| <<<Back 1 day (to 2018/04/11) | 20180412 |
tor8 | bilabila (for the logs): mupdf-x11 or mupdf-gl? | 09:24.58 |
| Robin_Watts: I've tagged a 1.13.0-rc1 and uploaded the source tarballs. could you do the windows build as usual? | 12:54.22 |
Robin_Watts | tor8: Sure. | 13:05.48 |
| tor8: The VSproject files are broken :( | 13:13.35 |
| tor8: http://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=4a7822d6750ecb6e0b63f4357738dc20ecaa58f6 | 13:15.41 |
tor8 | Robin_Watts: *sigh* | 13:15.58 |
| we *really* need to sort something out in this regard | 13:16.08 |
| Robin_Watts: LGTM | 13:16.14 |
Robin_Watts | tor8: premake should be a fairly high priority thing after the release, right? | 13:16.33 |
tor8 | Robin_Watts: yes. the main issue being thirdparty libraries needing weird -D flags that differ on various platforms. | 13:17.10 |
Robin_Watts | ok, with that it builds and tigers. | 13:17.46 |
tor8 | I've managed to get a working premake vs build | 13:17.56 |
| Harfbuzz breaks in more than one way for VS2005 | 13:18.13 |
Robin_Watts | We can stick with old harfbuzz though, right? | 13:18.27 |
tor8 | it started using msvc-specific functions that aren't in VS2005 | 13:18.31 |
| the stdint.h issue is easy enough to work around | 13:18.54 |
| harfbuzz commit 6c10328eb88b0641586f31a782b63f45eada5609 is worse | 13:19.15 |
Robin_Watts | Does the generated VS project have Release/Debug/Profile/Memento configurations? | 13:19.36 |
| How about DebugJava/ReleaseJava? Or DebugCommercial/ReleaseCommercial ? | 13:20.04 |
tor8 | Robin_Watts: yes. premake supports configurations. | 13:20.12 |
Robin_Watts | If yes to all of those, then great. | 13:20.32 |
| I've pushed the commit. | 13:20.53 |
| Do you want to move the tag? :) | 13:20.57 |
tor8 | I'll move the tag :) | 13:21.00 |
| and rebake the tarballs | 13:21.04 |
| I don't know how to get premake to work with the android build though. but maintaining the Makefiles is trivial in comparison. | 13:23.14 |
| even if we end up only using premake to generate the visual studio project files, I think that will buy us enough benefit | 13:23.56 |
Robin_Watts | tor8: VS and Xcode would be a win. | 13:24.39 |
tor8 | I did some experiments, and I could coax regular make to spit out a CMakeLists.txt that built the same stuff | 13:25.07 |
Robin_Watts | but if we can get linux/VS/Xcode out of the same thing, it'd solve the usual "I broke platform X cos I was only testing on platform Y" thing. | 13:25.22 |
| tor8: Yell when you move the tag. | 14:06.23 |
tor8 | Robin_Watts: I already have | 14:06.36 |
| but you may need to remove your local tag or it won't update | 14:06.45 |
Robin_Watts | Right, that was it. | 14:07.02 |
| git fetch --tags sorted it. | 14:07.34 |
| tor8: So... my CHANGES file says 1.12 still | 14:27.42 |
| Ignore me.! | 14:28.01 |
| tor8: mupdf-1.13.0-windows.zip in my homedir on casper now. Move it/Rename it to whereever you want. | 15:02.14 |
| tor8: Can we put a link to mupdf_explored.pdf on the documentation page of mupdf.com ? | 15:02.16 |
| I can do a new build now with the latest changes, and the number updated to 1.13 | 15:02.19 |
tor8 | I'm popping out for a few hours. If you put it somewhere on casper, I'll add a link to the docs.html page | 15:06.22 |
Robin_Watts | tor8: OK, latest version: http://ghostscript.com/~robin/mupdf_explored.pdf | 16:32.33 |
tor8 | Robin_Watts: I've linked to it from the mupdf.com/docs/ page under the "Library" heading. | 21:03.39 |
fredross-perry | tor8, robin, sebras - looking for debugging advice for #699213. In the logs you'll see crashes at or near fz_drop_page, and fz_free. What the user is doing is printing, which amounts to doing PDFDocument.save("some path", ""), and then passing the path to Android's printing system. I am not seeing the crashing. | 22:35.23 |
Robin_Watts | fredross-perry: Smells like heap corruption. | 22:58.52 |
| I'd reach for Memento. | 22:58.56 |
fredross-perry | Dr. Memento | 22:59.17 |
| This is (I think) that same Galaxy J7. | 22:59.45 |
Robin_Watts | I'd wonder if it's running out of memory. | 23:00.29 |
| That would make it throw, and possibly there is a catch clause that's wrong or something. | 23:00.54 |
| You could try making the fz_throws print ? | 23:01.17 |
fredross-perry | sure. But remember I am not crashing, it's only Silver at this point. I shoud ask him to try other device. | 23:17.42 |
Robin_Watts | fredross-perry: Yeah, but making fz_throws print in builds for silver won't hurt... | 23:20.36 |
fredross-perry | lemme see... | 23:21.09 |
| Is there a handy stack trace printer in the code somewhere? | 23:22.03 |
Robin_Watts | actually, it looks like they should already throw. | 23:22.40 |
| already print. | 23:22.43 |
| see fz_vthrow. | 23:22.52 |
| A handy stack trace printer? Well, Memento knows how to backtrace :) | 23:23.14 |
fredross-perry | I would imagine. | 23:23.30 |
Robin_Watts | (using 3 different methods depending on OS) | 23:23.32 |
fredross-perry | re fz_vthrow - looks like I need to define USE_ANDROID_LOG ? | 23:26.36 |
Robin_Watts | #ifdef __ANDROID__ | 23:33.59 |
| #define USE_ANDROID_LOG | 23:34.00 |
| #include <android/log.h> | 23:34.02 |
| #endif | 23:34.03 |
| That's at the top of error.c | 23:34.10 |
| Might be worth maybe testing to see whether __android_log_print(ANDROID_LOG_WARN, "libmupdf", "balls"); shows up. | 23:39.17 |
| No, wait ANDROID_LOG_ERROR | 23:40.54 |
fredross-perry | trying... | 23:42.15 |
GenkiSky | Hello, I'd like to submit a small patch for the x11 mupdf client. I couldn't find any mailing list, or other method, to do so. Am I missing something? | 23:42.30 |
Robin_Watts | GenkiSky: Hi. | 23:47.51 |
| go to bugs.ghostscript.com, open a bug, and attach a patch there. | 23:48.12 |
| Use "Mupdf" as the product, obviously :) | 23:48.26 |
| Genkisky: If you want to talk over the patch with a developer, try back here in about 12 hours. We're on european time mostly. | 23:49.04 |
| fredross-perry: bedtime, sorry. talk more tomorrow. | 23:49.31 |
fredross-perry | ok. | 23:49.39 |
GenkiSky | Robin_Watts: Great, thanks. I'll do that. I don't think I need to talk it over first, but I'll be happy to do so if people think otherwise on seeing it :) | 23:50.28 |
| Thanks for the heads up though. | 23:50.45 |
| Forward 1 day (to 2018/04/13)>>> | |