| | <<<Back 1 day (to 2013/12/20) | 2013/12/21 |
| veek | has anyone used gv for printing on windows?? I really hate the GSview print dialog box.. gv is beautiful in comparison | 09:20.02 |
| | it's a major pain trying to print a single page, you have to type the page number vs just scrolling and right clicking like in gv | 09:20.42 |
| | i can't use linux because i have a oldish laserprinter and the linux driver for some reason prints very slow whereas the GDI driver in windows is super fast | 09:21.43 |
| | though it's a postscript printer | 09:21.56 |
| Robin_Watts | veek: gsview is not written by us. | 09:22.15 |
| veek | yup, is there a gv port for windows? | 09:22.35 |
| Robin_Watts | I am not aware of such. | 09:23.21 |
| henrys | Robin_Watts: so we have everything including writing/saving? | 16:01.21 |
| Robin_Watts | henrys: everything including writing/saving, apps, graphics libraries, UI libraries, font libraries... | 16:23.19 |
| | essentially everything they ever developed. | 16:23.30 |
| | I'm still reeling :) | 16:23.39 |
| henrys | Robin_Watts: so we really need to think about what it means to GPL this? Is someone like Apache or some other army going to come around and fork the code. We won't be able to keep up with fixes resource wise. | 16:24.56 |
| Robin_Watts | I think we should think very carefully before GPLing the whole lot. | 16:29.50 |
| | I think we should carry on with the idea of carving out just the bits we need for MuPDF to read office docs to start with. | 16:30.27 |
| henrys | Robin_Watts: I'll wait until I see writing⦠this does seem a bit unlikely to me. | 16:34.12 |
| | Robin_Watts: like somebody is going to burst in the room and say wtf are you doing? | 16:37.48 |
| Robin_Watts | :) | 16:38.04 |
| henrys | Robin_Watts: how cares about MuPDF? We could gpl the viewing pieces and sell editing/saving technology and be players in the business, if we are getting everything. | 16:49.58 |
| Robin_Watts | possibly. I still need to get my head around this. | 16:51.07 |
| | This is like leaning on a door and finding it was push rather than pull. | 16:51.27 |
| mvrhel_laptop | wow. reading between the lines in your comments henrys | 17:03.09 |
| henrys | mvrhel_laptop: just got off the phone with miles | 17:13.14 |
| | just giddy about this (both of us) | 17:13.51 |
| mvrhel_laptop | nice | 17:14.11 |
| henrys | Robin_Watts: if you don't want to do the list I can probably talk Miles out of it. It seems reasonably clear to me we are getting everything. | 17:18.38 |
| Robin_Watts | henrys: I can certainly do a list of bullet pointed things that we are getting. | 17:18.58 |
| henrys | Robin_Watts: right he asked you to do that in email and I'm just saying I can probably talk him out of requiring that. | 17:20.57 |
| | Robin_Watts: assuming list writing is not a fun thing to do on Saturday ;-) | 17:21.26 |
| Robin_Watts | No, I meant that I am happy to a do a list of things that we should be getting. I would hate for me to have to capture everything with the list though. | 17:22.18 |
| | i.e. if we're saying "We are getting everything, including but not limited to ...." then I'm good. | 17:22.53 |
| | Helen is out doing a concert, so I'm going to take a run at the list shortly. | 17:23.58 |
| mvrhel_laptop | this is like christmas | 17:24.13 |
| henrys | That is how he said it would be worded yes, it does concern me a bit. I'd rather just say everything and audit "everything expected" is in the deliverable but the list is Miles' preference | 17:24.26 |
| mvrhel_laptop | sounds like the kids making their wish list | 17:24.33 |
| functorfun | hi | 17:25.13 |
| ghostbot | niihau | 17:25.13 |
| henrys | I have some errands to do, be back in a while | 17:26.17 |
| functorfun | https://gist.github.com/victoredwardocallaghan/8072322 I have issues with CUPS printing a document, I traced it down to the 'gs' filter core dumping here. | 17:27.29 |
| Docman | hi everyone, bots and non bots | 17:27.48 |
| functorfun | unfortantly the gs binary is stripped so no debug symbols :/ | 17:28.23 |
| Docman | anyone free to bounce a question off about ghostscript? | 17:28.24 |
| functorfun | Docman: Can try.. | 17:28.35 |
| Robin_Watts | functorfun: Why not build a debug gs binary ? | 17:29.19 |
| henrys | functorfun:reproduce it with a ghostscript command line and report a bug at bugs.ghostscript.com, you may not get a quick answer ⦠holiday and all | 17:29.28 |
| Docman | shouldnt be a problem, I just wanted to know if there was any way to change the image type (from png/jpg) or resolution in the rendering process | 17:29.32 |
| Robin_Watts | Docman: What command are you using? | 17:29.52 |
| henrys | functorfun: you can also report it to your distribution and then they report the ghostscript problem to us. | 17:31.42 |
| Docman | yikes, to be honest I'm just using it through one of these HTML5 / PDF converter tools. So I haven't explored the command line much | 17:32.28 |
| | is the output image type available via CLI? | 17:32.32 |
| henrys | functorfun: we need a ghostscript command line to start from we don't generally debug anything until we have that. | 17:32.55 |
| functorfun | henrys: yep, well its in the JPXDecode which I guess is a Haar wavelet implementation? | 17:34.43 |
| Robin_Watts | Docman: It certainly is. | 17:34.49 |
| functorfun | henrys: I have a pdf to reproduce with | 17:34.58 |
| henrys | functorfun: okay does gs PDFFILE from the command line reproduce the problem? | 17:36.20 |
| functorfun | interestingly no.. | 17:39.29 |
| | I did gs -o repaired.pdf -sDEVICE=pdfwrite -dPDFSETTINGS=/prepress corrupted.pdf and I am able to print the repaired.pdf | 17:39.44 |
| henrys | functorfun: perhaps the device color space from cups introduces the issue. What printer is it? | 17:41.25 |
| functorfun | I found that from here.. http://ghostscript.com/pipermail/gs-bugs/2013-August/032830.html what exactly does that do? | 17:41.29 |
| | oh its a HP LaserJet Pro 400 MFP M425dn | 17:41.54 |
| henrys | functorfun: of course it could be a fixed problem we have corrected a number problems with jpeg2000 is gs the same version on the command line? | 17:42.14 |
| functorfun | details here http://hplipopensource.com/hplip-web/models/laserjet/hp_laserjet_400_mfp_m425dn.html | 17:42.17 |
| | GPL Ghostscript 9.10 (2013-08-30) | 17:42.40 |
| henrys | functorfun: and you're sure cups is using the same? | 17:43.03 |
| functorfun | I have a fairly minimal system.. | 17:43.30 |
| | as far as I know.. I don't believe there is more than one binary on my system | 17:43.48 |
| | would you like the complete cups error log? | 17:44.07 |
| henrys | functorfun: I think a good next step is to report to the distribution, but I'll look at the error log if you post it. | 17:45.05 |
| functorfun | on it.. | 17:46.39 |
| | https://gist.github.com/victoredwardocallaghan/8072512 | 17:46.40 |
| henrys | tkamppeter__: does cups use a special ghostscript version or does it just use the exe a user would get running from the command line? | 17:46.45 |
| functorfun | henrys: i'll move /usr/bin/gs to /usr/bin/gs.foo for a sec and find out.. | 17:48.18 |
| | D [22/Dec/2013:04:48:45 +1100] [Job 251] PID 7561 (gs) stopped with status 1! | 17:49.08 |
| | henrys: ^ its using the normal system binary | 17:49.17 |
| henrys | functorfun: geez it would be nice if they gave you the command line for gs in the log | 17:50.51 |
| functorfun | henrys: I know :< I wacked the debug and whatnot flags up to full in the cupsd.conf file.. Trying to be a good 'user' here ;) | 17:52.21 |
| | I don't mind 'fixing it' myself if I knew what I was fixing.. | 17:52.57 |
| | why would ` gs -o repaired.pdf -sDEVICE=pdfwrite -dPDFSETTINGS=/prepress corrupted.pdf ` 'fix' the pdf as so that it can be printed? | 17:53.31 |
| Robin_Watts | functorfun: That decompresses and recompressed the images. | 17:54.11 |
| henrys | functorfun: I believe we would discard the jpeg2000 | 17:54.13 |
| Robin_Watts | We can't recompress the jpeg2000, so it will be recoded as normal jpeg. | 17:54.34 |
| henrys | what does gs -sDEVICE=ppmraw -o foo.ppm PDFFILE do? | 17:54.45 |
| functorfun | Well I know/or worked out JPXDecode is prob a Haar wavelet implementation of some-kind.. | 17:54.53 |
| henrys | with -r300 | 17:54.56 |
| Robin_Watts | As to why the decompression of the jpeg2000 would fail when going to cups, but not when being decompressed for recompression, I am not sure. | 17:55.27 |
| henrys | Robin_Watts: but it should have failed decoding it. | 17:55.33 |
| Robin_Watts | Yes, JPXDecode is haar wavelet based. | 17:55.38 |
| functorfun | gs -sDEVICE=ppmraw -o foo.ppm PDF.. is cooking along.. | 17:56.17 |
| | it has a few " **** ERROR: Unable to process JPXDecode data. Page will be missing data " around but continues.. | 17:57.15 |
| henrys | functorfun: of course that could be a red herring and have nothing to do with the crash | 17:58.35 |
| functorfun | -r300 makes it grow to a insane size but does continue.. | 17:59.14 |
| henrys | functorfun: I assume you've been printing merrily to this device until this file. | 17:59.18 |
| functorfun | yes | 17:59.26 |
| | well it does claim (gs) exit with sig 11 | 18:00.03 |
| | so that is a seg fault | 18:00.34 |
| henrys | functorfun: beautiful report it to bugs.ghostscript.com and we'll have a look command line and file | 18:00.36 |
| functorfun | Many thanks | 18:01.13 |
| henrys | we don't need the output ppm rater. | 18:01.15 |
| functorfun | I'll just upload the pdf, what I was doing and the chat log here? | 18:02.42 |
| henrys | the pdf and the command line | 18:03.20 |
| functorfun | from cups, ok then I am going to have to figure that out.. since if I knew that then I would compile ghostscript with symbols and get a backtrace | 18:04.56 |
| | then you would have a bt at worst or patch at best :p | 18:05.28 |
| | #cups is dead though :/ | 18:06.02 |
| henrys | functorfun: you reproduced it with ppmraw - I meant that command line | 18:06.43 |
| Robin_Watts | functorfun: No, you just said that you ran it with ppm... what henrys said. | 18:06.53 |
| functorfun | Ah, but that does not make gs sig 11 seg fault, it just reports that JPXDecode had issues and continues | 18:07.59 |
| | sorry. | 18:08.05 |
| | Are any signals past back to the caller when that "ERROR" is encountered.. Because it could be a bug in the wrapper filter from cups and not actually in the gs binary? | 18:09.35 |
| | Maybe its miss-handled or not handled at all? | 18:09.58 |
| henrys | functorfun: oh I misunderstood. | 18:10.17 |
| functorfun | sorry not to be clear. | 18:11.29 |
| henrys | functorfun: you can try the various cups devices -- gs --help, but I imagine a complete command line from the distribution folks would speed this along. | 18:12.09 |
| functorfun | I think its _more_ probable that there is a bug in the gs wrapper/filter cups uses perhaps? | 18:12.10 |
| henrys | functorfun: maybe | 18:12.32 |
| functorfun | the distro is archlinux so its not really much of a 'distro' if you know what I mean | 18:12.45 |
| | Ah ha, if I run pdf2ps PDFFILE it blows up | 18:14.02 |
| | is pdf2ps part of the gs source? | 18:14.24 |
| henrys | functorfun: is that a script or an exe? | 18:15.15 |
| | functorfun: should be our code | 18:16.18 |
| functorfun | good point.. I'll paste it, its a script | 18:16.51 |
| | https://gist.github.com/victoredwardocallaghan/8072840 | 18:17.06 |
| henrys | so it fails with the ps2write device and has nothing to do with cups. You can now report it ⦠say gs/lib/pdf2ps in the bugs report so there is no confusion | 18:17.51 |
| functorfun | ok, good thing we traced that last bit down.. Just waiting on a email for a account | 18:18.43 |
| | wish there was not a million bug tracker accounts O_o | 18:19.03 |
| henrys | finally off for errands bbiaw | 18:24.19 |
| functorfun | henrys: many thanks for all the advice | 18:24.37 |
| wart___ | Hi folks. I was reading the mupdf news, and it states that initial annotation support is available for "library" and android version in 1.3. I'm running mupdf under Linux, and I was wondering what "library" meant? Is it available on Linux? | 18:24.40 |
| | Sorry: library and android builds only. So question: What is a library build? | 18:25.11 |
| Robin_Watts | wart___: MuPDF is a C library for manipulating/rendering PDF files. | 18:25.34 |
| | It has a C level API for opening files, rendering them, annotating them etc. | 18:25.54 |
| | The example viewers we offer for different platforms are thin veneers over this. | 18:26.08 |
| | i.e. the viewers are as much as possible UI, with all the 'smarts' in the library. | 18:26.26 |
| functorfun | A library is effectively a clump of code that is a common function(s) between applications | 18:26.41 |
| Robin_Watts | The android viewer has the required UI stuff written to hook up to the new annotation functions. | 18:27.06 |
| wart___ | Robin_Watts: Right, so the Linux binary should then have annotation support in 1.3, if I understand correctly? I build mupdf via a gentoo ebuild, but I can compile it locally too, if needs be. | 18:27.06 |
| Robin_Watts | but the linux viewer doesn't have the required UI to do annotations. | 18:27.30 |
| wart___ | Sorry, I didn't mean what is a library in general (I get the concept) but I was unclear what the NEWS meant with "library and android build only" :-) | 18:27.39 |
| Robin_Watts | i.e. the core of MuPDF is capable of it, but the linux viewer won't do it. | 18:27.54 |
| | Does that make sense? | 18:28.02 |
| wart___ | Robin_Watts: Ah, ok. Thanks! | 18:28.07 |
| Robin_Watts | If you want to update the linux viewer, then we'd be very interested in a patch. | 18:28.24 |
| functorfun | wart___: the library implements _the_ functions and the gui allows you to _use_ the functions, basically. | 18:28.31 |
| | some gui's may have more 'expressed' function usage than others | 18:28.54 |
| wart___ | Robin_Watts: It's been on my wishlist for a while, since I don't use KDE or GNOME, so I really have no annotation option available (and I love mupdf) | 18:29.08 |
| functorfun | hence, some applications require you to update a library to pull in the extra functionality found in a newer library revision. | 18:29.26 |
| wart___ | Is there someone assigned to the linux viewer at present? | 18:29.59 |
| functorfun | hope I explained it ok? | 18:30.07 |
| Robin_Watts | wart___: The mupdf devs are all busy elsewhere for the short term. | 18:30.38 |
| wart___ | Robin_Watts: It's surprising to me that the linux community hasn't rallied around the idea by now. I guess everyone's focused on okular and the gnome pdf reader (evince) | 18:31.55 |
| | And I wish I had a second life to dedicate some code time :-) | 18:32.18 |
| functorfun | wart___: have you looked at zathura or xournal ? | 18:33.34 |
| | wart___: I use mupdf also, but they are useful for what you are asking | 18:33.51 |
| wart___ | functorfun: Yeah, I used to submit patches to the zathura folk. But iirc they don't implement annotations and they decided to make a gtk dep that I didn't like a couple years back. xournal, I think, doesn't annotate onto the pdf, last I checked, and I think it has funny dependencies. | 18:34.28 |
| | maybe I'll check back in with zathura, tho'. It has been a couple years. | 18:34.51 |
| functorfun | no idea.. just a idea | 18:35.05 |
| wart___ | functorfun: Thanks! | 18:35.13 |
| functorfun | no worries | 18:35.23 |
| tkamppeter | henrys, CUPS uses the standard gs executable (/usr/bin/gs in most Linux distributions), nothing special, no special fork, (at least in Ubuntu) patches are only upstream backports, nothing made especially for CUPS or the distro. | 21:01.27 |
| henrys | tkamppeter: thanks | 21:01.52 |
| tkamppeter | henrys, usually the gs calls are done by filters in the cups-filters package, I am not sure, but CUPS itself should not have any gs calls any more. | 21:02.39 |
| henrys | tkamppeter: yes in the op's case we tracked it back to pdf2ps | 21:03.22 |
| tkamppeter | henrys, filters calling it are pstopdf, pdftops, gstoraster, foomatic-rip, gstopxl, all in cups-filters. | 21:03.38 |
| | henrys, pdf2ps or pdftops? | 21:04.07 |
| henrys | well gs was crashing - so | 21:04.37 |
| | pdftops is poppler right? | 21:04.50 |
| | tkamppeter: ^^^ | 21:11.14 |
| tkamppeter | henrys, there are two legacy filters, simple wrapper scripts, names imagetops and texttops, only used when explicitly mentioned in the users individual config or PPD, never by the default configuration of CUPS or cups-filters. | 21:29.43 |
| | /usr/bin/pdftops is Poppler, the CUPS filter /usr/lib/cups/filter/pdftops (from cups-filters) allows the user/admin to use either Poppler or GS. | 21:30.49 |
| | henrys, ^^ | 21:31.00 |
| | henrys, pdf2ps is only called by the two legacy filters of cups-filters, by none of the standard filters of cups-filters and CUPS. | 21:33.04 |
| henrys | tkamppeter: oh interesting - so maybe cups wasn't using pdf2ps but nonetheless it was clear gs was was being used and the evidence points to a bug in ps2write or do you think something else was going on? | 21:40.21 |
| tkamppeter | henrys, it seems to be a bug in ps2write, so you should work on fixing it, but I cannot help much as the OP has a non-standard configuration which makes the legacy fallbacks getting triggered and therefore I cannot reproduce it with my system in standard configuration. I would need the OP's config. | 21:50.49 |
| | henrys, so the best solution is if you work with the OP to find out the GS command line and input file exactly used. Having hime following the instructions of https://wiki.ubuntu.com/DebuggingPrintingProblems could help, at least what he can do with his distro. | 21:52.09 |
| | henrys, it is perhaps also possible that he has an old system still not using cups-filters, but then I have no idea how pdf2ps gets called. | 21:53.13 |
| henrys | tkamppeter: thanks I'll add that link to my notes and point folks to it in the future. | 22:03.34 |
| Robin_Watts | henrys: ping | 23:09.59 |
| henrys | Robin_Watts: yes | 23:16.40 |
| Robin_Watts | The last mail I got from Darin said he hadn't got my list. | 23:17.01 |
| | I was expecting Miles to forward it on. | 23:17.11 |
| | (after checking he was happy etc) | 23:17.19 |
| | so I'm having visions of Darin sitting in his office on a saturday twiddling his thumbs and charging by the hour :) | 23:17.44 |
| | I've texted Miles but had no reply. | 23:18.08 |
| henrys | Robin_Watts: I wouldn't worry about Darin isn't like that. | 23:18.37 |
| | Robin_Watts: he just forward it, see your email | 23:19.03 |
| | Robin_Watts: heading out to shop some more | 23:19.27 |
| Robin_Watts | henrys: got it. Thanks. Have fun with your "shopping". | 23:19.48 |
| | I have heard tell of this, but only involving a web browser and a mouse... | 23:20.03 |
| sebras | hm.. I wonder if we should consider invalid page destinations as errors. | 23:20.06 |
| | I'm trying to show a pdf and get a long list of "error: invalid page object" at startup. | 23:20.32 |
| | it' is a due to loading the outline which tries to parse each action which in turn tries to resolve each destination. | 23:21.20 |
| | Forward 1 day (to 2013/12/22)>>> | |
IRC Logs