| <<<Back 1 day (to 2015/03/25) | 20150326 |
Guest9244 | Anyone else getting build errors with the master in Android? | 03:26.36 |
| [armeabi-v7a] Compile thumb : mupdf <= mupdf.c In file included from jni/mupdf.c:16:0: ../../include/mupdf/pdf.h:6:34: fatal error: mupdf/pdf/name-table.h: No such file or directory compilation terminated. | 03:28.11 |
| git status reports in line with master | 03:28.31 |
| Out of interest, how much of a donation would it require to have MuPDF add a two-page view feature to the core? | 03:37.15 |
| the core code looks like it should allow this and I know of another project which has done it on an older version of the code (but they've seriously wandered away from master so can't really be used as-is) | 03:38.03 |
kens | Another SO MuPDF question, this one looks like its probably solvable, I'm guessing the user needs to do some memory svaneging, but I have no idea how to answer it (its also possible its just Java being Java): | 09:56.40 |
| http://stackoverflow.com/questions/29270831/usingt-mupdf-and-a-list-view-in-two-seperate-activities-eventually-runs-out-of-m | 09:56.40 |
tor8 | Guest9244 (for the logs): if you get errors missing a file mupdf/pdf/name-table.h you need to run "make generate" again | 10:08.00 |
Robin_Watts | tor8: various small fixes on robin/master | 10:16.55 |
| tor8: How are the fz_off_t/LARGE_FILE changes coming? | 11:36.59 |
kens2 | chrisl ping | 11:54.04 |
Guest53153 | hi all | 11:54.24 |
chrisl | kens2: pong | 11:54.32 |
kens2 | chrisl is the GS Windows code compiled thread safe ? | 11:54.43 |
chrisl | kens2: no | 11:54.59 |
kens2 | And is there a way to tell if a Linux build is compiled thread safe or not ? | 11:55.03 |
chrisl | No | 11:55.08 |
kens2 | OK thanks | 11:55.14 |
chrisl | I suggested we should add an API call for it, but it never happened | 11:55.30 |
Guest53153 | i have a question about mupdf. how i can add image annotation in pdf file? | 11:56.08 |
kens2 | Ah, OK, got someone asking about multi-threading on SO< and I *think* they mean spawning multiple threads using GS, rather than using -dNumRenderingThreads. | 11:56.08 |
chrisl | The only way to be sure it's built threadsafe is to build it yourself and use the threadsafe option - none of our default builds are built that way. | 11:57.11 |
kens2 | Yeah that's basically what I just told him | 11:57.23 |
| Of course he's using Ghost4J so I can't help much anyway | 11:57.44 |
| Robin_Watts : paulgardiner Guest53153 is the questioner on SO about 'image annotations' | 11:58.23 |
chrisl | We should probably look at resolving the non-threadsafe devices and making threadsafe the normal build...... | 11:58.24 |
kens2 | It woudl seem reasonable to make it thread safe by default, unless there's some good reason not to. | 11:58.47 |
| non-threadsafe devices are one good reason of course | 11:59.01 |
chrisl | Exactly..... | 11:59.09 |
kens2 | Oh great Zoltan's back again | 12:00.09 |
paulgardiner | Guest53153: we discussed this a bit yesterday. We aren't sure why what you were doing doesn't work, assuming you are viewing the result using MuPDF. | 12:00.46 |
| Guest53153: But, probably using a Square annotation isn't your best option. The problem with doing so is that the result wont display in Acrobat. | 12:01.47 |
| Acrobat doesn't trust the appearance stream for annotations and regenerates it, so your image would be lost. | 12:02.36 |
| A better solution might be to add the image to the doc stream | 12:04.15 |
Guest53153 | but I trying read annotations using mupdf and later render it using direct2d. and native images in pdf document rendering correctly, but images that I added not rendering | 12:06.54 |
kens2 | Guest53153 : Images in the page content stream are images, and will be rendered correctly by any conforming reader. As I explained os Stack Overflow, there is *not* 'image' annotation type, so you can't simply add such a thing. There are annotation types which might be abused to create an annotation with an appearance which contains an image, but any consumer (such as Adobe Acrobat) which regenerates the appearance, instead of u | 12:08.56 |
| sing tghe one in the PDF file will not then display the image ofr your annotation. | 12:08.56 |
| As Paul said, you should first try rendering the PDF file you created with MuPDF and see what happens then. | 12:09.45 |
| Its not clear to me how you are 'reading annotations using MuPDF' and at the same time 'later render it using direct2d' | 12:10.29 |
Guest53153 | It means that I get fz_image file, get fz_pixmap from them and drawing bit map via direct2d | 12:14.17 |
kens2 | OK so the first thing to try, as I said, is to open the resulting PDF file with MuPDF and look at it, do you see your image ? | 12:15.01 |
| It would also be entirely sensible to post your PDF file (the one with the image added) somewhere that we cna look at it. | 12:15.27 |
Guest53153 | 'OK so the first thing to try, as I said, is to open the resulting PDF file with MuPDF and look at it, do you see your image ?' - no, I do not see | 12:17.44 |
kens2 | Then the image probably isn't there. | 12:18.06 |
Guest53153 | can I attach pdf to this chat? | 12:18.41 |
kens2 | No, put it somewhere public, like DropBox | 12:18.51 |
Guest53153 | ok, if I can't solve this problem today - I will put pdf and some code to GitHub | 12:20.44 |
kens2 | lunches | 12:32.14 |
Guest53153 | Thanks all! I'm solved this trouble. Its was very easy - it was problems in direct2d. I just used incorrect settings for bitmap) | 13:16.57 |
kens2 | GGood to hear you got it working satisfactorily | 13:17.29 |
Robin_Watts | Balls. I wanted to quiz Guest53153 about his usage and the license :( | 13:29.36 |
kens2 | Oops, I shoudl have done that, sorry, You oculd put a comment on the SO question, he *might* read it.... | 13:30.08 |
Robin_Watts | Morning mvrhel_laptop | 13:40.06 |
| bug 695861 looks odd to me. It looks like he's altering some of your code to scale stuff? | 13:40.52 |
henrys | Robin_Watts: the upload test file, where are the files going? I don't see my file in the tests. Looking at your code it looks like it adds the file directly to the repo? Is that legit or do you have to checkout a tree somewhere? | 14:14.08 |
Robin_Watts | henrys: It adds the files to tests_private/<type>/uploads/ | 14:17.19 |
| or something like that. | 14:17.23 |
| and the uploads dir should be in the tests. | 14:17.34 |
| I believe it's legit :) | 14:17.43 |
henrys | Robin_Watts: huh I'm sure I uploaded something but there is no uploads dir or the file. I'll try again. | 14:19.16 |
Robin_Watts | I'm just running an svn update now to see. | 14:19.48 |
henrys | Robin_Watts: don't spend much time on it I know you have a lot going on. | 14:20.25 |
Robin_Watts | henrys: hmm. the uploads dirs aren't showing up. | 14:24.04 |
tor8 | Robin_Watts: bah! there's no OFF_MAX to match SIZE_MAX :( | 14:28.36 |
Robin_Watts | tor8: That doesn't matter. We can't use off_t, hence we can't use OFF_MAX. | 14:29.11 |
tor8 | I might be dense, but why can't we use off_t again? | 14:30.08 |
Robin_Watts | Because we want to be able to do large file support on 32bit builds. | 14:30.42 |
| and windows does not have off_t. | 14:30.56 |
tor8 | off_t is 64-bit if LARGEFILE is defined | 14:30.58 |
| and if the answer is windows, my reply is #define :) | 14:31.09 |
Robin_Watts | We should define our own type: fz_off_t and use that. | 14:31.18 |
| If you want to make fz_off_t point to off_t on linux, I won't complain too loudly. | 14:31.55 |
tor8 | now we should probably namespace the windows define with fz_off_t, but on posix systems that *do* have off_t I'd expect fz_off_t to be a simple #define fz_off_t off_t | 14:31.59 |
Robin_Watts | tor8: right. | 14:32.14 |
| typedef, not #define, please :) | 14:32.29 |
tor8 | fair enough :) | 14:32.38 |
Robin_Watts | I've got to grab lunch. bbiab. | 14:32.38 |
tor8 | I'm hoping to get this to be as 'conventional' as possible, in line with what POSIX programs are expected to do | 14:32.49 |
| and at the same time maybe get us to be signed/unsigned comparison warning free | 14:33.13 |
| like MuJS is | 14:33.18 |
| Robin_Watts: ugh. rabbit hole :( | 16:07.01 |
Robin_Watts | tor8: mmm. | 16:15.41 |
| I'm looking at our transparency handling at the moment. | 16:15.59 |
| actually, I'm going to look at the svn upload stuff first. | 16:16.53 |
marcosw1 | I believe bmpcmp is working on i7 again, so I've added it back into the cluster. | 16:21.03 |
Robin_Watts | ok, so svn ls is showing that the uploaded files exist within svn. | 16:26.06 |
| why won't update check them out? | 16:26.11 |
| henrys: OK, so the upload script has been committing everything into the wrong place :( | 16:38.41 |
| should be fixed now. | 16:38.50 |
| give me a mo and I'll move everything over. | 16:38.56 |
henrys | Robin_Watts: thanks for fixing that. | 16:39.18 |
Robin_Watts | np. | 16:39.23 |
mvrhel_laptop | morning Robin_Watts | 17:02.28 |
Robin_Watts | morning. | 17:02.36 |
| I was running through the mupdf bugs last night, and found a patch for the winrt stuff. | 17:02.50 |
| possibly from someone on acid. | 17:02.56 |
mvrhel_laptop | was on just for a bit earlier. At orthodontist. Ethan is getting braces. $5K lighter in my wallet :( | 17:03.14 |
| Robin_Watts: ok. I can take a look at it | 17:03.33 |
Robin_Watts | mvrhel_laptop: ouch. | 17:03.40 |
| I suspect that he's abusing your code to do something that isn't intended. | 17:04.34 |
kens | OK I'moff, night all | 17:04.34 |
Robin_Watts | it looks to me like he's ignoring the 'scale' supplied and calculating one based on the width/height supplied. | 17:05.00 |
mvrhel_laptop | that is odd | 17:05.36 |
Robin_Watts | yeah, but my understanding of that function is minimal at best :) | 17:05.58 |
mvrhel_laptop | what is the bug #? | 17:06.05 |
Robin_Watts | bug 695861 | 17:06.50 |
fredross-perry | new 32-bit gsview installer: http://ghostscript.com/~fred/gsview/linux/GSView-6.0-Installer-32 | 18:04.05 |
| this one installs a minimum set of shared libs alongside the app. Works for me on Fedora, and an out-of-the-box Mint. Please report back. | 18:04.47 |
Robin_Watts | henrys: Was '2bars.pcl' the file you uploaded? | 18:38.27 |
mvrhel_laptop | ok finally back | 18:54.00 |
henrys | Robin_Watts: yes | 18:55.01 |
Robin_Watts | henrys: OK, hopefully that should be in future tests. | 18:55.32 |
| mvrhel_laptop: In case you didn't see before you disconnected: bug 695861 | 18:56.04 |
henrys | fredross-perry: did we fix the iconv issue in the mac build I think that came up before I left for Korea and I forgot about it? | 18:56.06 |
Robin_Watts | but it's not urgent. | 18:56.15 |
fredross-perry | yes, did that. | 18:56.21 |
mvrhel_laptop | thanks Robin_Watts | 18:57.02 |
Robin_Watts | tor8: I've got a couple of transparency fixes on robin/master | 20:38.36 |
| (along with the other stuff) | 20:38.56 |
| marcosw1: bmpcmp still failing :( | 20:49.15 |
| Forward 1 day (to 2015/03/27)>>> | |