| <<<Back 1 day (to 2013/02/03) | 2013/02/04 |
sebras | http://abstrusegoose.com/493 | 00:55.07 |
wizardxbl | hello everybody , i'm a freshman to mupdf | 02:24.40 |
| i want save doc to pdf but fz_write_document() seems not work well | 02:25.45 |
| i just find the function pdf_write_document to done this work | 02:49.29 |
kens | Hmm, my cluster push gave me 270 'file(s) have started producing errors' almost all of which are CUPS and nothing to do with the code I've changed.... | 09:49.59 |
chrisl | Quite likely they're "known errors" but the cluster lost the "known" part..... | 09:50.57 |
kens | I hope so, I can't believe that I've caused them. | 09:51.11 |
chrisl | Well, beats me...... | 09:57.32 |
kens | Well I've committed my change, its pretty simple and clearly was incorrect (for 32-bit builds) before. | 09:58.46 |
Robin_Watts | D'Oh. | 11:14.32 |
| I've had this window scrolled up for a while, so wondered why nothing was happening. | 11:14.56 |
chrisl | Robin_Watts: DVD arrived, thank you! | 11:17.28 |
Robin_Watts | chrisl: Ah, cool. | 11:17.39 |
| tor8: ping | 11:40.21 |
tor8 | Robin_Watts: morning | 11:41.16 |
Robin_Watts | morning | 11:41.34 |
| have you tried Pauls latest prototype? | 11:41.48 |
| He has it so that when you hit the 'search' button, each page goes through the text extraction device, and into a webview. | 11:42.32 |
| a webview just shows html. | 11:42.43 |
| Currently zooming doesn't work -that's the next thing for him to look at. | 11:42.57 |
| Styles don't seem to work either at the moment. | 11:43.11 |
| and (I'm looking at pdf_reference17.pdf) paragraph spacings seem to come through as 3 top bit set chars. | 11:44.28 |
| Are you available to look at the styling/paragraph issues ? | 11:44.46 |
| I'm trying to avoid touching any code until I get the quote done and dusted otherwise I'll get distracted. | 11:45.15 |
paulgardiner | oh yeah, nearly forgot: we need a reflow icon | 11:45.40 |
Robin_Watts | I could draw one, or we could look for one on the usual places. | 11:46.53 |
tor8 | paulgardiner: http://somerandomdude.com/work/iconic/ is what we currently use | 11:52.49 |
paulgardiner | Can you see a good one for reflow? | 11:53.20 |
tor8 | Robin_Watts: what needs to be done with the styles? | 11:53.27 |
Robin_Watts | We need to figure out why they aren't appearing in the webview. | 11:53.52 |
tor8 | paulgardiner: maybe one with the lines of text? like a "left-justified" icon? | 11:54.22 |
paulgardiner | I'm currently not using the code that produces html with styles. I guess what we need is the code that does that to produce a fz_buffer rather than assuming a FILE | 11:55.11 |
Robin_Watts | paulgardiner: Ah! | 11:55.22 |
paulgardiner | I'm just adding <html><body><div> plus lots of <p>s myself at the moment. | 11:56.05 |
Robin_Watts | Ideally we need that code to be generalised so that it can go to a variety of outputs; one of which is a FILE, one of which is a buffer. | 11:56.10 |
| That explains a lot. | 11:56.13 |
| You'd told me previously, and I'd forgotten it. | 11:56.27 |
tor8 | Robin_Watts: so, a fz_output that can write either to file or to a fz_buffer? | 11:57.03 |
paulgardiner | I thought we'd moved away from iconic | 11:57.17 |
tor8 | with fz_write and fz_print and fz_printf etc | 11:57.18 |
| paulgardiner: we will, but we haven't yet | 11:57.25 |
paulgardiner | ah ok | 11:57.30 |
Robin_Watts | Specifically whatever function does -tt. | 11:57.33 |
tor8 | paulgardiner: we liked http://adamwhitcroft.com/batch/ but that also has a limited (but less so) set of icons | 11:57.56 |
| paulgardiner: I have a repo with pre-rendered png files of all these sitting around | 11:58.13 |
paulgardiner | tor8: come to think of it an fz_buffer may be no good: we do want unicode... although utf8 would be ok | 11:58.30 |
tor8 | we use utf8 everywhere, unconditionally | 11:58.44 |
| http://ghostscript.com/~tor/stuff/mupdf-debug-batch.apk was the mupdf apk with the batch icon set | 11:59.32 |
paulgardiner | ok that's fine. I was using unicode because I was accessing the block/line/span/char heirarchy directly | 11:59.48 |
tor8 | http://ghostscript.com/~tor/stuff/mupdf-debug-awesome-trim.apk was with font-awesome | 11:59.49 |
| paulgardiner: yeah. anything that takes or returns a char* in mupdf is utf-8 | 12:00.09 |
| the fz_tex_char is a raw unicode int though | 12:00.19 |
| (and I need to trim my fingernails, keep slipping and missing keys) | 12:00.32 |
paulgardiner | Right. That makes sense. That's why it seemed to be working. | 12:00.45 |
| :-) Actually I am converting to utf8. I forgot | 12:01.31 |
| So if you can produce utf8 with styles in, it'sll slot straight in | 12:01.54 |
Robin_Watts | http://intranet.picsel.com/~robin/ic_reflow2.png | 12:02.23 |
| http://intranet.ghostscript.com/~robin/ic_reflow2.png | 12:02.34 |
| Ahem. Nothing freudian in that slip at all. | 12:02.43 |
| http://ghostscript.com/~robin/ic_reflow2.png | 12:02.57 |
paulgardiner | Robin_Watts: do you have that in the two sizes we need? 32 and 24 I think | 12:03.55 |
Robin_Watts | Doing the other one now. | 12:04.03 |
paulgardiner | Magic | 12:04.33 |
Robin_Watts | http://ghostscript.com/~robin/ic_reflow.png | 12:08.43 |
| http://ghostscript.com/~robin/ic_reflow2.png | 12:08.46 |
| Those are the l and the m variants respectively. | 12:08.53 |
| I'm sure we can find better ones at some point, but they'll do for now. | 12:09.43 |
paulgardiner | They look different from the previous ic_reflow2 you postef | 12:10.51 |
Robin_Watts | the previous one was white on black. | 12:11.14 |
| These are white on transparent. | 12:11.20 |
| which is what we actually need, right ? | 12:11.29 |
paulgardiner | ah yes, just as they need to be | 12:11.31 |
| perfect | 12:11.44 |
| The top menu will be a bit crowded but we can worry about that later | 12:12.14 |
| Any preference where in the sequence of icons to put this? | 12:12.35 |
Robin_Watts | paulgardiner: Leave search on the right. | 12:16.14 |
paulgardiner | :-) I just inserted it to the left of search before you said that | 12:16.51 |
Robin_Watts | but otherwise I don't care. | 12:16.57 |
| :) | 12:16.58 |
| Possibly we should have a 'Spanner' icon as the only icon in that top bar. | 12:17.58 |
| And when you click the spanner icon the tool bar appears. | 12:18.06 |
paulgardiner | Yeah, or the spanner and just the few that we think are high priority | 12:18.38 |
Robin_Watts | but it's a different tradeoff on phones/tablets. | 12:18.40 |
tor8 | a spanner with a text menu would be ideal, so we don't have to worry about so many icons! | 12:34.01 |
| just the basic few common operations + spanner | 12:34.16 |
| so bookmarks+spanner, and in search mode just the arrows and X | 12:34.37 |
| and maybe textual buttons for the annotation mode? | 12:35.05 |
Robin_Watts | tor8, paulgardiner: First attempt at a roadmap sent. Please give feedback. | 12:47.51 |
paulgardiner | will do | 12:48.10 |
Robin_Watts | I'm hoping that this is what Miles wanted. He can convert from times -> $$$ better than I can. | 12:48.14 |
paulgardiner | Robin_Watts: still reading so this may be already in there, but I better mention it now before I forget: "using the same fonts as in the document"; foxit seems to do that even if the doc uses an unusual font. | 12:52.45 |
Robin_Watts | hmm. It's conceivable that it's feeding fonts through as OpenWebFont things? | 12:54.45 |
paulgardiner | Robin_Watts: Improve out of order text recognition - I'd imagine there would be a research element to that, and I'd be surprised if we could achieve anything in 3 days | 12:55.18 |
Robin_Watts | paulgardiner: I was intending that to be just an extension of our search algorithm; rather than being 'either it goes on the end of what we have or it's a new span' | 12:55.56 |
paulgardiner | oh ok | 12:56.08 |
Robin_Watts | So just a post process pass through where we look to further collate chars/spans/lines etc, and reorder blocks. | 12:57.10 |
| yes | 12:58.51 |
paulgardiner | All looks good to me. I imagine that there's some possibility that for the sake of faster image and/or font handling, we might choose to create our own text-flowing view 1w/2w/3w | 13:00.11 |
Robin_Watts | paulgardiner: Right, but it's possible that the customer might be specifically looking for a PDF -> html process. | 13:06.38 |
paulgardiner | AH yes, true. I hadn't thought of that. | 13:07.10 |
| Quite likely now you mention it | 13:07.23 |
tor8 | Robin_Watts: maybe data: urls for embedding fonts in the html/stylesheet can work? | 13:09.45 |
Robin_Watts | tor8: I believe paul is sending the whole thing as one big data: url at the moment. | 13:10.41 |
| tor8: What formats can browsers generally accept fonts in? | 13:11.37 |
tor8 | Robin_Watts: otf and ttf I think | 13:19.15 |
Robin_Watts | ok, so that could be another item on the roadmap? | 13:19.32 |
tor8 | so type1 would have to be converted to cff and wrapped i an otf | 13:19.37 |
| and cff wrapped in an otf | 13:19.42 |
| type3 would probably just have to be substituted | 13:19.51 |
| Robin_Watts: paulgardiner: a few comments on the road map in the email | 13:20.23 |
Robin_Watts | tor8: Thanks, comments look good. | 13:23.06 |
| Care to write me a 'task' for RTL reordering ? | 13:23.23 |
| For 'undo hyphenation' I was expecting us to guess. | 13:24.06 |
tor8 | Robin_Watts: sumatrapdf has a patch | 13:24.11 |
Robin_Watts | If the next line starts with a Capital, no joining. | 13:24.16 |
tor8 | but it's ugly | 13:24.18 |
Robin_Watts | so Ankh-Morpork would be safe. | 13:24.29 |
| but city-state would not. | 13:24.35 |
tor8 | Robin_Watts: that takes care of Ankh-Morpork but not city-state. | 13:24.36 |
Robin_Watts | All this is 'best guess' of course. | 13:25.04 |
tor8 | a dictionary lookup for the word parts will help, but only if we know the language | 13:25.13 |
| so Caps to spot proper names, and then check if the split parts are words in themselves | 13:25.55 |
Robin_Watts | I think we just assume that any words split at the end of lines by a - where the second one isn't capitalised are hyphenated. | 13:26.27 |
| We'll get more things right than wrong, I suspect. | 13:26.36 |
| I may add "(imperfect, heuristic based, language independent)" to the roadmap. | 13:27.34 |
tor8 | in swedish there's even more ceaveats. there's a rule that if you have compound words with common parts in a list (like waterpower and windpower) you write them as "water- and windpower". | 13:29.25 |
| and in general there's a lot more hyphens used for non-hyphenation | 13:29.41 |
Robin_Watts | tor8: same rule in english. | 13:30.06 |
tor8 | oh. I've never seen that in use in english! | 13:30.21 |
| so joining hyphens is dangerous, but the best we can do is guess | 13:30.52 |
Robin_Watts | so maybe we should ignore hyphens. Or have it as an option. | 13:31.06 |
tor8 | I think looking at a dictionary and not join if both sides of the hyphen are complete words | 13:31.12 |
Robin_Watts | That assumes we have a dictionary. | 13:31.25 |
tor8 | and only do the joining if we can spot that it's english :) | 13:31.25 |
| there is one in /usr/share/dict/words :) | 13:32.22 |
Robin_Watts | only on unix, and only for english. | 13:32.40 |
tor8 | but that's another 2.4MB to pack in, and only for english | 13:32.47 |
Robin_Watts | Having it as an option may be worthwhile. | 13:33.02 |
tor8 | in practice, I don't think hyphenation is all that common in pdf documents people pass around | 13:33.59 |
Robin_Watts | it's on page 2 of the pdf_reference17.pdf hence me spotting it :) | 13:35.29 |
| tor8: Can you write me a task for the RTL reordering please? | 13:39.16 |
sebras | Robin_Watts: so fixing copying of arabic text is on the agenda then? cool. :) | 13:39.57 |
tor8 | Add RTL reordering pass to text extraction for both copy/paste, search and reflow. 1w/2w/1m. | 13:40.23 |
Robin_Watts | And what section does that go in? | 13:40.46 |
| does Adobe Reader get that right? | 13:41.12 |
tor8 | but that's just reordering from visual LTR to logical RTL so that the text extraction actually gets the correct logical order | 13:41.39 |
Robin_Watts | Ok, updated version sent with tweaked timings. | 13:42.36 |
| Let me know if I missed any comments, or if you have further ideas. | 13:42.46 |
Robin_Watts | lunches | 13:42.51 |
tor8 | I think it should go in the early part, since it's pretty crucial and we do have open bugs for it. | 13:43.35 |
kens2 | Does anyone know if we have a routine to 'unescape' a PostScript string ? THat is to convert a PostScript string with non-ASCII characters represented with \xxx octal escapes back into hex binary representation ? | 13:50.17 |
Robin_Watts | tor8: I put it just post "as good as foxit" | 14:20.44 |
| tor8: ping | 14:53.58 |
tor8 | Robin_Watts: pong. | 15:20.17 |
Robin_Watts | tor8: Until I hear back from henrys, I'm assuming that the Roadmap is done. | 15:20.38 |
| and therefore I'm starting to help out. | 15:20.46 |
| I'm assuming that you're doing the html -> fz_buffer stuff? | 15:21.03 |
| if you've not started I'll start to look at it. | 15:21.17 |
| otherwise I'll try some simple code to gather line break distances and to guess paragraphs based on it. | 15:21.52 |
tor8 | Robin_Watts: add a fz_output struct and abstract the printing functions to go to buffers or FILE*, I can do that but I doubt I'll be able to get it done today | 15:22.05 |
| tomorrow for certain | 15:22.13 |
Robin_Watts | ok, I will do that then. | 15:22.14 |
| I don't think it's brain surgery ;) | 15:22.30 |
tor8 | no, it's just a chore fixing all the places that could currently use it | 15:22.50 |
henrys | kens:do you know offhand the ps->pdf behavior with blank pages, I want PCL to do something similar? - see http://bugs.ghostscript.com/show_bug.cgi?id=693608 | 16:28.33 |
| kens:with duplexing | 16:28.50 |
| Robin_Watts: the list looks like it is ready to be seen by Miles and Raph either one of you can copy them in or I will, let me know. | 16:30.28 |
Robin_Watts | henrys: I wasn't sure if we wanted to reveal timings to Raph | 16:30.54 |
| or whether Miles would like to convert them to $$$ beforehand. | 16:31.06 |
kens2 | henrys, sorry, wasn busy. ps->PDF should preserve blank pages no problem | 16:31.14 |
| I know PCL can 'remove' blank pages, PostScript never doess | 16:31.44 |
Robin_Watts | But if you're happy with the roadmap, I can forward it to Miles (or you can) | 16:31.49 |
henrys | no raph can be considered "one of us" | 16:32.00 |
kens2 | henrys I'm assuming that the PCL interpreter is eliding the blank pages, or am I wrong ? | 16:32.43 |
Robin_Watts | Is it worth delaying going back to Raph for a day or so to see if we can say "and here is an apk that shows it working"? | 16:33.03 |
| paulgardiner worked over the weekend to get us something. | 16:33.27 |
| We have a version that goes from pdf -> html and displays in the app in webviews. | 16:33.51 |
paulgardiner | We may have zoom in a minute. It's going better than expected | 16:34.03 |
Robin_Watts | Zooming the webviews doesn't work currently (he's battling that now...) | 16:34.08 |
| ah, fab. | 16:34.12 |
henrys | okay let's hold off until that is ready. | 16:34.29 |
Robin_Watts | And we don't have styles yet (but that's an output thing, I'm working on that now, and hope to have something to pass to paulgardiner in an hour or so) | 16:34.41 |
| But I'll send the roadmap to Miles. | 16:34.55 |
henrys | okay | 16:35.05 |
aksr | hi to all, does mupdf supports pdf annotation? | 16:35.14 |
Robin_Watts | We display them, if they are in the document and have an appearance stream. | 16:35.43 |
aksr | i know it's possible to see already annotated pdfs | 16:35.49 |
henrys | kens2:nto sure I didn't look at the job yet. | 16:36.06 |
Robin_Watts | We are working on code to allow creation and editing of them. | 16:36.11 |
aksr | Robin_Watts: nice to hear ;) | 16:36.19 |
Robin_Watts | The first couple of commits for that have gone in, but it's not complete yet. | 16:36.25 |
kens2 | henrys, I'm assuming that page 2 in the PCL is blank and being omitted, so pdfwrite never sees it, hence teh final PDF has only 3 pages | 16:36.46 |
aksr | any ETA for this? Robin_Watts | 16:36.48 |
Robin_Watts | aksr: Well, it was the next thing on our priority list, but it just got shunted down it a bit. | 16:38.35 |
aksr | Robin_Watts: ok, thank you | 16:38.48 |
Robin_Watts | being as this is open source though, you can always pick up what we've got an run with it. | 16:39.16 |
henrys | Robin_Watts:can you keep all of us on the recipient list or just use tech | 16:41.45 |
| ? | 16:41.52 |
Robin_Watts | I used tech | 16:41.54 |
henrys | paulgardiner: gets tech? | 16:42.13 |
Robin_Watts | I figured that keeping everyone up to date was worthwhile. | 16:42.17 |
| oh, I assumed he did, but... | 16:42.24 |
| Shall I mail marcosw and ask him to add paul to tech@ | 16:42.54 |
| ? | 16:42.58 |
paulgardiner | Don't think I do | 16:43.14 |
Robin_Watts | I'll resend with paul added as a new recipient. | 16:43.42 |
fdncred | Robin_Watts: FYI - Been testing mutool clean -ggg <inpdf> <outpdf> <pagenum> to split PDF files. It's doing great at keeping the annotations however.... | 16:53.21 |
Robin_Watts | fdncred: fab. | 16:53.39 |
fdncred | Robin_Watts: however, i noticed that it removes the bookmarks. any tips on maintaining them per page? | 16:53.55 |
Robin_Watts | by bookmarks do you mean "outlines" ? | 16:54.09 |
fdncred | ya | 16:54.13 |
Robin_Watts | In order to do that, we'd need to walk the outlines structure culling things off the page. | 16:55.48 |
| We don't currently do that. | 16:55.56 |
| I guess we'd like to handle outlines better, but I can't see us doing it any time in the next few months. | 16:57.49 |
fdncred | alternatively - i could see enumerating them, storing that info in a text file, and then i could insert them myself if there were a way to insert outlines with mupdf | 16:58.57 |
paulgardiner | Here's the version with zooming: http://intranet.glidos.net/~paul/MuPDFReflow20130204b.apk | 17:01.45 |
kens2 | Robin_Watts : ping (in the absence of a Marcos) | 17:03.10 |
Robin_Watts | pong | 17:03.19 |
kens2 | I'm trying to queue a cluster test, and it seems to be ignoring me | 17:03.31 |
Robin_Watts | paulgardiner: Excellent. Will try in a mo. | 17:03.32 |
| Using git cluster ? | 17:03.39 |
kens2 | Yes | 17:03.44 |
| usr/bin/gitpush.sh | 17:03.59 |
| It worked fine thismorning, now, nothing.... | 17:04.07 |
| Where should I go look to ifnd the problem / | 17:04.20 |
Robin_Watts | reload the dashboard? | 17:04.22 |
kens2 | ? | 17:04.23 |
| Hm, OK I'll try gthat | 17:04.31 |
| In fact I'll use Chrome | 17:04.41 |
| And there it is, Firefox screwing up :-( Thanks Robin_Watts | 17:05.26 |
Robin_Watts | no worries. I like that kind of fix :) | 17:05.37 |
kens2 | I htink something is broken in the more recent builds of Firefox, this never used to happen | 17:06.14 |
Robin_Watts | paulgardiner: OK, I've got a first version of the stuff to allow output to buffers on line. | 17:13.18 |
fdncred | Robin_Watts: is there anyway to create bookmarks/outlines with mutool? | 17:17.00 |
Robin_Watts | no. | 17:17.39 |
fdncred | thank you | 17:17.48 |
| for answering my q's | 17:18.00 |
Robin_Watts | np. | 17:18.10 |
| We can't 'recognise' the structures to make. | 17:18.23 |
| marcosw: Can we add Paul to the tech@artifex.com list please? | 17:55.27 |
marcosw | sure. | 17:55.42 |
kens2 | is off, goodnight all | 17:56.43 |
Robin_Watts | night kens2 | 17:56.49 |
marcosw | Robin_Watts: btw, you are on the tech@artifex.com list as @wss.co.uk. Should I change to you @artifex.com email address? | 18:00.11 |
Robin_Watts | marcosw: Makes no difference to me - it all comes to wss in the end and gets filtered. | 18:00.54 |
ray_laptop | marcosw: I don't see why you assigned bug 693612 to me | 18:03.32 |
| marcosw: see my comments. | 18:03.41 |
henrys | hi ray_laptop I was down in your neck of the woods (Palm Springs) visiting my cousin. | 18:05.53 |
ray_laptop | henrys: good weather this weekend | 18:07.58 |
| at least about 70 miles west of PS | 18:08.35 |
| PS is a great place to visit (and play golf) but I wouldn't want to live there | 18:09.03 |
henrys | yes I like the tram they have there, it is supposed to be famous | 18:09.08 |
| too hot in summer? | 18:09.27 |
ray_laptop | henrys: did you go up (and play in the snow). That's an attaction for SoCal residents. Probably less so for you | 18:09.54 |
henrys | we did | 18:10.10 |
ray_laptop | henrys: you bet it's too hot in the summer. they will have a streak of days in the +120 degrees F | 18:10.43 |
mvrhel_laptop | we are going to PS for spring break in April to get out of the rain and clouds here and let the kids swim in the pool. probably do some desert hiking at Josuah tree | 18:10.46 |
ray_laptop | henrys: where it will cool down at night to 90 or so | 18:11.04 |
henrys | yes my cousin was telling me somedays you really have to stay in. | 18:11.30 |
| mvrhel_laptop: I liked it but I didn't see a lot they have 2 year old twins and I went to visit the kids. | 18:12.33 |
mvrhel_laptop | oh henrys I am going to out tomorrow | 18:12.34 |
henrys | I should change my IRC nick to Oh Henry | 18:13.01 |
mvrhel_laptop | I am off to Olympia with my sons class for a field trip | 18:13.21 |
henrys | mvrhel_laptop:okay | 18:13.26 |
| how the windows app coming? | 18:13.36 |
| not too hurry you just curious | 18:13.54 |
mvrhel_laptop | henrys: It is coming along ok. I am hoping to have something to show by the meeting | 18:13.56 |
| I think once I get through this UI issue, things will come together pretty easily | 18:14.26 |
| the UI issuing being finding the best UI control to get smooth scrolling and allowing me to dynamically update the flipview list | 18:15.02 |
| which is the structure that is in the windoze UI that I am trying to use | 18:15.22 |
| overall though it has not been too painful | 18:15.35 |
| I also need to work on the color/black and white thing for the new support customer | 18:15.51 |
| bug 693583 | 18:16.08 |
| ray_laptop: I am going to work on adding the check on this in the CMM and then hopefully hand it off to you for you to add the device to do the mapping | 18:16.40 |
ray_laptop | mvrhel_laptop: what do you mean by "add the device to do the mapping" ? | 18:22.00 |
mvrhel_laptop | ray_laptop: I thought you were going to add in a device that would map the CMYK buffers to gray | 18:22.26 |
| if needed | 18:22.29 |
| let me check the IRC logs... | 18:23.30 |
ray_laptop | mvrhel_laptop: OK. Right. That's pretty easy if we assume a "simple" CMYK->K mapping (a'la 100% BG 100% UCR) | 18:24.09 |
mvrhel_laptop | right | 18:24.13 |
marcosw | henrys: I'm looking at the email from Guillaume. I'll open a bug if necessary. | 18:27.04 |
henrys | marcosw:okay great thank you. | 18:27.31 |
ray_laptop | marcosw: Did you see my comment on bug 693612 ? | 18:27.59 |
marcosw | yes. I reassigned it to alexcher, per your suggestion. Don't know why I didn't' realize it wasn't a clist issue, working on too man different bugs at once while watching the Super Bowl :-) | 18:28.53 |
ray_laptop | marcosw: note, I tried ALL of the 32-bit build problem bugs with Windows and can't get any failures (which is good since 32-bit windoze is more common that 32-bit linux) | 18:29.09 |
| marcosw: thanks. It was a good game (even if SF lost) | 18:29.41 |
marcosw | ray_laptop: as I said in the bug report, it's an indeterministic problem that comes and goes under linux, so am not surprised it doesn't fail on windows. | 18:30.25 |
henrys | marcosw: do you have access to any HP printer with a duplexer? | 18:31.56 |
marcosw | henrys: as it happens I have access to four. Two at uni (where I am currently) and two at home. | 18:32.31 |
alexcher | marcosw: I've reproduced the problem on x86. | 18:32.36 |
marcosw | henrys: though none of them are particularly current models. | 18:32.52 |
| alexcher: great, maybe I should assign 693613 to you as well. | 18:35.06 |
henrys | can you send the pcl file associated with http://bugs.ghostscript.com/show_bug.cgi?id=693608 and just describe the output in particular which pages/sides are blank? | 18:35.15 |
| he made obfuscated the file so it won't be attractive to look at | 18:36.07 |
| s/made// | 18:36.14 |
alexcher | marcosw: I need to reproduce it first. | 18:36.21 |
ray_laptop | alexcher: which problem were you able to reproduce on Windows (or is it just x86 32-bit on linux) ? | 18:38.07 |
alexcher | ray_laptop: I didn't try Windows. The bug 693613 is reproduced on x86 Linux but worx fine on ppc32. | 18:39.40 |
ray_laptop | alexcher: I see. That corresponds to what I've seen. It fails on peeves (which is i7) | 18:40.31 |
marcosw | henrys: how long is that document!?! I stopped it after ~20 pages, is that enough? | 18:40.35 |
alexcher | ray_laptop: -dNOGC helps. | 18:40.48 |
ray_laptop | alexcher: helps reproduce the problem, or avoid the problem ? | 18:41.23 |
marcosw | henrys: do you want me to summarize the duplex test output in a bug comment? | 18:41.27 |
henrys | marcosw:oh sorry - that should be enough just the first five page - at least according to his description we should see a problem in the first few pages. | 18:41.30 |
| marcosw:exactly. | 18:41.40 |
alexcher | ray_laptop: -dNOGC helps to avoid the problem. | 18:41.46 |
ray_laptop | alexcher: not too surprising given that the SEGV call stack seems to often be in the GC (for some of the bugs, anyway). Not the case with bug 693592, however | 18:43.41 |
marcosw | henrys: I've updated 693608 | 18:48.06 |
henrys | marcosw:thank you | 18:53.18 |
marcosw | henrys: np | 18:53.23 |
| seems to be the day for P1 customer bugs... | 19:21.12 |
ray_laptop | henrys: I think a support contract would not allow us to provide the PCL fonts, would it? They'd need an actual OEM contract | 19:59.00 |
henrys | ray_laptop:hmm hadn't really thought about it. I'll make scott aware of the issue and he can discuss it with miles if need be. | 20:00.45 |
| what is happening - product name AFPL Ghostscript | 20:11.19 |
| ? | 20:11.25 |
| marcosw:looks like something wrong with the cluster | 20:41.30 |
marcosw | I'll take a look. | 20:41.40 |
henrys | marcosw:humph just got my report but when I looked everything was on standby | 20:44.17 |
marcosw | your macpro was stuck, so i manually aborted it and the jobs it had been sent were redistributed to the other nodes. Since it hadn't received any jobs that meant the cluster job finished immediately. | 20:45.14 |
| I've been meaning to add code to the cluster to detect this case; if one machine get's stuck and hasn't received any jobs just abort it. | 20:45.41 |
| the cluster already sort of does this, but the logic is pretty simplistic and relies on pretty long time outs. | 20:46.50 |
mvrhel_laptop | bbiaw | 21:42.11 |
tor8 | Robin_Watts: sorry, had to pop out for a bit there. I see things have progressed nicely with paulgardiner and the zoom. | 22:29.12 |
| Robin_Watts: should probably name the file stm_output.c and also add a raw write function to the struct | 22:33.40 |
| other than that, patch looks good | 22:33.54 |
Robin_Watts | tor8: ok. thanks. | 22:45.43 |
tor8 | Robin_Watts: just rename the file and it's okay to commit. adding a write() function can be done later when we need it. | 22:46.32 |
| I'm thinking we'd want to use this for the pdfwrite/xpswrite devices too | 22:46.53 |
| where we need to write binary data and have a tell() function as well | 22:47.12 |
Robin_Watts | I started on the xpswrite device a while ago., | 22:48.05 |
| xpswrite we need to write zipfiles. | 22:48.15 |
| I was using fz_printf_buffer in those, IIRC. | 22:48.47 |
| Forward 1 day (to 2013/02/05)>>> | |