| <<<Back 1 day (to 2013/01/20) | 2013/01/21 |
mvrhel_laptop | Robin_Watts: tor8 for the logs (its late there) I was able to get everything working with the suggestions Robin made. Now to add in some UI for multiple pages and zooming | 00:18.59 |
BW^- | can I use ghostscript to convert a PDF to XPS? | 01:44.59 |
| (as in, OpenXPS, the stuff Windows' print spooler eats) | 01:45.08 |
henrys | BW^-: not yet but we're working on it. | 02:21.34 |
BW^- | henrys: cool! is there any beta level support for it or alike? | 02:22.30 |
| henrys: anything i could put my hands on now? | 02:22.37 |
henrys | no I'd check back in a couple of weeks though. | 02:23.19 |
BW^- | henrys: aha ok. in the meantime, are you aware of any bsd/linux command line program that could do this? | 02:24.39 |
henrys | not that I know of have you looked if mono does anything with printer drivers? | 02:26.57 |
| BW^: curious why you want this, to support a printer? we jumped on the xps bandwagon early with an interpreter but really haven't seen much business. | 02:31.37 |
BW^- | henrys: appears that any printing you want to do in windows, is best made with a ready XPS source | 02:32.21 |
| henrys: at least if it's about taking a doc from a ready-made source and just feeding the OS with it, with as-little-code-as-possible | 02:32.43 |
| henrys: indeed I see in ghostscript the "windows printer" driver too, that requires the GS to be bundled on client machine which may be completely fine of course, just a couple of megs. | 02:33.21 |
| henrys: ah, an off-topic thing, if I have a .ps file, can I parse out which the rightmost position of it has black [text] content easily? (as to crop it) | 02:58.56 |
henrys | BW^-: not quite following your question you can use the bounding box device to find the extant of graphics on the page. You probably want to look at ghostscript stack overflow, they field all sort of questions like that. | 03:52.28 |
BW^- | henrys: aha - the bounding box is wrong and i need to manually detect where the black content is. i believe i'll use some rasterization program and then check the pixel coords there. thx. | 04:03.33 |
deleet | Robin_Watts: feel free to pm me if I'm not around | 06:05.36 |
| (and leave a msg ofc) | 06:05.48 |
Robin_Watts | deleet: The recycle stuff is cauging crashes. I'm going to work on rejigging it today. | 09:13.17 |
| If you could be around to test what I come up with, I'd be grateful. | 09:13.30 |
| but right now, I need to go dig my van out of the snow. | 09:14.18 |
kens | goes to kick network repeatedly until it co-operates | 09:14.26 |
Robin_Watts | hi paulgardiner. I am about to go out to dig the van out/take helen to the station, but I did find the source of the crashes etc. | 09:22.56 |
| The recycle() calls are indeed breaking things. | 09:23.04 |
| so I think I have a plan to fix it. | 09:23.14 |
| I'll call you to talk about it when I get back. | 09:23.23 |
paulgardiner | Ah great | 09:23.37 |
Robin_Watts | I think basically, we will have to implement our own reference counting system on BitmapHolders. | 09:23.56 |
| and only recycle when they become free. | 09:24.09 |
paulgardiner | sounds sensible. Strange though. I thought I could reason that any render to a recycled bitmap would be for a page we no longer needed | 09:25.33 |
Robin_Watts | paulgardiner: http://pastebin.com/uqbZPaxE | 11:18.23 |
paulgardiner | :-) "Trying to use a recycled bitmap" That's a give away | 11:37.42 |
tor8 | well, that's curious. business is cheaper than economy plus... | 11:47.31 |
kens | We had that once | 11:49.37 |
| business cheaper than economy in fact | 11:50.00 |
tor8 | kens: that's even odder :) we're flying into MIA right? | 11:50.13 |
kens | Yes, Miami international | 11:50.22 |
| Of course, that was Virgin, so its not called business, its premium economy there | 11:52.37 |
tor8 | well, BA calls it world traveler or something silly | 11:53.15 |
kens | Yeah, silly names all round | 11:53.25 |
Robin_Watts | got to go jump start a neighbours car. brb. | 11:58.16 |
tor8 | effing useless airlines... they want to charge 25 gbp to reserve a seat, but none of the exit and bulkhead seats are available. so what's the point then? | 12:02.25 |
kens | BA ? | 12:02.34 |
tor8 | BA. | 12:02.37 |
kens | Starting to sound like RyanAir | 12:02.41 |
tor8 | the only ones with a decent flight schedule I've found | 12:02.57 |
kens | lunches, mainly to warm up, bbiab | 12:05.06 |
Robin_Watts | http://git.ghostscript.com/?p=mupdf.git;a=commitdiff;h=58dafe4c6280390b9e6fc93848ec522b46019966 | 12:37.35 |
| http://git.ghostscript.com/?p=mupdf.git;a=commitdiff;h=2e66d56cd8d3e5461b2733df89b9453db5db783e;hp=0c12559e1f12121d0e414a42fa99a386f9c906ac | 12:42.08 |
| paulgardiner: Pushed your fix. | 13:06.36 |
| Going to get lunch now. | 13:06.47 |
paulgardiner | Robin_Watts: ok. Nearly done | 13:07.33 |
| Robin_Watts: change is up on paulg/master | 13:29.32 |
Robin_Watts | paulgardiner: The blur fix works. | 14:08.02 |
| The forms stuff hasn't crashed yet. | 14:08.09 |
paulgardiner | Can't take the tension! :-) | 14:08.40 |
Robin_Watts | We still might want to consider updating the search so that it ensures that if there is a match on the page, that that match is at least partially on screen. | 14:08.40 |
| Don't know if that's an utter bitch to code though. | 14:08.53 |
paulgardiner | Yeah. I'll give that some thought. | 14:09.06 |
Robin_Watts | OK, fetching the transformer to try that now. | 14:09.31 |
ioo | hi all | 14:14.48 |
| i got i file i cannot convert | 14:14.57 |
| pdf file, don't know if it's a bug or a malformed file | 14:15.10 |
| a sample is here | 14:15.23 |
| http://www.wikiupload.com/VNWC70MSWWWDF8X | 14:15.29 |
| and the output from the conversion is here | 14:15.45 |
| http://pastebin.com/GFv28zJb | 14:15.52 |
| can you help? | 14:15.57 |
| it only happens with the pdfwrite device | 14:16.11 |
kens | MuPDF/GS ? versaion ? command line ? | 14:16.15 |
ioo | cli gs 9.06 | 14:16.29 |
| linux binaries from the site | 14:16.38 |
| 64 bit | 14:16.42 |
Robin_Watts | kens: command line and version are in the pastebin. | 14:16.48 |
ioo | ./gs-906-linux_x86_64 -dPDFSTOPONERROR -sDEVICE=pdfwrite -sOUTPUTFILE=out.pdf annotated.pdf | 14:16.57 |
| the file was a blank document exported from ooffice | 14:17.34 |
Robin_Watts | paulgardiner: seems fine on the transformer too. Woo Hoo! | 14:17.36 |
kens | 50% of the time I don't have to look if I ask the question :-) | 14:17.36 |
ioo | and the note was added with nitro pdf | 14:17.48 |
paulgardiner | Robin_Watts: brilliant | 14:17.50 |
| Robin_Watts: Oh yeah, forgot to mention: I managed to get EH to update their file manager. Now xps files open directly in MuPDF. Still working on OI | 14:19.41 |
Robin_Watts | paulgardiner: ES presumably ? Great news. | 14:20.37 |
paulgardiner | Yeah ES | 14:20.52 |
kens | ioo it works OK for me on 64-bit Windows with latest code (not latest release, but hEAD of Git) | 14:21.12 |
| I suspect it may have already been fixed | 14:21.44 |
| 9.05 gives me a typecheck in le | 14:22.42 |
ioo | kens: i'll try with the latest git | 14:22.56 |
kens | Let me know how you get on | 14:23.02 |
ioo | so, is (was) this a bug? | 14:23.30 |
kens | If it works now, but not before, then 'probably' | 14:23.52 |
Robin_Watts | paulgardiner: Something is still unhappy; I can't run under the emulator. | 14:26.54 |
paulgardiner | And ok without recycle calls? | 14:27.53 |
Robin_Watts | I see the file list, but as soon as I start test.pdf it returns to the file list again. Logging shows a SIGILL when loading the lib. | 14:27.55 |
| No idea. | 14:27.57 |
| I haven't used the emulator for ages, but I need to take some screenshots, so I thought "Aha!" :) | 14:30.12 |
henrys | US holiday today I'll be in and out. | 14:32.16 |
Robin_Watts | OK. recycle makes no odds to emulator. | 14:42.10 |
| So, the worry I have is that the emulator is Froyo. None of our other devices are that old, right? | 14:42.39 |
| My phone is 2.3.5 (Gingerbread) | 14:43.32 |
paulgardiner | I don't think I've used the emulator other than on your original version. | 14:47.07 |
| I wonder if the loading and storing of lib globals is problematic on the emulator | 14:48.23 |
Robin_Watts | paulgardiner: ooh, could be. | 14:52.01 |
| I am experimenting with the emulator set to gingerbread. | 14:52.30 |
paulgardiner | Just a wild guess. Don't be too hopeful | 14:52.32 |
Robin_Watts | ok. gingerbread emulator didn't help. | 14:55.14 |
fdncred | I'm looking for a tool to flatten all interactive content in a PDF. Not necessarily burn to raster. Just flatten. I have scripts that work in acrobat but i need a command line util or api. Any ideas? | 15:07.51 |
Robin_Watts | fdncred: What sort of thing needs flattening? | 15:13.14 |
fdncred | Hey Robin, I'm specifically talking about annotations, stamps, text boxes, lines (drawing stuff), comments | 15:15.01 |
| all form fields | 15:15.15 |
| most of the scripts call an acrobat function called flattenPages() that flattens "everything" | 15:16.05 |
Robin_Watts | so essentially, all the form fields/annotations become regular mage markings ? | 15:16.44 |
fdncred | that's what i'm looking to do specifically because most of the APIs don't work right when pages are split i.e. they puke on annotations or don't copy the annotations,e tc | 15:16.47 |
| yes | 15:16.54 |
Robin_Watts | Well, 2 ideas. | 15:17.29 |
fdncred | i have 3 pdfs that demonstrate "problem pdfs" that i could show you | 15:17.31 |
Robin_Watts | 1) Use "mutool clean" to do the splitting of PDFs. That should keep annotations etc correctly. | 15:18.00 |
| and thus avoid the need to flatten. | 15:18.09 |
| 2) Use gs with a pdfwrite device. That will render the annotations (at least the ones with appearance streams), and flatten forms data (as long as it has an appearance stream) | 15:18.52 |
fdncred | i'll look at mutool clean - i thought i tried that and it didn't work. perhaps i'm mistaken. all the flattening is only because splitting doesn't work properly so if i can split properly no need to flatten. thanks. | 15:22.11 |
Robin_Watts | If you can find a case where splitting with mutool clean fails, then please let us know! | 15:22.39 |
sebras | fdncred: if you are having problems with splitting pdfs using mutool clean then we'd like to hear about it. | 15:22.45 |
| Robin_Watts: tgmta. | 15:22.53 |
Robin_Watts | tgmta? | 15:24.50 |
sebras | Robin_Watts: two great minds think alike. | 15:26.49 |
Robin_Watts | There is a slashdot report that says that android emulator only supports armeabi, not armeabi-v7 or armeabi-v7a. | 15:26.52 |
| Ah! | 15:26.55 |
ioo | kens: with current git the error is gone | 15:30.15 |
| thank you | 15:30.23 |
kens | OK that's good news :-) | 15:30.28 |
| thanks for letting us now | 15:30.34 |
ioo | now, i am using gs in a server to automatically batch-convert some documents. How stable is the git tree? can i use it without problems? | 15:32.40 |
| or is it better to try to locate the commit that fixes the problem and try to backport it to the stable version? | 15:33.16 |
Robin_Watts | ioo: HEAD is fairly stable. | 15:34.35 |
| We have a release due in a week or two, so you can always hold on for that. | 15:34.49 |
kens | ioo in my opinion, stick with the code as is | 15:34.51 |
ioo | ok, thank you for your work | 15:35.50 |
kens | NP | 15:35.58 |
sebras | q | 15:44.52 |
| that was meant for gdb, not for you! | 15:45.02 |
Robin_Watts | Damn. Just died on me :( | 16:17.11 |
fdncred | Robin_Watts: i can't figure out how to split pages with mutool clean. Any tips? i tried "mutool clean <mypdfname> page-1.pdf 1-1" to split page 1 out. no dice. | 16:33.26 |
Robin_Watts | mutool clean in.pdf out.pdf 1 | 16:33.45 |
| that should work. | 16:33.50 |
fdncred | Robin_Watts: somethign wiggy is going on. original pdf=19 pages 6.3mb, single page output=6.4mb | 16:39.39 |
| maybe the other pages will be smaller | 16:40.08 |
| nope - every page is 6.464 mb | 16:40.57 |
Robin_Watts | fdncred: Right, you need to do -gggg to remove unused stuff. | 16:41.24 |
fdncred | ahhhh... trying now | 16:41.36 |
Robin_Watts | mutool clean -gggg in.pdf out.pdf 1 | 16:42.13 |
fdncred | so far so good. now trying my other types of interactive pdfs. sorry for being so obtuse. | 16:42.42 |
Robin_Watts | no worries. | 16:42.50 |
paulgardiner | Robin_Watts: pushed | 16:45.39 |
Robin_Watts | paulgardiner: Could you review http://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=d8f4cf7e73bd18a665887c48c88fb18cb99c4945 please? | 16:46.30 |
| tor8: If you have opinions on that, please speak up too. | 16:46.40 |
paulgardiner | sure | 16:46.48 |
tor8 | Robin_Watts: if it's the same as this morning, LGTM | 16:47.14 |
Robin_Watts | Oh, you already looked at the rename review? | 16:47.30 |
| Yes, no changes. | 16:47.37 |
| Thanks. | 16:47.42 |
paulgardiner | How come git hasn't worked out these are renames rather than deletions and creations. | 16:49.00 |
| ? | 16:49.14 |
Robin_Watts | Has it not ? | 16:49.21 |
paulgardiner | Maybe it has. Not sure how to read the changes | 16:50.14 |
| There is a git-mv I think | 16:50.39 |
Robin_Watts | It must have done, because it's correctly got all your changes into the moved files. | 16:50.51 |
paulgardiner | Oh ok | 16:51.11 |
tor8 | Robin_Watts: it "detects" git moves and file renames by looking at file similarity between commits. there's a command line flag to git diff that does the detection, I'm not sure why it doesn't do it by default. | 16:51.51 |
Robin_Watts | having said that, blame won't work through this review :( | 16:51.58 |
| tor8: I thought it had detected the fact they were moves. | 16:52.18 |
tor8 | Robin_Watts: git blame -M | 16:52.19 |
Robin_Watts | ok, git blame -M works. | 16:53.13 |
| Shame it doesn't offer a 'blame' link in the git web view, but hey... | 16:53.27 |
paulgardiner | Sounds like that's ok then, and besides that it looks fine. | 16:53.44 |
Robin_Watts | Thanks. | 16:53.51 |
kens | [For the logs] mvrhel_laptop Is there an appropriate CMS routine to convert a buffer full of data, rather than a sample at a time ? If os, what is it please ? I'm looking at images and would like to handle a scan line at a time. | 17:03.28 |
Robin_Watts | kens: There is. | 17:08.30 |
kens | Robin_Watts : I rather thought there would be ;-) | 17:08.55 |
Robin_Watts | gscms_transform_color_buffer | 17:09.02 |
| I'm guessing you've been using gscms_transform_color ? | 17:09.13 |
kens | I'll loook at it tomorrow, thanks | 17:09.43 |
| I've been using icc_link->procs.map_color | 17:10.03 |
| OK I'm heading off now, goodnight all | 17:20.29 |
Robin_Watts | paulgardiner: Yeah, still broken, still caused by recycle. | 17:31.06 |
apineda | Just wondering what happens when a pdf has a spot color with a secondary CMYK device color using a gs device like tiff32, color managemen is on by default, but what working space profile is used, and what output profile is used? | 17:50.57 |
| nm I googled | 19:17.12 |
| like a good chap should | 19:17.15 |
sebras | Robin_Watts: oh, so you didn't get access to com.artifex.mupdf at google play? | 20:00.53 |
Robin_Watts | sebras: No :( | 20:05.27 |
sebras | Robin_Watts: oh, why? did they argue that the string used is not trademarkable or what? | 20:07.39 |
Robin_Watts | They didn't argue anything. | 20:07.50 |
sebras | ok. | 20:07.55 |
Robin_Watts | They removed the offending app, but we still can;t upload to com.artifex.mupdf. | 20:08.06 |
| I suspect they apply their 'once a name is used, it's gone forever' rule. | 20:08.26 |
sebras | right. | 20:08.33 |
| or it could be a matter of time. | 20:08.42 |
Robin_Watts | could be. | 20:08.48 |
| mvrhel_laptop: kens asked you a question earlier. | 20:37.58 |
| deleet: ping | 20:38.26 |
mvrhel_laptop | oh hi Robin_Watts | 21:12.48 |
Robin_Watts | deleet: Can you test http://ghostscript.com/~robin/MuPDF.apk please? | 21:13.00 |
| We've had to rework the recycle stuff to avoid crashes. | 21:13.08 |
mvrhel_laptop | Robin_Watts: looks like you answered the question for him | 21:13.38 |
Robin_Watts | In particular we DO NOT call recycle on the old bitmap before allocating the new bitmap. This is because the old bitmap is still being used. | 21:13.38 |
mvrhel_laptop | thanks | 21:13.39 |
Robin_Watts | mvrhel_laptop: So my answer was right? | 21:13.48 |
mvrhel_laptop | yes | 21:13.51 |
Robin_Watts | I suspect his next question will be "but I'm calling icc_link->procs.map_color, how does that relate to gscms_transform_color_buffer" ? | 21:14.24 |
| Oh, he wants to call icc_link->procs.map_buffer | 21:15.13 |
| kens: (For the logs): You want to call icc_link->procs.map_buffer | 21:15.30 |
mvrhel_laptop | yes | 21:15.35 |
Robin_Watts | :) | 21:15.39 |
mvrhel_laptop | I thought I had done both | 21:15.53 |
Robin_Watts | You may well have done. he said he hadn't looked hard. | 21:16.17 |
mvrhel_laptop | that way we can have different CMMs handle stuff (for the case when CM is off or for the odd cases for customer 330) | 21:16.18 |
Robin_Watts | I think I am now in a position where I can hit 'publish' and we'll be on the android app store. | 21:16.58 |
| mvrhel_laptop: Do you have a Nexus 10 ? | 21:24.59 |
mvrhel_laptop | Robin_Watts: I do | 21:25.48 |
Robin_Watts | Could you test http://ghostscript.com/~robin/MuPDF.apk please? | 21:26.07 |
mvrhel_laptop | oh ok. hold on a couple minutes | 21:27.19 |
Robin_Watts | actually, let me upload a new version. I'll say when it's there... | 21:27.33 |
mvrhel_laptop | ok | 21:27.37 |
Robin_Watts | mvrhel_laptop: New version there now. I'm off for the night, so no hurry! | 21:41.41 |
| We've had reports (well, a report) that on hi-res tablets mupdf can run out of memory due to not being able to get enough bitmaps. | 21:42.08 |
| and I'd like to know if you can see that. | 21:42.21 |
mvrhel_laptop | Robin_Watts: ok I will beat on it a some tonight | 21:42.26 |
Robin_Watts | (basically, does it die after a while of panning/zooming etc). | 21:42.39 |
| Thanks. | 21:42.40 |
mvrhel_laptop | ah ok | 21:42.44 |
| have a good night | 21:42.50 |
deleet | Robin_Watts: hey, got your mail | 22:34.45 |
| will give it a try sure | 22:34.56 |
Robin_Watts | deleet: Thanks | 23:31.29 |
| Forward 1 day (to 2013/01/22)>>> | |