| | <<<Back 1 day (to 2012/12/12) | 2012/12/13 |
| JakeSays | anybody know which source file[s] implement the text writer device? | 04:46.57 |
| kens | JakeSays : gdevtxtw.c | 08:17.36 |
| Robin_Watts | tor8: Morning | 10:58.59 |
| tor8 | Robin_Watts: hey | 11:00.41 |
| Robin_Watts | Turns out that the places where I needed to pass the transformed text bbox were just after the place I'd just changed, so it was an easy fix. | 11:01.24 |
| | There are therefore now 3 patches up for review :) | 11:01.35 |
| | paulgardiner: There are 2 patches of yours up for review, right? Looking now. | 11:06.52 |
| paulgardiner | ta | 11:06.58 |
| Robin_Watts | paulgardiner: OK. In The refactoring of reset fields... | 11:15.14 |
| | You write 'Exclude' to the dictionary sometimes. | 11:15.36 |
| | Then check for that in add_field_hierarchy_to_array | 11:15.57 |
| | Then remove it again. | 11:16.01 |
| paulgardiner | yeah. That's how it worked before, but it was NoReset instead of Exclude and I wasn't putting the result in an array | 11:17.01 |
| Robin_Watts | I'm wondering if you still need to write to the array/delete it later. | 11:17.26 |
| paulgardiner | Sorry, don't understand | 11:18.14 |
| Robin_Watts | "form" is an array of entries. | 11:19.04 |
| | Some of which you process, others of which you don't. | 11:19.21 |
| paulgardiner | It's a hierarchy of feilds which makes it more complicated, but yes... | 11:20.01 |
| tor8 | Robin_Watts: *nod* | 11:20.14 |
| paulgardiner | Hi Tor | 11:20.23 |
| Robin_Watts | Rather than writing the details of whether to process them or not into form and then rubbing those details out later, why not keep that information in a separate array? | 11:20.25 |
| | Not changing the PDF objects would be preferable, I think, right? | 11:21.13 |
| paulgardiner | I'd need to look up in the array rather than just test whether a fields is marked | 11:21.14 |
| | I don't really like marking them, but it seemed the most efficient | 11:21.38 |
| | I could use an ordinary array and binary search I suppose. | 11:22.18 |
| Robin_Watts | Oh, I see. You'd need a seach in forms within the fields loop (or vice versa) | 11:22.58 |
| paulgardiner | Yeah | 11:23.13 |
| Robin_Watts | OK. I follow. That makes sense. | 11:23.14 |
| | OK, both of those look good. | 11:24.01 |
| | tor8: Thanks. I'll push paulgardiners commits and mine then. | 11:24.12 |
| paulgardiner | Thanks Robin | 11:24.25 |
| | tor8, Robin_Watts: I need some icons for the file picker, directory and one for each file type we support. Any thoughts? | 11:43.29 |
| Robin_Watts | What format? PNG? | 11:52.04 |
| | and what res? | 11:52.48 |
| paulgardiner | yeah png. We have 24x24 and 32x32 versions for the others we use, but not so clear these want to be square | 11:54.24 |
| tor8 | paulgardiner: probably rectangular document shaped with the acrobat logo (or maybe not?) and the xps blue paper plane logo | 11:55.05 |
| kens | I wouldn't use acrobat logo | 11:55.16 |
| | trademarked I think | 11:55.23 |
| Robin_Watts | The 3 lobed red loop? | 11:55.34 |
| kens | Robin_Watts : yes | 11:55.50 |
| tor8 | kens: yeah. that's what gives me hesitation. but it's used in gnome file browsers so that's why I mentioned it. | 11:55.52 |
| kens | understood, but Adobe might choose not to sue Gnome, we might be a bit different | 11:56.33 |
| | we'd have to show the logo had become generic. Of course, other people using ti would help | 11:57.20 |
| tor8 | paulgardiner: http://somerandomdude.com/work/iconic/# scavenge there for placeholders (same set as our toolbar icons) | 11:57.30 |
| | there's a generic document and a folder icon we could use | 11:58.13 |
| paulgardiner | Oh ok, that'll do nicely for now. | 11:58.32 |
| tor8 | and maybe the up arrow for going up a level | 11:58.37 |
| | I'd suggest to pick the black ones, for the "light" theme I switched the file browser to | 11:59.16 |
| paulgardiner | Right | 12:00.58 |
| Robin_Watts | tor8: OK. The PDF cmap stuff is good to go on my repo, I think. | 12:03.22 |
| tor8 | Robin_Watts: have you tried compiling it? ;) | 12:05.31 |
| | also, I'd suggest dropping the (%d/%d) part so the warning rate limiter can kick in to reduce the spammage when generating files | 12:06.17 |
| Robin_Watts | picky picky picky :) | 12:06.18 |
| tor8 | Robin_Watts: check tor/master for revised version | 12:08.57 |
| | with better warning messages | 12:09.02 |
| Robin_Watts | ah, I've just revised mine. Let's see how they compare... | 12:09.38 |
| | ok, let's go with yours. | 12:10.48 |
| tor8 | weee! \o/ I win! | 12:11.28 |
| | Ahem. I'm still jet lagged. :) | 12:11.38 |
| | ... and I have a cold coming on :( sore throat. | 12:11.53 |
| Robin_Watts | Pushed. | 12:14.05 |
| | Do we have a 'deep compare object' function? | 12:20.11 |
| | pdf_objcmp | 12:20.55 |
| | tor8: use after free fix on my repo. Tests running now. | 12:29.20 |
| | tor8: And PNG image fuzzing fixes. | 12:56.20 |
| | tor8: And a null deref fix. | 13:00.50 |
| | tor8: yeah, jet lag is killing me too. Thankfully no cold (yet) | 13:01.12 |
| JakeSays | kens: thanks | 13:43.15 |
| kens | NP | 13:43.20 |
| Robin_Watts | tor8: And another patch moving TR2 detection. | 14:19.25 |
| paulgardiner | Where can I find some .cbz files? | 15:20.47 |
| Robin_Watts | looking... | 15:22.17 |
| | http://www.contrapositivediary.com/?p=1197 | 15:25.12 |
| | tor8: Another patch on robin/master | 15:26.56 |
| kens2 | thinks Robin_Watts is grump today | 15:27.23 |
| Robin_Watts | mmm? | 15:27.37 |
| kens2 | reply to gs-bugs about ligatures | 15:27.51 |
| | Normally you are more polite I htink | 15:28.03 |
| | even to idiots | 15:28.06 |
| Robin_Watts | Well, what sort of stupid bloody bug report is that? "It's a blocker!" with no example file, or even a command line. | 15:28.34 |
| | I wasn't trying to be rude, just direct. | 15:29.13 |
| kens2 | Oh I agree, just that it seemed lmore like a reply I would write | 15:29.15 |
| paulgardiner | Ah no. It's some XPS examples that I need. | 15:32.12 |
| Robin_Watts | paulgardiner: Have you got an svn copy of our tests_private stuff yet? | 15:34.03 |
| paulgardiner | I have what you stuck on a Bluray | 15:35.01 |
| Robin_Watts | ok, so no. | 15:35.11 |
| | http://www.ecma-international.org/publications/standards/Ecma-388.htm | 15:35.50 |
| | That's got an xps doc on it. | 15:36.01 |
| paulgardiner | ta | 15:36.39 |
| Robin_Watts | http://msdn.microsoft.com/en-gb/library/windows/hardware/gg463429.aspx and that has more. | 15:37.10 |
| | Do you have a portable HD? | 15:37.26 |
| | At some point I should copy my svn checkout for you. | 15:37.45 |
| paulgardiner | Robin_Watts: that would be handy. Once I have it, provided I maintain the time stamps when copying it, I should be able to keep it up to date. | 15:42.01 |
| | I have a spare 1/2 GB disc here. I should bring it over to your sometime. | 15:42.29 |
| Robin_Watts | even without timestamps, I believe. | 15:42.34 |
| tomtaylor | hello all - has anyone got a second to help me work out why gs is rastering some text in a PDF to PDF conversion? | 15:42.44 |
| Robin_Watts | tomtaylor: Sure. I suspect the answer is that you are asking for a version of PDF that doesn't contain transparency. | 15:43.26 |
| kens2 | WHat version of GS tomtaylor ? | 15:43.31 |
| tomtaylor | 9.06 | 15:43.40 |
| kens2 | Robin_Watts : there can be other reasons | 15:43.45 |
| tomtaylor | Robin_Watts: that's what I suspected | 15:43.46 |
| kens2 | tomtaylor : what command line are you using ? | 15:44.00 |
| paulgardiner | Robin_Watts: I just managed to open that XLS file in MuPDF from ES File Explorer. The only oddity was how many different programs offered to open it, thus making MuPDF way down on the list | 15:44.03 |
| kens2 | and cna you post the PDF somethwere so I cna look at it please# | 15:44.14 |
| Robin_Watts | kens2: Sure, but that's my first guess. | 15:44.15 |
| | paulgardiner: I really hope you mean xps :) | 15:44.27 |
| paulgardiner | :-) I do | 15:44.37 |
| Robin_Watts | What was wrong with the intent line? | 15:44.47 |
| kens2 | you mena MuPDF soesn't open excel spreadhseerts ? :-) | 15:44.49 |
| Robin_Watts | kens2: Hey, we were aware for the weekend. Maybe paulgardiner got bored... | 15:45.14 |
| paulgardiner | Must work on that. | 15:45.16 |
| tomtaylor | -sDEVICE=pdfwrite -dPDFX -dCompatibilityLevel=1.4 -sProcessColorModel=DeviceCMYK -dColorConversionStrategy=/LeaveColorUnchanged -dPDFSETTINGS=/prepress -dHaveTransparency=true -dSubsetFonts=false -dEmbedAllFonts=true -dNOPAUSE -dQUIET -dBATCH -P | 15:45.17 |
| | and some other image resolution options | 15:45.28 |
| kens2 | tomtaylor : that 'looks' OK if rather verbose | 15:45.47 |
| tomtaylor | would -dPDFX take priority over -dHaveTransparency? | 15:45.47 |
| kens2 | Yes it does, and that means no transparency | 15:46.06 |
| Robin_Watts | I win! \o/ | 15:46.19 |
| kens2 | Its forbidden in the PDF/X spec | 15:46.22 |
| Robin_Watts | is just as jetlagged as tor8 :) | 15:46.28 |
| tomtaylor | yep | 15:46.31 |
| paulgardiner | Robin_Watts: the problem is that I managed to open the XPS without making any changes to the app. :-( | 15:46.31 |
| kens2 | Robin_Watts : still no idea if the original has transparency | 15:46.37 |
| Robin_Watts | paulgardiner: oh. | 15:46.41 |
| tomtaylor | the thing is: I don't think the original has transparency | 15:46.49 |
| paulgardiner | Robin_Watts: maybe the problem occurs only with some file managers | 15:46.51 |
| kens2 | It could be a CIDFont problem | 15:46.51 |
| | tomtaylor : So I need to see the original | 15:47.04 |
| tomtaylor | one second | 15:47.11 |
| Robin_Watts | paulgardiner: I was using ES file explorer. | 15:47.16 |
| paulgardiner | It did look like MuPDF wasn't in the list at first because the list was so long | 15:47.21 |
| | That's what I have too | 15:47.34 |
| kens2 | also just because you can't see any transparecny doesn't mean it isn't there, Cairo sets *everything* in a transparenmcy group, even if its opaque.... | 15:47.53 |
| Robin_Watts | paulgardiner: I just get "open as: Text Audio Video Image" | 15:48.19 |
| paulgardiner | Weird. So it's device specific | 15:48.53 |
| tomtaylor | kens2: http://files.tomtaylor.co.uk.s3.amazonaws.com/data/history.pdf | 15:49.08 |
| kens2 | fetching now | 15:49.14 |
| Robin_Watts | That's on my phone. I believe I get the same on the transformer. | 15:49.17 |
| tomtaylor | that is coming out of an OSX PDFContext | 15:49.25 |
| | it warns about the broken xref table | 15:49.39 |
| kens2 | not a good start | 15:49.48 |
| paulgardiner | Robin_Watts: could it be that an app has become the default viewer for XPS on your device? Seems unlikely | 15:50.06 |
| tomtaylor | i don't have much control of the output from the PDFContext, so I'm using Ghostscript to clean the file up and optimise image resolution for printing | 15:50.14 |
| ademmler | Hi, I am new to use an irc channel, so please be patient. | 15:50.27 |
| Robin_Watts | I get "file open failed" from the File explorer on the transformer. | 15:50.45 |
| ademmler | Chris form artifex recommended me to join in order to gain infos about "screening in ghostscript" .... | 15:50.58 |
| Robin_Watts | ademmler: Welcome. We'll be gentle. | 15:51.08 |
| paulgardiner | Maybe if I'd downloaded the file via the browser it would fail | 15:51.10 |
| kens2 | tomtaylor : which of the (12!) pages, and where, do you see a problem ? Its too big to rnadomly search | 15:51.11 |
| ademmler | Who can point me to example, docu etc .... | 15:51.15 |
| tomtaylor | page 2 | 15:51.17 |
| | is entirely rastered | 15:51.22 |
| ademmler | I like to get accurate screening in tiffsep1 device :-) | 15:51.29 |
| kens2 | tomtaylor : OK let me look at that page | 15:51.43 |
| ademmler | thx to Robin_Watts | 15:51.48 |
| chrisl | ademmler: accurate screening has a specific meaning in Postscript, is that what you're after? | 15:52.01 |
| ademmler | yep ecatly .. or something similar | 15:52.16 |
| | I want to evaluate GS for doing tiffg4 files for offset printing | 15:52.29 |
| chrisl | Well, AccurateScreens is a Postscript user parameter | 15:52.42 |
| ademmler | I need a Mac OS X or Linux RIP .... most vendors offer just windows | 15:52.45 |
| | Is accurate screening honored by GS? | 15:52.59 |
| chrisl | <shrug> Like I said on the bug - not my area..... | 15:53.25 |
| ademmler | I am slo interested in stochastic methods ... | 15:53.32 |
| | I ment "also" not "slo" | 15:53.49 |
| Robin_Watts | paulgardiner: OK, I've just installed the latest version of ES file manager on the transformer, and as far as I can see, it offers EVERY app on the device for .xps | 15:54.52 |
| | That rather suggests to me that it has failed to find a matching intent. | 15:55.23 |
| | Let me try updating the ES file explorer on my phone. | 15:55.32 |
| chrisl | ademmler: You're probably a little early for folks who know the most about screening in GS, they are in the US | 15:56.03 |
| ademmler | ok, so I try again later. Are they west or east cost? | 15:58.00 |
| Robin_Watts | west | 15:58.17 |
| ademmler | thx guys ..... | 15:58.26 |
| kens2 | ? | 15:58.26 |
| chrisl | ademmler: btw, I believe we do honour the AccurateScreens user parameter, but I've no idea *how* we honour it | 15:58.27 |
| Robin_Watts | paulgardiner: Yeah, new version does the same on my phone. | 15:58.28 |
| ademmler | have fun and a nice "advent" | 15:58.36 |
| chrisl | Cheers! | 15:58.54 |
| Robin_Watts | ademmler: and you. | 15:59.00 |
| ademmler | @chrisl: I rmember this also, but I cant find the options in the docu anymore ... | 15:59.16 |
| | However I let you know how far I come .... | 15:59.25 |
| | bye to all:-) | 15:59.32 |
| kens2 | tomtaylor : its definitely the PDFX that causes it | 15:59.54 |
| chrisl | Hmph, got Postscript on the brain - I keep typing "exch" instead of "echo" in a shell script...... | 16:00.28 |
| kens2 | I suspect its because you have used a Multiple Master font | 16:00.38 |
| tomtaylor | kens2: interesting - is that out of PDFX spec? | 16:01.16 |
| kens2 | Not sure, I'm still investigating | 16:01.30 |
| | Now to the debugger | 16:01.34 |
| tomtaylor | kens2: thank you, much appreciated | 16:01.44 |
| kens2 | Having reduced the file to something manageable | 16:01.52 |
| tomtaylor | kens2: do you know why that would only be happening on page 2? is there a specific use on that page that triggers it? | 16:04.25 |
| kens2 | DOn't know yet | 16:04.33 |
| | well its not getting to text processing so its being rasterised at a higher level | 16:06.01 |
| paulgardiner | Robin_Watts: OI File Manager says "Application not available" | 16:07.42 |
| tomtaylor | kens2: the images in the PDF are cropped, but the cropped area hasn't been removed from the file, so it's conceivable that the uncropped image area touches some of the text | 16:09.18 |
| kens2 | Doesn't matter | 16:09.27 |
| | Well, if I remove some more stuff, it stops getting rendered | 16:11.04 |
| | Soits something in the content stream | 16:11.25 |
| | need to binary chop to see what | 16:11.35 |
| | Ah, it is one of the images | 16:12.12 |
| tomtaylor | kens2: is there something wrong with it, or does it look like a bug? | 16:13.00 |
| kens2 | I doubt its a buig, nor is there necessarily anything 'wrong' wiht the image | 16:13.33 |
| | But I need to find outr hwich one and look at it | 16:13.34 |
| | aseems to be the lower image | 16:14.04 |
| tomtaylor | the lower one has no filter | 16:14.14 |
| | the top one is zlib/deflate | 16:14.21 |
| chrisl | Does the image have a softmask? | 16:14.35 |
| kens2 | Giver me a minute | 16:14.43 |
| Robin_Watts | paulgardiner: OK, no useful debug from Asus file manager. | 16:15.00 |
| | paulgardiner: I guess the thing to do is to pare mupdf back to see what is triggering on pdf files. | 16:15.24 |
| | is it triggering on name, or mimetype, or both etc? | 16:15.35 |
| | Then to see if just putting .xps in works (i.e. if only 1 intent thing is being accepted) | 16:16.01 |
| paulgardiner | Yeah. I'll fiddle | 16:16.02 |
| Robin_Watts | tor8: ping | 16:16.14 |
| kens | tomtaylor : your file does indeed contain transparency. As chrisl says, the image contains a soft mask: | 16:17.09 |
| | 18 0 obj | 16:17.09 |
| | << | 16:17.09 |
| | >> | 16:17.12 |
| | Duh, irc doesn't like '/' | 16:17.24 |
| tomtaylor | oh, interesting | 16:17.38 |
| kens | "18 0 obj | 16:17.49 |
| | "<< | 16:17.49 |
| | " /BitsPerComponent 8 | 16:17.49 |
| | " /ColorSpace 20 0 R | 16:17.49 |
| | " /Width 640 | 16:17.49 |
| | " /Height 360 | 16:17.49 |
| | " /Interpolate true | 16:17.49 |
| | " /Length 921600 | 16:17.50 |
| | " /SMask 17 0 R | 16:17.50 |
| | " /Subtype /Image | 16:17.51 |
| | " /Type /XObject | 16:17.51 |
| | ">>" | 16:17.52 |
| | PDF/X doesn't allow tyransparency so we have to render it, thus your text is rasterised. | 16:18.11 |
| tomtaylor | i think I know why... it originally came from a GIF | 16:18.20 |
| | it is possible to tell ghostscript to just flatten the image, not the entire page? | 16:18.53 |
| kens | tomtaylor : No | 16:18.58 |
| Robin_Watts | tomtaylor: That could be represented as an image + imagemask. | 16:19.11 |
| | i.e. without PDF 1.4 transparency stuff. | 16:19.24 |
| tomtaylor | Robin_Watts: yep | 16:19.35 |
| Robin_Watts | So it's down to the original creator being less than brilliant. | 16:19.39 |
| tomtaylor | oh, I know :) | 16:19.44 |
| | it's coming out of PDFContext in Quartz, part of OS X's Cocoa stack | 16:20.02 |
| | well, I can try and remove the alpha component in the source images, and see whether it spits out a flat image in the PDF | 16:21.10 |
| | thank you everyone, I wouldn't have had a clue where to start diagnoising that | 16:22.23 |
| chrisl | ;-) "PDF: becuase the world was not confusing enough already......." | 16:23.11 |
| | kens: The Ubuntu/Toshiba printer problem seems to be either an LZWDecode bug, or a buggy interaction between the filter and the rest of the rip - another cr*p PS printer, I wonder who the rip vendor is.... | 16:25.15 |
| kens | sorry went to get a coffee | 16:26.17 |
| chrisl | Wasn't urgent.... | 16:26.30 |
| kens | chrisl well its not the first | 16:26.33 |
| | probably an old Xionics rip or something ;-) | 16:26.58 |
| chrisl | Well, this experience has told me only to buy a PS printer on sale-or-return, and to run *lots* of tests as soon as I get it! | 16:27.54 |
| kens | By now you probably have a nice selection that show up problems on particular interpreters :-) | 16:28.21 |
| Robin_Watts | chrisl: Is this a printer you've just bought? | 16:28.25 |
| kens | no its a Ubunutu bug report | 16:28.33 |
| Robin_Watts | ah. | 16:28.37 |
| kens | against ps2write | 16:28.46 |
| chrisl | Robin_Watts: no, it's a Ubuntu bug, after switching to ps2write for the PS filter | 16:28.55 |
| kens | Of course there is nothing wrong with the ps2write output | 16:28.56 |
| chrisl | Although with the number of applications emitting CIDFonts in their PDFs for the cups print queue, it would be good to find a way of getting a ps3write going..... | 16:31.06 |
| Robin_Watts | Coo. Color lasers have dropped in price: http://colour-laser-printers.co.uk/printer/1314B014AA/i-SENSYS%20LBP5360.php | 16:31.40 |
| kens | chrisl indeed it would, one day when I've done everything else.... | 16:31.55 |
| Robin_Watts | Color, Duplex and ethernet. | 16:31.58 |
| kens | Hmm, tempting | 16:32.14 |
| | I wonder what the per page cost is | 16:32.29 |
| | ah, 6000 page ink pack is £400 for a £230 printer.... | 16:33.00 |
| Robin_Watts | 86+VAT for 6000 pages (black toner), 113+vat for 6000 pages for each of C, M, Y | 16:33.01 |
| | ouch! | 16:33.07 |
| kens | At least inks can be bought separately | 16:33.33 |
| Robin_Watts | but given that most of my printing is black, that would mean 86+VAT for 6000 pages. | 16:33.42 |
| kens | has a teenage daughter at college, lots of colour pages.... | 16:34.02 |
| Robin_Watts | which is probably cheaper than 35ish quid for 2500 pages now. | 16:34.06 |
| | I suspect color costs are cheaper than inkjet too. | 16:34.30 |
| chrisl | Geez, a lot of printers advertise "manual duplex" - what a crock! | 16:34.56 |
| Robin_Watts | So, why wouldn't I throw the printer away when the cartridges run out, and just buy another one? | 16:39.27 |
| kens | cos it comes wiht a very small ink cartridge | 16:43.47 |
| Robin_Watts | kens: The brochure does not state that (I specifically looked) | 16:44.09 |
| kens | Robin_Watts : bet it does though | 16:44.24 |
| Robin_Watts | It has a footnote that suggests that some printers come with just 2000 sheets worth of toner. | 16:44.50 |
| kens | I owuld imagine that's the case | 16:45.03 |
| Robin_Watts | but I couldn't see that footnote applied to that printer, but I fear you're right. | 16:45.24 |
| tor8 | Robin_Watts: pong. | 16:59.32 |
| Robin_Watts | Hi tor8. Many patches up for review. All little though. | 16:59.46 |
| chrisl | I wonder if I can un-register myself from the Ubuntu launchpad bug tracker - let someone else deal with these loons for a while :-( | 16:59.58 |
| tor8 | Robin_Watts: | 17:01.38 |
| | -for (y = 0; y < info->height; y++) | 17:01.38 |
| | +for (y = info->height; y > 0; y--) | 17:01.39 |
| | not equivalent or am I still too jetlagged to think? | 17:01.40 |
| Robin_Watts | Perfectly equivalent as long as we don't use the value of y in the loop, I believe. | 17:02.07 |
| | The first counts from 0 to info->height-1. The second from info->height to 1. | 17:02.59 |
| tor8 | y will range 1..info->height in the latter, 0..info->height-1 | 17:03.07 |
| | in the first | 17:03.12 |
| | oh, we never use x and y as values? | 17:03.41 |
| Robin_Watts | such is my belief. | 17:03.52 |
| mvrhel_laptop | good morning | 17:04.07 |
| chrisl | mvrhel_laptop: good morning..... | 17:04.23 |
| mvrhel_laptop | chrisl: sounds like you are having fun.... | 17:04.26 |
| tor8 | I don't mind counting down, but it's a bit tricky having an "off" range of values there | 17:04.26 |
| chrisl | mvrhel_laptop: on the whole, I'd rather be strumming guitars with Miles and Scott ;-) | 17:04.47 |
| tor8 | (and I usually do the counting down loops with a while, rather than for, but that's just my personal habit) | 17:04.55 |
| mvrhel_laptop | yes | 17:04.59 |
| chrisl | mvrhel_laptop: there was a guy on here earlier wanting to talk about halftoning with Ghostscript - he said he would check back later | 17:05.29 |
| | His nick was "ademmler" | 17:06.03 |
| ray_work | chrisl: if he shows up, I can probably help as well | 17:06.04 |
| | chrisl: thanks | 17:06.09 |
| | chrisl: I'm reviewing the logs | 17:06.21 |
| Robin_Watts | I could see the argument for using a do { } while(); if I was sure that the condition held at the top of the loop first time. | 17:06.24 |
| tor8 | Robin_Watts: the batch LGTM, but unless you feel strongly I'd suggest for (y=height-1; y--; y >= 0) or s/y/some-other-variable-name/ | 17:06.42 |
| Robin_Watts | But otherwise I prefer a for to a while if all 3 components of the for are filled out. | 17:06.44 |
| tor8 | Robin_Watts: I never use do/while though! | 17:06.49 |
| chrisl | ray_work: yep, I figured either of you would be able to talk to him - but you popped just as I hit the return key! | 17:06.53 |
| Robin_Watts | I much prefer for (y=height; y >0; y--) to for (y=height-1; y >= 0; y--) | 17:07.19 |
| tor8 | n = count; while (n-- > 0) { ... } is my preferred construct for counting down | 17:07.20 |
| chrisl | Anyway, I'm off for a *double* session of squash training.... I'm really not sure I'm up for this :-( | 17:07.27 |
| tor8 | Robin_Watts: yeah, it looks a bit nicer. oh well, just go ahead and commit. | 17:08.06 |
| Robin_Watts | tor8: I see no advantage to that over a for, and it's less readable, IMHO. | 17:08.15 |
| | tor8: OK, thanks. | 17:08.22 |
| | Is that for all the commits ? | 17:08.26 |
| mvrhel_laptop | chrisl: sorry I stepped away for sec, but it looks like ray_work volunteered | 17:16.33 |
| | back to the transparency salt mine | 17:17.41 |
| kens | off now, goodnight all | 17:23.08 |
| ray_work | ademmler: (for the logs). Ghostscript doesn't support "AccurateScreens" (at least not well). What I would recommend for ordered dither (traditional screening with angle/freq and spot shape) is to use mvrhel's utility to generate threshold arrays with the desired parameters using a fairly large supercell size. See gs/toolbin/halftone/ordered_diither/README | 17:39.37 |
| Robin_Watts | tor8: Was your earlier nod for all the commits on my repo ? | 17:41.01 |
| tor8 | up to and including "X11 app startup" | 17:41.22 |
| Robin_Watts | fab, thanks. | 17:41.33 |
| ray_work | ademmler: this can generate a PostScript ThresholdArray type of output. For stochastic screening, there is a sample stochastic threshold array shipped with gs (see gs/lib/stocht.ps). This is a 167x167 mask that is just an example. | 17:42.28 |
| | ademmler: The toolbin/halftone/gen_stochastic/gen_stochastic.c program (only documentation is in the source, sorry) is my program that was used to generate the 167x167 and can be used to generate any size stochastic array (large may take a while). | 17:44.00 |
| | ademmler: This generates a 'turn on sequence' designed to be run through the toolbin/halftone/threshold_remap/threshold_remap.c program to linearize and convert to PS. | 17:45.30 |
| | ademmler: since large stochastic arrays have so many levels intrinsically, the threshold_remap lets you take the TRC (tone response curve) into account and make nice linear 256 level threshold arrays. | 17:47.55 |
| Robin_Watts | tor8: And one more patch, again from zeniko. | 17:57.07 |
| | Someone should explain that if he keeps adding new fixes to a bug, he'll never get the bounty paid out :) | 17:57.31 |
| | tor8: What's our policy on changes to libopenjpeg etc? | 19:01.08 |
| | Can we just commit them back into our git submodule repo? | 19:01.18 |
| tor8 | Robin_Watts: put them on the artifex branch on our submodule repo | 19:21.16 |
| Robin_Watts | tor8: That's the branch the submodule checks out automatically? | 19:21.42 |
| tor8 | Robin_Watts: nope, the submodule needs to be pointed at a sha1 | 19:21.58 |
| Robin_Watts | ah, I see. Thanks. | 19:22.12 |
| | tor8: 1 last review online. | 19:22.24 |
| Robin_Watts | pops to shop/ | 19:22.39 |
| tor8 | so, git checkout artifex; edit; git commit; cd ..; git add thirdparty/libopenjeg (without trailing /); git commit | 19:22.42 |
| | you can double check with git status that it's updating the sha1 for the submodule | 19:23.10 |
| | Robin_Watts: *nod* to "eek" patch | 19:24.21 |
| mvrhel_laptop | arg | 19:43.31 |
| Robin_Watts | tor8:Thanks | 19:45.04 |
| | mvrhel_laptop: mmm? | 19:45.11 |
| mvrhel_laptop | oh Robin_Watts you are still here | 19:45.23 |
| | so I was fooling with my Ubuntu install again | 19:45.40 |
| Robin_Watts | ok... | 19:46.20 |
| mvrhel_laptop | and just discovered that I dont have the clusterkey in my .ssh directory | 19:46.22 |
| | trying to find where it came from. Is that something marcos would have given me | 19:46.40 |
| Robin_Watts | possibly. just a tick. | 19:47.04 |
| | on casper. | 19:47.42 |
| | /home/mvrhel/.ssh/cluster_key | 19:47.55 |
| mvrhel_laptop | ok thanks | 19:48.00 |
| Robin_Watts | np. | 19:48.15 |
| | casper is ubuntu so that should be in the right format | 19:48.33 |
| mvrhel_laptop | wtf. I am getting the stupid (public key) issue now trying to get to casper. I thought I had that resolved | 19:51.43 |
| | sigh | 19:51.45 |
| | oh wrong user | 19:52.01 |
| Robin_Watts | any luck? | 19:53.48 |
| mvrhel_laptop | almost there... | 19:53.54 |
| | well. now I get an rsync issue | 19:55.34 |
| Robin_Watts | whilst attempting to push ? | 19:55.54 |
| mvrhel_laptop | yes | 19:55.57 |
| | is there a possible user issue there too? | 19:56.05 |
| | I am michaelv locally | 19:56.12 |
| | and mvrhel else where | 19:56.19 |
| Robin_Watts | export CLUSTER_USER=mvrhel | 19:56.24 |
| mvrhel_laptop | aha | 19:56.29 |
| | ok now it is working | 19:57.30 |
| | I think what I will do is create a new mvrhel account on here. thanks for helping Robin_Watts | 19:58.13 |
| Robin_Watts | mvrhel_laptop: Or just add export CLUSTER_USER=mvrhel to your .profile file. | 19:58.36 |
| mvrhel_laptop | I will do that now | 19:59.15 |
| Robin_Watts | Anyone understand the internals of openjpeg ? | 19:59.28 |
| mvrhel_laptop | not me sorry | 20:02.47 |
| | off to lunch while the cluster chugs. | 20:02.58 |
| Robin_Watts | will look more at this tomorrow. | 20:03.09 |
| henrys | torn, Robin_Watts:were you guys going to help out zeniko with mupdf ;-) | 20:24.15 |
| | tor8 ^^^ | 20:24.25 |
| | seriously we should figure out fair bounty compensation | 20:25.24 |
| mvrhel_laptop | hmm. That is a odd log output | 20:40.57 |
| | henrys: you there? | 20:41.44 |
| henrys | yes | 20:41.58 |
| mvrhel_laptop | can you click on my user name on the regression dashboard | 20:42.12 |
| | what does it mean log file too big? | 20:42.54 |
| henrys | is it just coming from inches? | 20:43.44 |
| mvrhel_laptop | yes | 20:43.48 |
| | and what does this mean | 20:44.00 |
| | timeoutFail: log file(s) too big: | 20:44.03 |
| | 28 temp/tests_private__comparefiles__055-01.ps.ps.ppmraw.300.0.log | 20:44.04 |
| | 28 temp/tests_private__comparefiles__Bug689815.pdf.pdf.ppmraw.300.0.log | 20:44.07 |
| | 28 temp/tests_private__comparefiles__Bug690559.pdf.pdf.ppmraw.300.0.log | 20:44.08 |
| | 28 temp/tests_private__comparefiles__Bug690559.pdf.pdf.ppmraw.72.0.log | 20:44.09 |
| | 40 temp/tests_private__comparefiles__NECPNTD.pdf.pdf.ppmraw.300.0.log | 20:44.11 |
| | 40 temp/tests_private__pdf__sumatra__1479_-_does_not_render.pdf.psdcmyk.300.1.log | 20:44.13 |
| | 48 temp/tests_private__comparefiles__0.pdf.ps.pkmraw.300.0.log | 20:44.15 |
| | 120 temp/tests_private__pdf__sumatra__692248_-_wrong_ObjStm_ref.pdf.plank.72.0.log | 20:44.17 |
| | 20756 temp/tests_private__comparefiles__Bug690637.pdf.plank.300.1.log | 20:44.19 |
| | 118800 temp/tests_private__pdf__sumatra__x_-_renders_slowly.pdf.pbmraw.300.0.log | 20:44.21 |
| | in its last gasp | 20:44.22 |
| | 118800 | 20:44.27 |
| | tests_private__pdf__sumatra__x_-_renders_slowly.pdf | 20:44.35 |
| | is that likely the problem/ | 20:44.43 |
| | ? | 20:44.44 |
| | I don't see how my change could have caused a problem here, but I can take a look at the file and see | 20:45.16 |
| henrys | I'd just report it to marcosw | 20:45.33 |
| mvrhel_laptop | ok | 20:45.38 |
| henrys | but I guess you are stuck until he fixes it. | 20:45.50 |
| mvrhel_laptop | right well let me do a clean push to see if that does it too for me | 20:46.11 |
| henrys | you can also disable inches and go again perhaps inches in low on disk space. | 20:46.54 |
| mvrhel_laptop | that might be better | 20:47.06 |
| | let me do that | 20:47.08 |
| marcosw | a potential customer asks if pdfwrite is threadsafe. Does anyone know? (other than ken, who is presumably sleeping) | 22:01.07 |
| ademmler | Chris Lidl recommended me to join into this IRC in order to get infos about "screening in ghostscript" .... | 23:06.53 |
| | Is somebody here, who is faimilar with it? | 23:07.08 |
| | sotty forgot tosay HI to all | 23:07.20 |
| | it is my second time in an IRC - so in this I am a newbie | 23:07.37 |
| | Forward 1 day (to 2012/12/14)>>> | |
IRC Logs