz_strsep instead." comment... strsep is nonstandard BSD extension it's hardly fair to blame MS for not providing it | 14:35.20 |
sebras | ok, small patch on sebras/master. it clusters well. | 14:55.10 |
| Robin_Watts: small, silly patch on sebras/master. | 17:07.08 |
Robin_Watts | the jpeg one? makes sense. | 17:08.35 |
sebras | yes, that one. | 17:15.57 |
| I took the liberty of uppercasing dct in the commit message and rebasing onto master. I'll go a head and push unless you stop me. :) | 17:16.43 |
| Robin_Watts: where you is you. | 17:16.52 |
Robin_Watts | go for it. | 17:22.17 |
sebras | Robin_Watts: done. | 17:25.51 |
user___ | @tor8 : Hello, I am the person who had trouble copying text from pdf yesterday, on linux mupdf | 18:52.06 |
| I am using mupdf-gl package from the community repository on Manjaro | 18:52.43 |
| The freeglut package is indeed the one from the repository, how can I check if it behaves the same as your enhanced version ? | 18:54.34 |
tor8 | user___: if copying text doesn't do anything, it's not our enhanced version | 19:07.31 |
| user___: our enhanced version also supports unicode input. stock freeglut only supports latin-1. | 19:07.50 |
| user___: so if you can enter "Ω" in the search field, you've got a good version. if you can't then you've got stock freeglut. | 19:08.41 |
user___ | @tor8 : Any tips on inputing unicode characters in the search field. For instance I open the search field with </> key, then type : <Ctrl>-<Shift> <U> <0> <0> <C> <6> <Enter> for Ã, but all it does is close the search field | 19:22.05 |
| @tor8 : it does seem that the package version is the stock one. Do you know whether an update is planned, that would take into account your enhancements, or should linux users all compile mupdf-gl to have correct behaviour ? | 19:28.51 |
moolc | user___: tor8 is no longer here | 19:30.20 |
| Forward 1 day (to 2018/05/10)>>> | |