| | <<<Back 1 day (to 2016/04/17) | 20160418 |
| Maori | Again | 01:49.27 |
| | Hi,there.I'm trying to get the font information of the special character by using the Mupdf.Using the fz_text_char->style->font->name.But the result seemed to be wrong.Can someone tell me how to get it? | 01:49.35 |
| | http://pan.baidu.com/s/1coLoYU | 01:49.45 |
| tor8 | Maori (for the logs): you have to give us access to the PDF file, or we can't really help. | 09:10.15 |
| sebras | tor8: so today is 1.9? | 09:31.57 |
| tor8 | sebras: that's the hope. | 09:32.09 |
| sebras | tor8: right, so once the release is done I guess the fixed on sebras/master are scheduled to go in? | 09:32.57 |
| tor8 | sebras: correct. once the release is out, all our branches will rush to get pushed to master :) | 09:33.25 |
| sebras | tor8: so what is hindering the release at this moment in time? | 09:33.50 |
| Robin_Watts_ | tor8: OK, so what needs to be pushed to master... | 09:33.52 |
| | I've got a leak fix, and the 2 sanitize fixes. | 09:34.18 |
| tor8 | Robin_Watts_: and the last MuJS update fix | 09:34.27 |
| | then I think we're good | 09:34.33 |
| Robin_Watts_ | You push your mujs fix. | 09:34.40 |
| tor8 | Robin_Watts_: done. | 09:38.13 |
| Robin_Watts_ | And that's my fixes pushed. | 09:52.36 |
| tor8 | Robin_Watts_: And that's tagged. | 09:54.29 |
| sebras | tor8: oh no! | 09:54.59 |
| | tor8: Robin_Watts missed a trailing period at the end of his commit subject line! | 09:55.20 |
| tor8 | sebras: .... | 09:55.25 |
| sebras | tor8: :) | 09:55.26 |
| tor8 | don't scare me like that! | 09:55.29 |
| sebras | tor8: >;) | 09:55.37 |
| | Robin_Watts_: what do you mean about interchangeability in the BDC patch? | 09:57.45 |
| tor8 | sebras: there are two forms of BDC operators | 09:58.24 |
| sebras | tor8: yes I know. | 09:58.29 |
| tor8 | /Name BDC and <<...dictionary>>> BDC | 09:58.38 |
| Robin_Watts_ | sebras: Previously I was always looking the Name up and converting it to the inline form. | 09:59.12 |
| | So that the operator was always called with <<dictionary>> BDC | 09:59.24 |
| | Which is great for execution, cos you don't care which form it was. | 09:59.37 |
| | But for the sanitise stuff, the operator has to write the source version out again. | 10:00.10 |
| | and Acrobat won't accept /PlacedPDF <<dictionary>> BDC | 10:00.38 |
| | It will only accept /PlacedPDF Name BDC | 10:00.49 |
| | i.e. the two forms are not interchangeable (even though the spec says they should be ) | 10:01.10 |
| sebras | Robin_Watts_: hm, so basically this is a workaround for acrobat not accepting the <<dict>> form then. | 10:01.25 |
| Robin_Watts_ | Not accepting the dict form *in all cases*. Yes. | 10:01.39 |
| sebras | ah I see. | 10:01.49 |
| | Robin_Watts_: so that means that prior to your bugfix we had <<dict>> and the same dict entries then appeard in the Properties subdictionary, right? | 10:02.34 |
| Robin_Watts_ | Yes. | 10:02.41 |
| sebras | Robin_Watts_: couldn't it be this that set acrobat off? | 10:03.02 |
| Robin_Watts_ | sebras: I would be surprised. | 10:18.18 |
| | but regardless, we now preserve files in a nicer way. | 10:18.28 |
| sebras | Robin_Watts_: ok. | 10:19.57 |
| | Robin_Watts_: prompted by your pdf_dict_put() fix I looked at all other calls to pdf_dict_put[l](). would you mind reviewing the top patch of sebras/master? I think there are a few leaks in pdfposter and one in pdf-write.c | 11:11.07 |
| Robin_Watts_ | sebras: Sure. | 11:11.43 |
| | sebras: Not sure about the first one. | 11:12.46 |
| | Oh, wait... yes. | 11:13.08 |
| sebras | Robin_Watts_: because it is a copy then I think we do leak... ok. :) | 11:13.33 |
| Robin_Watts_ | Yes, I think they are all correct. | 11:14.04 |
| sebras | Robin_Watts_: thanks. then I'll wait for tor8 to merge my master branch. | 11:14.38 |
| | Robin_Watts_: but I guess you guys will now be busy with building all the release binaries? | 11:14.57 |
| Robin_Watts_ | sebras: tor8 is doing that, I think. I've got my head in some gs stuff. | 11:22.18 |
| sebras | Robin_Watts_: ok. | 11:23.07 |
| Van_ | Hi, could someone help me with building mupdf? I followed the steps of the ReadMe.txt but on the step 10 "make" it throws an error because of /cygdrive/d/Downloads/mupdf-1.8-source/platform/x11/win_main.c:1235: undefined reference to `fz_argv_from_wargv' /cygdrive/d/Downloads/mupdf-1.8-source/platform/x11/win_main.c:1291: undefined reference to `fz_free_argv' | 14:09.47 |
| | I read that errors are to be expected. The "generated" is created and filled. On step 12 "ndk-build.cmd" the system can't find make: *** [obj/local/armeabi-v7a/objs/mupdf/mupdf.o] Error 2 | 14:13.33 |
| sebras | Van_: the sample viewer for Windows is normally compiled using VS2005 (I think it's that version at least), not cygwin. | 14:13.58 |
| | Van_: oh! you are trying to build the android app on Windows. | 14:15.27 |
| Van_ | Ah yes, forgot to add that bit of information | 14:15.47 |
| sebras | Van_: so in step 10 what command did you run? | 14:17.28 |
| Van_ | I run in /cygdrive/d/Downloads/mupdf-1.8-source "make" and i get the undefined reference error | 14:18.54 |
| Robin_Watts_ | Van_: You're on windows? | 14:19.04 |
| Van_ | Yes, windows 7 | 14:19.18 |
| Robin_Watts_ | So, it looks to me like you're trying to use cygwin to build on windows. It's possible that that is broken- it would explain the fz_argv_from_wargv and fz_free_argv thing. | 14:20.07 |
| | But that's fine, because generated was set up before then. | 14:20.21 |
| | So... the problem is the error you are seeing on Step 12. | 14:20.34 |
| sebras | Robin_Watts_: Van_: wouldn't it be enough to run "make generate" in step 10? | 14:20.46 |
| Robin_Watts_ | sebras: Yes. | 14:20.56 |
| sebras | Robin_Watts_: if Van_ simply used "make" then this is the cause of the first batch of errors above. | 14:21.21 |
| Robin_Watts_ | So, I'm not sure why it's trying to make mupdf.o in ndk-build. | 14:21.57 |
| Van_ | Alright, but as you said the "generated" folder ist created and filled with 9 header files | 14:22.02 |
| sebras | Van_: seems about right. | 14:22.18 |
| Robin_Watts_ | Van_: Yes. ignore that problem, and we'll just talk about the last one. | 14:22.27 |
| Van_ | Alright | 14:22.41 |
| Robin_Watts_ | Can you ndk-build clean, then ndk-build and paste the whole results into pastebin.com please? | 14:22.49 |
| sebras | Robin_Watts_: LOCAL_SRC_FILES := mupdf.c is in jni/Android.mk though. and I actually build mupdf.o here too. | 14:23.55 |
| Robin_Watts_ | D'Oh. I was thinking it should be called mupdf_native.c, but that's a different thing. | 14:24.24 |
| Van_ | Here it is http://pastebin.com/sPebhePG | 14:24.41 |
| Robin_Watts_ | Van_: So, it should be trying to build platform/android/jni/mupdf.c | 14:24.45 |
| | Do you have platform/android/jni/mupdf.c ? | 14:25.44 |
| Van_ | Yes, it exists there | 14:26.28 |
| Robin_Watts_ | Ok, no idea why that's not working then. | 14:26.48 |
| | Oh. Try running ndk-build.cmd from a DOS prompt rather than from cygwin. | 14:27.36 |
| Van_ | Same error message | 14:28.24 |
| Robin_Watts_ | Van_: I am confused then, sorry. | 14:29.47 |
| sebras | Van_: are you sure you have the compiler in D:/TM3/android-ndk-r11c/build//../toolchains/arm-linux-androideabi-4.9/prebuilt/windows-x86_64/bin/arm-linux-androideabi-gcc ? | 14:30.11 |
| Robin_Watts_ | It might be worth you swapping to use 1.9rc2 that's out now. | 14:30.15 |
| | sebras: ooh, good thought. | 14:30.46 |
| sebras | Van_: in case you have updated your android environment and forgotten to update local.properties. | 14:30.57 |
| Van_ | Alright, this seems on something, the path above does not exists like that | 14:32.02 |
| | It's D:\TM3\android-ndk-r11c\toolchains\arm-linux-androideabi-4.9\prebuilt\windows without "_x86_64" | 14:32.15 |
| Robin_Watts_ | Van_: Are you trying to use the 64bit ndk to do 32bit builds? | 14:33.07 |
| Van_ | windows/bin/arm-linux-androideabi-cpp.exe does exist though | 14:33.09 |
| Robin_Watts_ | Cos Google had that screwed up for a while. | 14:33.25 |
| Van_ | If I'm not mistaken I downloaed 32bit ndk. I will recheck it | 14:33.46 |
| Robin_Watts_ | Van_: Your best bet these days is to install Android Studio 2.0, and get the sdk and ndk that comes with that. | 14:34.20 |
| sebras | Van_: check your platforn/android/local.properties as well. | 14:34.39 |
| Robin_Watts_ | Our project doesn't build using AS2.0, but we're moving in that direction. | 14:34.39 |
| Van_ | The zipfile is android-ndk-r11c-windows-x86.zip, I will try Android Studio then | 14:34.50 |
| | sebras: I set then up like that: sdk.dir=D:\\TM3\\EclipseAndroid\\sdk | 14:35.30 |
| | Seems there is the ndk line missing... | 14:35.57 |
| | Trying the Android Studio approach | 14:38.18 |
| | During the download/install of Android Studio I tried with the 1.9 version. Same error message on ndk-build.cmd. It's expecting something on windows-x86_64 but I have only windows. Should I just rename it? | 14:47.34 |
| sebras | Van_: could you try doing "export NDK_LOG=1" and then running ndk-build and pasting it to pastebin.com? | 14:52.48 |
| | Van_: on linux this will cause extra debug output, hopefully letting us know how it detects what compile to use. | 14:53.21 |
| Van_ | sebras: Here it is http://pastebin.com/bHYAjMK9 | 14:59.10 |
| | sebras: I think I missunderstood the bit compatibility. I downloaded the 64bit NDK and its doing something now. | 14:59.37 |
| | sebras: This is the new error message http://pastebin.com/6WWWjsq6 | 15:03.04 |
| sebras | Van_: hm... I don't understand why the error messages are shown. | 15:09.52 |
| | Van_: when I build under linux libmupdf.so doesn't have isinf or isfinite defined either. | 15:10.24 |
| | Van_: but I'm using a stone age android build environment. | 15:10.37 |
| Van_ | sebras: Hmm, that's strange. Could you point me to the NDK of Android Studio 2? Can't find it | 15:11.05 |
| sebras | Van_: no, unfortunately not. I have never used it and so I would simply google for it myself. | 15:13.23 |
| Robin_Watts_ | Van_: When you run AS2.0 it should offer to download it. | 15:13.27 |
| | It'll be in the list of 'updates'. | 15:13.37 |
| Van_ | +Robin_Watts_: Thanks, I found it under System settings > Android SDK > SDK Tools | 15:14.59 |
| | +Robin_Watts_: Seems AS2 does the work. ndk-build.cmd went through without error messages! | 15:24.23 |
| Robin_Watts_ | Fab. | 15:24.31 |
| Van_ | Pretty strange but I'm happy. Thank you and sebras for you help and time | 15:25.20 |
| Robin_Watts_ | np. Let sus know how you get on. | 15:26.22 |
| | Can I ask what you're using MuPDF for please? (Just being nosey) | 15:26.38 |
| Van_ | +Robin_Watts_: I will keep you guys posted. I was delegated to test it out as an in-app pdf previewer | 15:40.36 |
| Robin_Watts_ | Van_: Cool. Just be aware of the licenses. | 15:42.16 |
| | Either it's GNU AGPL (which rules it out of use for most commercial apps), or you need a commercial license from us. | 15:42.49 |
| | That's not a block to you evaluating it of course! | 15:43.00 |
| Van_ | +Robin_Watts_: Thanks for the reminder, I think that went under the radar | 15:44.46 |
| | I will be testing out the implementation tomorrow and report back how it went. Thanks again for the help! | 15:48.25 |
| | Forward 1 day (to 2016/04/19)>>> | |
IRC Logs