| <<<Back 1 day (to 2015/03/31) | 20150401 |
KunMyt | i'm sory, i have a question for MuPDF on android. When i'm open pdf file with ContentProvider (success), why function as draw, highlight, underline, strike mode not appear. tsk | 02:03.43 |
mvrhel_laptop | ok adding -I as an argument fixed my issues | 03:52.02 |
| henrys: ok I think I am finally to a beta point | 04:28.38 |
| barring some issue that someone has or finds | 04:28.48 |
| seems to be pretty stable on my windows 8 64 bit machine and my window 7 32 bit machine | 04:29.38 |
| hmm and of course I find one small issue that I want to fix already... | 04:36.16 |
| ok. fixed the issue, made a new installer and pushed it to casper. | 05:02.54 |
| I am going to have to figure out a way to automate some testing for this | 05:03.20 |
| done for the night | 05:14.18 |
KunMyt | Hi all, i have question: how to save pdf after edit it when read file from ContentProvider using mupdf for Android. | 08:38.32 |
tor8 | Robin_Watts: your 2 commits LGTM | 09:36.54 |
| fix for the MinGW build on tor/mingw | 09:37.04 |
| and updated mudraw commit on tor/master (with manpage changes) | 09:37.25 |
Robin_Watts | tor8: cool. will look now. | 09:37.42 |
| I was trying to have a look at 694367, but my first attempts at it have not been encouraging. | 09:38.06 |
| The lines that disappear are filled rectangles, not strokes. | 09:38.34 |
tor8 | Robin_Watts: yeah. I have a nasty suspicion we should rewrite the rasterizer to fix that one properly :( | 09:38.44 |
| scan converter* | 09:39.06 |
| it might be hackable to force features to be at least 1 pixel wide/high | 09:39.56 |
| forcing them to be taller is tricky | 09:40.07 |
| but ideally we should use some other pixel filling rule | 09:40.16 |
Robin_Watts | to be at least 1/fz_aa_vscale pixels high. | 09:46.59 |
| tor8: Does curl not build under mingw or something? | 09:53.54 |
tor8 | our submodule doesn't build ... it tries to include amigaos.h and sys/socket.h | 09:54.47 |
| I'm sure it's solvable with some configuration hacking, but I don't think it's worth the effort | 09:55.15 |
Robin_Watts | fair enough. | 10:01.30 |
| both look fine to me then. | 10:01.37 |
| I'll rebase and push them. | 10:02.30 |
| 1 more on robin/master | 10:03.26 |
tor8 | LGTM | 10:06.04 |
Robin_Watts | Ta. | 10:12.25 |
| tor8: bug 695660 points out that the man page for mutool is missing documentation of flags etc. | 10:15.38 |
| yeah mutool info has no docs in the man page. It just says "TODO" | 10:17.13 |
tor8 | it's also missing some new flags | 10:18.41 |
Robin_Watts | I shall leave bugs 696656, 696657, 696659, 696660 to you then. | 10:19.35 |
tor8 | Robin_Watts: the strstr(output, "%d") tests in mudraw for PWG and PCL output (that set append=0) won't work if you use %03d to get zero-padded page numbers | 10:29.46 |
Robin_Watts | tor8: Indeed. | 10:29.56 |
| Maybe a smarter test? | 10:30.22 |
tor8 | if only we had regexec() :) | 10:30.34 |
| a test to look for %[0-9]*d should be trivial to code | 10:31.03 |
Robin_Watts | exactly. | 10:31.09 |
tor8 | Robin_Watts: %d scan fix and more manpage updates on tor/master | 10:52.19 |
Robin_Watts | both lgtm. | 10:54.20 |
tor8 | Ta | 10:54.42 |
Robin_Watts | http://bugs.ghostscript.com/show_bug.cgi?id=695317 ? | 11:01.15 |
| oh, crap. | 11:01.55 |
| the cluster calls mudraw -d. | 11:02.04 |
tor8 | oh, crap. | 11:02.28 |
| can you fix? | 11:02.31 |
| I have a feeling that there's more keybindings in mupdf than that one missing. | 11:03.37 |
Robin_Watts | will try. | 11:04.27 |
tor8 | Robin_Watts: have you looked at http://bugs.ghostscript.com/show_bug.cgi?id=695321 ? it sounds familiar, so I wonder if we've already fixed it | 11:04.32 |
Robin_Watts | tor8: The bug I just closed as "WORKSFORME"? | 11:06.35 |
| :) | 11:06.36 |
tor8 | ah :) | 11:06.42 |
Robin_Watts | tor8: OK, I think the cluster is fixed now. I'm off for a run while it runs through. | 11:31.12 |
inkbottle | Hi, I've extracted and uncompressed a page of an pdf document, containing images, that using qpdf | 13:19.52 |
| know I'd like to extract one stream corresponding to an image | 13:20.39 |
| preferably using a simple text editor | 13:20.57 |
kens | You'll need a binary editor, an image will be stored as binsry | 13:21.20 |
inkbottle | problem is, what I see doesn't look like an image at all | 13:21.33 |
| hi | 13:21.40 |
ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 13:21.40 |
Robin_Watts | inkbottle: I would try using mupdf, specifically mutool. | 13:21.57 |
| Then mutool clean -difggg in.pdf out.pdf 1 | 13:22.14 |
| (where 1 is the page number you are interested in) | 13:22.24 |
| That will leave out.pdf being the appropriate page from in.pdf with the content stream decompressed so it's as easy to read in an editor as you are going to get. | 13:22.52 |
| Then you can use mutool extract to get the actual images. | 13:23.15 |
inkbottle | kens: after the page extraction and the decompression, I piped it through hd, and the result was strange for an image: http://paste.debian.net/164288/ | 13:24.14 |
| Robin_Watts: I'm reading you | 13:24.39 |
kens | I'm afraid 'hd' means nothign to me | 13:25.09 |
inkbottle | Robin_Watts: ok, I try that | 13:25.16 |
Robin_Watts | will be back in 40 mins or so. | 13:25.36 |
kens | hd appearsw to be ahard drive mapping tool, I can't see how that's relevant to a PDF steram, so it must be something else. | 13:26.10 |
chrisl | hexdump, hd â ASCII, decimal, hexadecimal, octal dump | 13:30.37 |
kens | And what format does it normally dump the hex in ? The paste is hex, just decorated | 13:31.22 |
inkbottle | kens: on my computer it does this: http://paste.debian.net/164290/ | 13:31.48 |
kens | Well that doesn't look *anything* like inkbottle's output | 13:32.08 |
chrisl | inkbottle: what were you expecting it to look like? | 13:33.16 |
inkbottle | kens: sorry my pastebin was just about hd | 13:33.43 |
kens | It doesn't look like the output chrisl gets from hd | 13:34.01 |
chrisl | I didn't post any output? | 13:34.27 |
kens | Your paste | 13:34.34 |
chrisl | Just a description | 13:34.49 |
kens | oops sorry | 13:34.52 |
inkbottle | the stream I had from qpdf was actually: | 13:34.53 |
| stream | 13:34.54 |
| <F0><FF><FF><FC><EE><E5><F9><FF><FF><FF><FF><FA><E1><CC><E5><FF><FF><FF><FD><FC>ܽ<D4><FB><FC><F9><FD><FF><E9><CC><D8><F8><FA><F8><FF><FF><F7><EA><EA><F6><FF><FE><FF><FD><FC><FF><F7><F7><FF><FE><FF><FC><FE><FF><FB><FA><FE><F7><FF><FE><FF><FF><F9><FD>endstream | 13:34.56 |
kens | Got you 2 confused there | 13:34.57 |
chrisl | inkbottle: so what's wrong with that? | 13:35.10 |
kens | Well that'se qpdf for you possibly. We don't deal with qpdf here so can't help you | 13:35.14 |
inkbottle | kens: yead I'm presently trying the solution with mupdf, above | 13:35.57 |
kens | I suspect MuPDF will get you something closer to what you want | 13:36.14 |
chrisl | Depends what you want..... which you haven't actually stated | 13:36.57 |
inkbottle | chrisl: you are write, that seems to be actually an image | 13:37.22 |
tor8 | inkbottle: mutool extract will extract the image (and re-encode it as png if necessary) | 13:38.01 |
chrisl | Well, it is encoded binary data, from a PDF stream, which could well be an image..... or something else entirely...... | 13:38.25 |
kens | Its binary data , it *could* be an image, or it could be the data for a function. Can't tell without seeing the PDF | 13:38.29 |
tor8 | inkbottle: if you just want the raw stream data, "mutool show -be file.pdf <objectnumber>" | 13:38.34 |
inkbottle | however I'll continue with mupdf and see what it gives | 13:38.52 |
| chrisl: what I'm trying to do is the following: I have a pdf file which doesn't show neither with okular nor with mupdf. but it does with chromium. It yields with okular errors about jpx stream... what I want to do is try and see what yields the error | 13:43.07 |
kens | If you try it with GS and -dPDFDEBUG -dPDFSTOPONERROR it'll tell you what image is causing the problem | 13:44.05 |
chrisl | Hmm, debugging JPX..... good luck!! | 13:46.03 |
inkbottle | chrisl: thanks | 13:47.22 |
| chrisl: imagemagick renders the document correctly too (so there may be hope) | 13:48.50 |
chrisl | That's Ghostscript, then..... | 13:49.13 |
kens | Which has a modified OpenJPEG | 13:49.34 |
chrisl | As does mupdf, which is a little odd.... | 13:49.56 |
kens | Yeah, do we use the same OpenJPEG ? | 13:50.09 |
| inkbottle : can you make the PDF public ? | 13:50.22 |
chrisl | kens: not sure..... | 13:50.44 |
inkbottle | kens: I could tell you where I found it... (not sure...) | 13:51.26 |
kens | If you can, then we could poke it here | 13:51.40 |
inkbottle | chrisl: kens: "bookzz.org/s/?q=light+on+yoga&t=0"; first file in the list, the one with "Dipika" | 13:59.11 |
kens | OK let me look | 13:59.26 |
| Which page has a problem ? | 14:00.18 |
inkbottle | I used to be able to see it with okular, then some six month ago, no images anymore; use chromium to view it as it should be | 14:00.49 |
kens | Acrobat is happy with it | 14:01.00 |
inkbottle | p. 66, | 14:01.04 |
| eg | 14:01.07 |
| but basically most pages | 14:01.19 |
kens | Let me go to page 66 | 14:01.27 |
| MuPDF 1.6 displays it fine for me | 14:01.42 |
| As does Acrobat | 14:01.48 |
inkbottle | kens: ok | 14:02.28 |
| ... | 14:02.30 |
| kens: so you can see the yoga psture all right? | 14:03.06 |
kens | Ghostscript is happy with it too | 14:03.11 |
| Yes, yoga posture and text | 14:03.21 |
inkbottle | so it must be a problem with my installation | 14:03.40 |
kens | 2 black and white pictures on page labelled 65 | 14:03.48 |
inkbottle | I gives you the errors it gives | 14:03.50 |
chrisl | Where did you get the mupdf install? | 14:04.03 |
kens | Text begins 'Technique 1. Stand in Tadasana. (Plate 1.)' Obviously I can't do the accents from my keyboard | 14:04.46 |
inkbottle | the errors: http://paste.debian.net/164294/ | 14:05.06 |
| chrisl: kens: it's debian | 14:05.17 |
| jessie | 14:05.20 |
chrisl | Those aren't mupdf errors..... | 14:05.37 |
kens | OK, but where did you get MuPDF (and what version) ? | 14:05.38 |
inkbottle | there shouldn't be anything special in my installation | 14:05.51 |
kens | Its possible that these are OpenJPEG errors which we have patched (I'm not sure if all patches have been accepted yet upstream) | 14:06.20 |
chrisl | poppler is probably being linked to an old (probably unpatched) openjpeg library | 14:06.35 |
Robin_Watts | kens: they have not. | 14:06.44 |
| I checked yesterday | 14:06.49 |
kens | Then I owuld guess its somethign wrong in OpenJPEG, and we've fixed it, but that *is* a guess | 14:07.06 |
inkbottle | chrisl: kens: actually it works with mupdf, it used not to and I haven't tryied recently, but now it works :-) | 14:07.31 |
kens | If inkbottle sees the same problem in MuPDF is it possible that someone has packaged MuPDF, and made it use syatem shared libraries ? | 14:07.38 |
| Oh OK, so maybe its something we fixed in OpenJPEG. | 14:07.51 |
inkbottle | but I use okular to takes notes in documents | 14:07.54 |
chrisl | Rehash the PDF with Ghostscript, and Okular will probably be happy with it | 14:08.30 |
kens | True, because hte images won't be JPX any more :-) | 14:08.57 |
inkbottle | chrisl: how do i do that? | 14:08.59 |
chrisl | Use the pdfwrite device | 14:09.10 |
kens | gs -sDEVICE=pdfwrite -sOutptuFile=new.pdf input.pdf | 14:09.15 |
| -sOutputFile that should be | 14:09.29 |
| I wouldnt' normally suggest it but.... | 14:09.42 |
chrisl | It looks like a very simple PDF, I doubt there's much to go missing in this case | 14:10.17 |
kens | Of course, if the Debian packagers have diddled with GS to make it use system shared librareis it will fail with GS as well | 14:11.28 |
chrisl | gs may still be using jasper <shudders> | 14:12.32 |
inkbottle | :-) | 14:13.13 |
| we'ii se in a mnute | 14:13.27 |
| kens: chrisl: It works, thanks :-) | 14:15.55 |
kens | NP | 14:16.00 |
inkbottle | Thanks to for having introduced me to mutool | 14:18.48 |
| So do you say the reason it doesn't work with okular/poppler is because the later is linked to an old OpenJPEG lib? | 14:20.54 |
kens | Not neccesarily an old version | 14:22.29 |
| Possibly one that does not yet include some fixes we have supplied | 14:22.40 |
inkbottle | I've checked: the poppler used by okular uses libopenjpeg5=1:1.5.2-3; mupdf uses libopenjpeg2-7=2.1.0-2 | 14:25.16 |
kens | Well, its older | 14:26.29 |
inkbottle | even in "experimental" does poppler use libopenjpeg 1.5... | 14:30.38 |
| Thanks, see you | 14:30.52 |
Robin_Watts | tor8: You may have noticed that the commits this morning broke the cluster tests (even after the -d/-D thing). | 14:32.50 |
| It seems that asking pdf_is_dict() during the repair is a bad thing. | 14:33.11 |
| I'm going to try a different fix for that problem | 14:33.27 |
mvrhel_laptop | kens: what gsprint dll do you have in you Artifex Software /gsview 6.0/bin directory? | 14:34.59 |
kens | 1 moment | 14:35.07 |
mvrhel_laptop | likely a screw up in the installer | 14:35.10 |
kens | I did uninstall the previous version before starting | 14:35.28 |
mvrhel_laptop | thats good | 14:35.41 |
| I dont have 64 bit windows 7, I guess I should set up this other machine to dual boot | 14:36.26 |
kens | I have gsprint64.dll, 9Kb (8704 bytes) created 31 March 20:15 23:52:52 no version number | 14:36.31 |
mvrhel_laptop | odd then | 14:36.37 |
| that is def. the 64bit one | 14:37.33 |
kens | Hmm, well I can only go by what Windows tells me... | 14:37.47 |
mvrhel_laptop | no it is the correct one | 14:37.57 |
kens | I'll start it from the folder in case there are others somewhere | 14:38.02 |
mvrhel_laptop | the 32bit one is much smaller | 14:38.02 |
kens | If I start it form the bin folder, it seems to be wroking | 14:39.11 |
mvrhel_laptop | oh | 14:39.28 |
kens | Its doing 'COnvert to XPS@ with hindsight the PDF reference manual perhaps wasn't the best choice | 14:39.30 |
mvrhel_laptop | that is odd | 14:39.32 |
| oh I should add a cancel option on that | 14:39.56 |
kens | Yeah, I just killed the app :-) | 14:40.06 |
mvrhel_laptop | kens, it has not gotten to the gsprint part though if that is where you are | 14:40.14 |
| oh yes it did | 14:40.19 |
| sorry | 14:40.22 |
kens | Its further than it got before.... | 14:40.29 |
mvrhel_laptop | gsprint is just used for bringing up the unique printer dialog ui stuff | 14:40.43 |
| at this point | 14:40.49 |
kens | But if I select the XPS document writer its back to BadImageFormatException | 14:40.52 |
mvrhel_laptop | oh | 14:41.00 |
| ok that is where it is used | 14:41.13 |
| in the selection of different printers | 14:41.18 |
kens | Printing to the Xerox Phaser seems to work OK | 14:41.21 |
| Selecting the Adobe PDF prtiner throws the error | 14:41.48 |
mvrhel_laptop | kens: so you get the error anytime you change the printer | 14:41.51 |
| yes | 14:41.53 |
kens | Yeah every time I change printer | 14:42.04 |
| If I don't do that it seems to be better. | 14:42.12 |
| Let me change my default pritner | 14:42.21 |
| Yeah if I set Adobe PDF as the default and pritn to that its just fine, produces a nice new PDF file | 14:43.31 |
mvrhel_laptop | ok, I wish I had win7 on a 64 bit machine here. my 32bit machine with win7 seems to have died. it won't get out of bios start-up | 14:43.40 |
kens | Its just chaning printer that seems to break | 14:43.44 |
mvrhel_laptop | kens: ok I will see what I can figure out. thanks for the testing | 14:44.01 |
kens | NP | 14:44.06 |
chrisl | mvrhel_laptop: why not use a virtual machine for Win7 x64? | 14:44.54 |
mvrhel_laptop | chrisl: I think that is what I will do. | 14:45.24 |
| I don't have a copy of win7 around here, so I will need to pick that up | 14:46.04 |
Robin_Watts | mvrhel_laptop: I am using a 64bit Win7 machine here if you want stuff tested. | 14:46.34 |
| (presumably you think kens bug only occurs on a 64bit Win7 machine?) | 14:46.58 |
kens | Its worth checking to see if another installation will show it | 14:47.13 |
| Or if its somethign weird I have done here | 14:47.19 |
Robin_Watts | Let me grab the installer. | 14:47.28 |
chrisl | My 64 bit Win7 VMs don't have printers attached..... | 14:47.44 |
| I suppose I could install one | 14:48.01 |
kens | You can put a printer on FILE: chrisl | 14:48.06 |
| I don't really have a Xerox Phaser after all :-) | 14:48.27 |
chrisl | I wonder why MS saw fit to have a "Fax" device in Win7..... | 14:49.07 |
Robin_Watts | cos previous versions of the OS have had one, and removing functionality will make *someone* complain? | 14:49.51 |
| Also, fax is still very useful for some things. | 14:50.03 |
chrisl | But I reckon very few people actually use it - seems strange that it isn't optional | 14:50.52 |
| Same error kens saw: BadImageFormatException: incorrect gsprint DLL | 14:53.24 |
kens | But only when switching printers ? | 14:53.35 |
| If you just print straight to the default its OK ? | 14:53.47 |
mvrhel_laptop | ok. I will get win7 then and figure this out | 14:53.50 |
chrisl | Yes, only when switching. Oddly, click okay, and printing still works...... | 14:54.45 |
kens | Oh, I didn't try that | 14:54.57 |
Robin_Watts | will not bother installing the version I just downloaded then. | 14:55.32 |
| This way there will still be at least 1 fresh machine to try the next version on :) | 14:55.48 |
chrisl | mvrhel_laptop: you have Win7 32 bit? | 14:56.02 |
mvrhel_laptop | yes, but the machine died this morning | 14:56.29 |
| wont get past bios startup ;( | 14:56.39 |
chrisl | Oh, so should I try that, too? | 14:56.40 |
mvrhel_laptop | that would be good | 14:56.45 |
chrisl | There'll be a short delay while I switch VMs...... | 14:57.56 |
| Hrm, this will take a little longer - it's having to download the .net runtime stuff | 15:00.27 |
henrys | just running the mac installer and it occurred to me we don't have a click through license - not to hold up a beta but one for the todo list. | 15:00.37 |
chrisl | Really?? It's there on the Linux one, I think..... | 15:01.19 |
henrys | chrisl: upon install? | 15:02.09 |
chrisl | henrys: yes | 15:02.30 |
henrys | chrisl: okay so a mac bug for fred | 15:03.04 |
| I was replacing but that shouldn't matter. | 15:03.27 |
chrisl | After the "select components" screen, the next screen is the license, and there's a radio button for accept/not accept | 15:03.34 |
Robin_Watts | tor8: 1 fix on robin/master | 15:04.50 |
henrys | chrisl: nothing like that on the mac. I think that would be a problem because you can't run anything on the mac until you okay the "unidentified" developer. | 15:06.16 |
| I assume chrisl has inux covered and kens windows, so I'll just focus on mac. | 15:07.57 |
chrisl | I haven't done as much as I would have liked - busy with other things...... | 15:08.19 |
henrys | I got a crash opening the pdf reference the 2nd time, is there a crash log somewhare ? mvrhel_laptop ? | 15:17.17 |
mvrhel_laptop | henrys: is this on the mac? | 15:17.33 |
henrys | mvrhel_laptop: yes | 15:17.39 |
mvrhel_laptop | henrys: we are going to need to pull fred in on this then | 15:18.01 |
henrys | mvrhel_laptop: okay he'll be here in a bit. | 15:18.27 |
mvrhel_laptop | we probably need to log these in bugzilla | 15:18.31 |
| I know there is an entry for gsview but does that go to russell? | 15:19.25 |
kens | it gets assigned to russel | 15:19.44 |
mvrhel_laptop | ok that needs to be changed then | 15:19.51 |
| bbiab | 15:20.16 |
henrys | mvrhel_laptop: I wonder if we should have artifex gsview for now? | 15:20.31 |
| or I can just change it, let me know. | 15:20.46 |
tor8 | use pdf_is_indirect instead? /Root *should* be an indirect reference to an object | 15:22.47 |
| Robin_Watts: ^ | 15:22.55 |
| though I expect your #2 fix is more robust | 15:23.27 |
Robin_Watts | tor8: The problem in the bug was that we had /Root 10 0 R and object 10 was [] | 15:23.37 |
| (or something like that). | 15:23.47 |
tor8 | Robin_Watts: right. so your #2 fix is the best way. | 15:23.56 |
Robin_Watts | yeah, it's the best I can think of. | 15:24.05 |
chrisl | mvrhel_laptop: the DLL error does not happen on Win7 32bit when switching printers in the dialogue. | 15:28.02 |
Robin_Watts | tor8: so can I push it? | 15:28.46 |
mvrhel_laptop | I think gsview should be artifex gsview. ghostgum gsview should be russell's | 15:29.39 |
Robin_Watts | mvrhel_laptop: I might argue for "gsview" and "legacy gsview" | 15:30.08 |
chrisl | But if we're not letting Russell do new releases, then surely there's no point having it in bugzilla at all? | 15:30.20 |
mvrhel_laptop | that sounds resonable | 15:30.21 |
| that true too | 15:30.28 |
chrisl | Seems pointless letting people report stuff that can't ever be addressed | 15:31.11 |
mvrhel_laptop | right | 15:31.17 |
henrys | sigh ... this is an intermittant crash. Still have all the white space on the right of the page if I open the file from finder, I open it from gsview the file is centered in too much white space. Strange | 15:33.15 |
Robin_Watts | chrisl: yeah, I think I agree. | 15:34.06 |
| So change the gsview category to point to mvrhel, and when we release, close all the bugs with "use the new version". | 15:34.43 |
henrys | okay I'll change it to michael | 15:34.54 |
chrisl | Or (also) change it to "artifex gsview" so we're explicit about it..... | 15:35.23 |
henrys | default assignee is michael but I've changed the name. | 15:37.08 |
chrisl | And I've now stopped tbird asking me "have you forgotten to attach a file" on almost every e-mail....... | 15:43.00 |
mvrhel_laptop | bbiab | 15:43.58 |
henrys | fredross-perry: just asking a gsview mac question, is there a crash log anywhere for qt? I'm getting an intermittant crash with a large pdf. | 15:46.17 |
tor8 | Robin_Watts: go ahead | 15:46.27 |
fredross-perry | Look in the Console app under User Diagnosic Reports. | 15:48.00 |
| ⦠and pls send me the PDF file. | 15:48.55 |
henrys | fredross-perry: will do, yes there is a nice backtrace I'll create a bug. | 15:53.29 |
fredross-perry | lovely. | 15:53.41 |
| henrys: good time to talk? | 16:03.18 |
henrys | sure fredross-perry | 16:03.50 |
fredross-perry | OK hang on | 16:04.08 |
mvrhel_laptop | chrisl, so did you have the printer selection issue with the 32 bit version of windows 7? | 16:11.59 |
chrisl | mvrhel_laptop: no, it worked fine on 32 bit, as did printing. So it is a 64 bit only issue | 16:13.45 |
mvrhel_laptop | chrisl: ok thanks a bunch for testing | 16:14.00 |
chrisl | And as the printing on 64 bit still actually worked, it seems the error is an error - so to speak..... | 16:14.25 |
mvrhel_laptop | I agree | 16:14.34 |
kens | TIme to go, goodnight all | 16:20.18 |
henrys | fredross-perry: http://bugs.ghostscript.com/show_bug.cgi?id=695895 | 16:20.50 |
| fredross-perry: also do you see the huge amount of whitespace with that pdf? | 16:56.23 |
fredross-perry | what seems to be happening is that the scrolling area has a bunch of space at the top (and the bottom). If you use the toolbar Home and ENd buttons, you get to the first and the last pages. Something is up, Iâll have to see. | 16:59.59 |
| Also, I get the crash and can fix it with a change to pdf-annot.c. pdf_parse_file_spec is being given a null file_spec. | 17:01.05 |
henrys | fredross-perry: I do know I did something weird to that pdf, experimenting with annotations but I can't clearly remember. | 17:03.45 |
fredross-perry | henrys: in Preview, if you enable âHilights and notesâ, it crashes. The issue happens on page 1139. | 17:24.54 |
Robin_Watts | If MuPDF is dying on that PDF, we should look to see why. | 17:29.41 |
henrys | Robin_Watts: I don't get a preview crash. | 17:35.09 |
Robin_Watts | henrys: have you enabled Highlights and notes ? | 17:35.32 |
henrys | Robin_Watts: I did but we may have different versions | 17:35.51 |
| fredross-perry: with preview 8.0 I don't see a problem with highlights & notes on. | 17:41.10 |
fredross-perry | henrys: yes, Preview 8.0. open doc, scroll to page 1139, turn on H&N, bech ball, crash. | 17:43.01 |
| *beach* | 17:43.18 |
henrys | fredross-perry: right I don't see it. maybe you have less memory which causes the problem to happen. I notice preview sometimes does odd things when constrained resource wise. not sure though... | 17:46.22 |
fredross-perry | I have 8 GB. Yosemite 10.10.2. | 17:47.00 |
henrys | fredross-perry: anyway if mudf crashes independently something robin or tor should investigate ... just reassign and continue on, all the same here except I have 10.10.1... it's a macpro quad core xeon. | 17:49.26 |
fredross-perry | OK, Iâll add something to the bug abut the code and move on. Thanks. | 17:50.15 |
Robin_Watts | fredross-perry: Urm... Before you get to pass us the bug, you have to tell us how to regenerate it with normal mupdf :) | 17:52.33 |
Kingihai | Hi there | 17:52.41 |
ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 17:52.41 |
Kingihai | I have a question about submitting data on pdf | 17:53.59 |
henrys | or it could be a good problem for fred to work on Robin_Watts , you decide. | 17:54.13 |
| Robin_Watts: I'd decide but I'm leaving for lunch ;-) | 17:54.45 |
Robin_Watts | henrys: If you wanted to transition fred into working within mupdf, then it would seem reasonable to let him run with it until he hits a wall. | 17:54.54 |
henrys | okay when he pops back tell him we agreed on that. bbiab Robin_Watts | 17:55.27 |
Robin_Watts | ok. | 17:55.36 |
Kingihai | Any | 17:55.44 |
| Any one? | 17:56.05 |
Robin_Watts | Kingihai: Ask the question, don't ask to ask :) | 17:56.23 |
Kingihai | Yes | 17:56.38 |
| Yes, thank you. How can I suceed to send data in android device through submit button? Mupdf can do it? In android device... | 17:58.32 |
Robin_Watts | Kingihai: Currently, MuPDF does not support submitting data in any form. | 17:58.56 |
| What it can do is to save the file you are editing to disc, and you could then extend the android specific app to submit that saved file. | 17:59.32 |
| but I can't give you a simpler answer than that, sorry. | 17:59.55 |
Kingihai | Impossible? Or not yet? | 18:00.02 |
Robin_Watts | It's not impossible. It's not done yet. It's not on our roadmap though. | 18:02.13 |
Kingihai | It is impossible just like Adobe form-central? | 18:02.14 |
Robin_Watts | I am not familiar with Adobe form-central. | 18:02.29 |
Kingihai | You say that it is not impossible , right? | 18:03.42 |
Robin_Watts | It's a simple matter of programming. | 18:04.23 |
Kingihai | Really... Wow | 18:04.38 |
Robin_Watts | I did not say that it was a matter of simple programming :) | 18:05.06 |
Kingihai | Can I have that ? | 18:05.35 |
Robin_Watts | Can you have what? | 18:05.43 |
Kingihai | Dynamic PDF data can be moved to another server. | 18:06.40 |
Robin_Watts | Sorry, I think we're hitting a language barrier here. | 18:06.59 |
| MuPDF cannot currently do submission of data. | 18:07.09 |
| If you want it to do such, then you will need to write some new code. | 18:07.23 |
Kingihai | I'm Korean.. Bad English ... | 18:08.34 |
| I create a dynamic pdf with check box, radio button... I want to send form data to server... | 18:09.48 |
Robin_Watts | I understand what you want. | 18:10.07 |
| MuPDF cannot do it. | 18:10.12 |
Kingihai | In case expert PDF in Apple that is ok.. | 18:10.44 |
| But mupdf can not.. | 18:10.57 |
| Impossible.... | 18:11.37 |
Robin_Watts | Currently, impossible. | 18:11.54 |
Kingihai | How long does it take ? | 18:12.14 |
| To be possible? | 18:12.33 |
Robin_Watts | I don't know. | 18:12.39 |
Kingihai | You are not developer? | 18:13.02 |
Robin_Watts | I am. | 18:13.08 |
| But I have no plans to add that functionality. | 18:13.19 |
Kingihai | Money can do? | 18:14.28 |
mvrhel_laptop | ok. gsview had an issue with temp files not getting deleted. that is now fixed | 18:14.47 |
Robin_Watts | Kingihai: Yes. | 18:14.58 |
mvrhel_laptop | my neighbor has a win 7 64 bit copy that i will borrow and see if I can replicate the issue you all are seeing | 18:15.14 |
Kingihai | Can you tell me how much? | 18:15.27 |
Robin_Watts | If you contact sales@artifex.com they can discuss paid work. | 18:15.30 |
| I cannot tell you how much. sales@artifex.com can tell you how much. | 18:15.44 |
| Do you have a budget? | 18:15.51 |
Kingihai | No, but I have to get it.. | 18:16.38 |
Robin_Watts | I cannot think that this will be less than a couple of weeks work, possibly more. I would need to look into it to be sure. | 18:17.05 |
Kingihai | Can I send to you sample file? | 18:18.21 |
Robin_Watts | Kingihai: go to bugs.ghostscript.com and open a bug. | 18:18.43 |
| Add the sample file to the bug. | 18:18.50 |
Kingihai | Already last year.. | 18:19.16 |
Robin_Watts | Kingihai: what bug number? | 18:19.43 |
Kingihai | It is long time ago... Can you wait for a while? | 18:20.32 |
Robin_Watts | sure. | 18:20.40 |
| fredross-perry: Henry and I discussed the bug briefly. | 18:22.23 |
| We think it would be a good way to start you on the internals of mupdf development. | 18:22.47 |
fredross-perry | Whatâs happening is that pdf_parse_file_spec is getting called with a null file_spec. | 18:22.54 |
Robin_Watts | Do you want to run with it and see how far you can get? | 18:23.01 |
| oh, ok. | 18:23.08 |
| sounds like you're most of the way there then. | 18:23.19 |
fredross-perry | Itâs easy to protect against that inside pdf_parse_file_spec, but I donât know whats happening further upstream. | 18:23.35 |
Robin_Watts | pdf_parse_file_spec is only called from 2 places. | 18:24.13 |
| both inside pdf_parse_action | 18:24.25 |
fredross-perry | I think the logic is sound, but thereâs an uninitialized var issue. | 18:25.19 |
Robin_Watts | you'd need to get pdf_parse_file_spec to return NULL in this case, and then ensure that callers to pdf_parse_action can cope with null file_spec. | 18:25.36 |
| I don't think much of our code uses this at the moment, so it may not be a big job :) | 18:25.52 |
| if (pdf_is_string...) ... *else* if (pdf_is_dict ... ) ... *else return NULL;* | 18:26.43 |
| or better... else filename = NULL; | 18:27.23 |
fredross-perry | vry first line is âpdf_obj *filename;â should be âpdf_obj *filename=NULL;â. Thatâs enough. | 18:27.57 |
Robin_Watts | That would work too. | 18:28.22 |
mvrhel_laptop | Robin_Watts: I will look at 695861 now | 18:28.52 |
Robin_Watts | mvrhel_laptop: Thanks. | 18:28.59 |
| I suspect the short answer is that he is trying to abuse one of your functions in a way that's not intended. | 18:29.24 |
| but the long answer might be that we can remove bottom_right from the args. | 18:29.39 |
mvrhel_laptop | right | 18:31.16 |
fredross-perry | now the procedural question. Does a beta gsview need to be built with a released version of mupdf? | 18:32.15 |
Robin_Watts | fredross-perry: No. | 18:32.43 |
fredross-perry | ok then. | 18:32.53 |
Robin_Watts | But we should get your fix into mupdf asap, so it makes it into git at least. | 18:33.03 |
| Do you want to do a commit, and I'll review it? | 18:33.31 |
fredross-perry | sure | 18:34.14 |
| fixed with commit f4590b1fe1dac863d725447d4f90fe8f2048580a | 18:40.26 |
| in fred@ghostscript.com:repos/mupdf.git | 18:40.27 |
Robin_Watts | I don't see that commit on any branch I've fetched here.. | 18:41.31 |
| Ah, yes I do | 18:41.49 |
mvrhel_laptop | Robin_Watts: so I think I would rather not do this the way he has done it. I can understand the need to render a subregion, but I would rather not have that done in the function. I will rip out the unneeded params and I guess add an procedure that includes a clipping rect | 18:41.54 |
| s/the function/this function/ | 18:42.10 |
| s/an/a/ | 18:42.18 |
Robin_Watts | mvrhel_laptop: even for rendering a subregion I don't follow his code. | 18:42.24 |
| You ignore bottom_rect. | 18:42.32 |
mvrhel_laptop | yes | 18:42.35 |
| that makes no sense to me either | 18:42.39 |
Robin_Watts | As far as I could tell, his new code ignored scale. | 18:42.40 |
| and it calculated separate X and Y scales based on bottom_right (and top_left) | 18:43.06 |
| So I'd be tempted to just make the code do what you need. | 18:43.17 |
| and leave him to put his bastardised function in his own code. | 18:43.30 |
mvrhel_laptop | ok will do | 18:43.42 |
| thank ;) | 18:43.48 |
fredross-perry | be back in 15⦠| 18:43.48 |
mvrhel_laptop | thanks | 18:43.52 |
Robin_Watts | no worries. | 18:43.55 |
mvrhel_laptop | has tor ripped the gsview code out of the mupdf project yet? | 18:44.17 |
Robin_Watts | mvrhel_laptop: I have not seen it disappear yet. | 18:44.29 |
mvrhel_laptop | I saw that one of the customers had asked him a question about some of the code | 18:44.39 |
Robin_Watts | I may have missed that. | 18:45.01 |
mvrhel_laptop | Junghen from yesterday | 18:45.21 |
Robin_Watts | Junghen was asking about structured text extraction, I think. | 18:46.23 |
| I don't think that's your code, is it? | 18:46.35 |
mvrhel_laptop | the code he pasted in is mine | 18:46.53 |
Robin_Watts | Oh! He is using your code. I hadn't twigged. | 18:47.01 |
mvrhel_laptop | the sooner it is ripped out of the mupdf project, probably the better.... | 18:47.31 |
Robin_Watts | yeah. I suspect tor will get to that soon (tomorrow?) as he starts looking at making the rc. | 18:47.56 |
| but I'd keep committing changes there until he says otherwise. | 18:48.18 |
mvrhel_laptop | oh that ship has sailed. gsview has its own repository | 18:48.37 |
| we did that at the last staff meeting | 18:49.11 |
| http://git.ghostscript.com/?p=gsview.git;a=summary | 18:49.25 |
Robin_Watts | mvrhel_laptop: Oh, ok, so he just needs to expunge it from mupdf.git. | 18:50.13 |
mvrhel_laptop | so mupdf and ghostscript are submodules | 18:50.14 |
| yes | 18:50.17 |
Robin_Watts | mvrhel_laptop: That sounds ideal. | 18:50.25 |
mvrhel_laptop | it has worked out rather nice | 18:50.37 |
| when changes happen in the mupdf or ghostscript project solutions the conversion to vs2013 pretty much takes care of the changes | 18:51.01 |
| so keeping everything in-sync is pretty easy | 18:51.22 |
Robin_Watts | fredross-perry: commit looks good to me. about to push it. | 18:53.10 |
fredross-perry | lovely thx | 18:53.20 |
mvrhel_laptop | Robin_Watts: so do you want to close Bug 695861 and let him know we are not taking that patch or shall I | 18:53.29 |
| I will clean up the params in the function though | 18:53.48 |
Robin_Watts | go for it. You can speak with more authority than me. | 18:53.56 |
mvrhel_laptop | ok | 18:54.16 |
Kingihai | Re hi.. | 19:06.21 |
| I uploaded sample file, bug695896 | 19:07.20 |
Robin_Watts | Kingihai: Thanks. | 19:07.30 |
Kingihai | In Desktop and laptop Data sending is ok | 19:09.36 |
| In Apple expertpdf it is ok. | 19:10.24 |
| I want to be ok in android by mupdf.. | 19:11.07 |
mvrhel_laptop | ok I need to add the ability to do annotations and form filling with gsview. AR supports both | 19:11.08 |
| so I need to pull in mujs.... | 19:12.27 |
Robin_Watts | Kingihai: OK. That's using javascript internally. | 19:12.39 |
Kingihai | Yes | 19:12.49 |
mvrhel_laptop | bbiaw lunch time | 19:12.57 |
Robin_Watts | Our current javascript lib does not have SOAP and Http integration currently, AIUI. | 19:13.11 |
Kingihai | Need much time? | 19:14.30 |
Robin_Watts | Kingihai: The guys that know about the forms/javascript support have gone for the night. Try again here in 15-18 hours or so. | 19:14.34 |
| Kingihai: They might be able to tell you - I cannot. | 19:14.53 |
Kingihai | Ok.. I got it | 19:15.11 |
| Thank you a lot | 19:15.24 |
| Can you tell me his name? | 19:16.24 |
Robin_Watts | tor8 and paulgardiner. | 19:16.37 |
Kingihai | Maybe nickname? | 19:17.48 |
Robin_Watts | tor8 = Tor Andersson | 19:18.12 |
| paulgardiner = Paul Gardiner | 19:18.17 |
Kingihai | You Robin Watts? Thank you for your helping. After 15 hours, I will come again. | 19:20.27 |
Robin_Watts | see you then. | 19:20.46 |
Kingihai | Bye Robin.. | 19:21.31 |
pipitas | I want to use a font that is part of a TrueType Collection (Avenir-Heavy in Avenir.ttc on a Mac) with Ghostscript/PostScript. Is there any way to address it directly (without extracting it as a separate TTF from the TTC file)? | 19:41.13 |
| I *think* I know how I should add it to cidfmap under its own name, following the Use.htm document. But I'm looking for something *direct* for the command line like -c "/Avenir-Heavy <</FileType /TrueType /Path (/System/Library/Fonts/Avenir.ttc) /SubfontID 9>> findfont 180 scalefont setfont â¦[more PostScript code]â¦" | 19:50.12 |
henrys | pipitas: I imagine kens would know off the top of his head. I know you can add truetypes to the font map and I assume collections would work as well, thus there is likely a command line incantation. | 19:53.50 |
pipitas | In font map you would add a line "/Avenir-Heavy <</FileType /TrueType /Path (/System/Library/Fonts/Avenir.ttc) /SubfontID 9>>;" â but I'm not sure how to translate that into proper PostScript for CLI invocation with "findfont"⦠| 19:57.01 |
mvrhel_laptop | found my windows 7 disc | 20:02.32 |
| now to see if I can do remote debug through vmware.... | 20:06.06 |
rakm | hi, i'm trying to install gs 9.0.7 through brew and not sure how | 20:24.14 |
rayjj | mvrhel_laptop: Robin_Watts: My optimization for customer 532 tripped over a L...O...N...G... standing bug in the calculation in RECT_ENUM_INIT. It gets "nbands" *WRONG*, but it looks like my code is the first to actually use it | 20:24.16 |
| rakm: 9.16 was just released. I recommend that you don't bother with 9.07 | 20:24.43 |
| rakm: I don't know what 'brew' is. Googling.. Googling... Googling... | 20:25.35 |
rakm | rayjj: I'm using 9.15 right now and some legacy code is failing with the error: `Unknown device: psrgb` | 20:25.55 |
| i know we were using 07 originally and just wnat to go back to a working version | 20:26.08 |
| don't relaly have the time to debug why or deal with externalities of doing it a different way right now :/ | 20:26.28 |
rayjj | rakm: Google just talks about beer related stuff. What is brew | 20:27.03 |
rakm | brew is the osx package manager. i'll need to do it for our rhel deploy also, but i think it's easier to install specific versions with yum | 20:27.05 |
| rayjj: beer related stuff right now would be nice :) | 20:27.18 |
rayjj | rakm: I think we removed the psrgb device to discourage its use | 20:27.44 |
rakm | i'm sure there are great reasons. and i'd love to do it the right way soon | 20:28.12 |
| but don't have the luxury right now | 20:28.22 |
rayjj | rakm: oh, yeah. That created a PS file that was really just a wrapper on a 24-bit image | 20:28.27 |
rakm | just want to get to working state | 20:28.33 |
rayjj | rakm: the 'psim' code is still there / sort of / but it looks like it was removed | 20:35.18 |
| rakm: as far as "getting back to a working 9.07" why can't you just pull that source in ? | 20:35.47 |
rakm | rayjj: that's what i'm trying to do. do you know what sha is tagged at 9.0.7? | 20:36.17 |
| git tag doesn't have anything | 20:36.24 |
rayjj | rakm: my "git tag" shows ghostscript-9.07 | 20:38.07 |
rakm | hmm, i'm looking in the wrong place | 20:38.25 |
rayjj | rakm: and git log -t ghostscript-9.07 shows commit 76b4193e5d8cb280f4795b0e959657ad695ad54d | 20:38.59 |
rakm | ughh, package management | 20:39.04 |
| nice, thanks rayjj | 20:39.14 |
| i don't know know how to actually check that out / install it with brew | 20:39.27 |
rayjj | git co -t ghostscript-9.07 should check it out | 20:39.51 |
| or if you don't have 'co' aliased maybe git checkout -t ghostscript-9.07 | 20:40.21 |
henrys | Robin_Watts: so now the cluster test has to actually look in the uploads dir when testing. My upload is there where it is expected but there is evidence it doesn't get looked at when the cluster runs. | 21:48.28 |
| is that for marcosw1 ? | 21:48.58 |
Robin_Watts | henrys: I believe I added the uploads dirs to the tests. | 23:17.07 |
| hmm, evidently not. Doing that now. | 23:18.40 |
| done. | 23:19.25 |
| Forward 1 day (to 2015/04/02)>>> | |