| | <<<Back 1 day (to 2014/02/10) | 2014/02/11 |
| R4v3n_ | Hi, anyone out there who can help with a ghostscriptAPI problem, for the .net framework .. Id really appreciate the help | 05:03.55 |
| | Hi, anyone out there who can help with a ghostscriptAPI problem, for the .net framework .. Id really appreciate the help thanx | 05:06.43 |
| ray_laptop | R4v3n_: The person that works on ghostscript.net is here occasionally -- Check the logs. I think the developer signs on here as "jhabjan" | 05:26.43 |
| | R4v3n_: as far as general gs API issues, feel free to ask here | 05:27.43 |
| R4v3n_ | thanx ray_laptop i'll keep a look out for jhabjan | 05:28.26 |
| ray_laptop | R4v3n_: and as long as your question isn't a total noob "how do I use this" telling us that you haven't even read the docs, we'll try to help ;-) | 05:28.56 |
| R4v3n_ | lolz sure | 05:29.09 |
| | im having trouble with the getimages function when i access it thorugh a generic handler (.ashx) it throws an exception | 05:29.51 |
| ray_laptop | and, of course if it is specific to something that someone else developed (like the .NET interface) we can help | 05:29.54 |
| | R4v3n_: sounds like something that is a .NET function since there is no "getimages" for ghostscript | 05:30.45 |
| R4v3n_ | the exception is "failed to process ghostscript command" error code :100 which is a generic fatal error | 05:30.46 |
| ray_laptop | R4v3n_: so either find the .NET developer somewhere else, or wait 'til they show up here | 05:31.38 |
| R4v3n_ | i'll do that | 05:31.52 |
| ray_laptop | R4v3n_: we can only help if you have the ghostscript command that fails and the input file(s) | 05:32.10 |
| R4v3n_ | thanx for your help :) at least i am a step further to the solution . much appreciate your help | 05:33.14 |
| ray_laptop | AFAIK, the .NET interface uses the API interface (gs/doc/API.htm) and sends special PostScript to provide the extra functionality | 05:34.04 |
| R4v3n_ | i never saw it like that ... may be i should look in the source for the script thats being sent out and walk through the code there so if i can find if something is off | 05:36.06 |
| ray_laptop | sounds like we need to make closer contact with the .NET developer as we did with the developer of GSView (Russell Lang) since we want to make sure things aren't being developed that will be difficult to maintain (as CUPS was until we took it over) | 05:36.11 |
| | since we maintain and develop Ghostscript future features, we know how to steer other developers that use the gs library/DLL | 05:37.25 |
| | R4v3n_: if you can insert a "-Z:" (case sensitive) argument to the gsapi_init_with_args argument list it might help (assuming you can capture the stdout and stderr logs) | 05:40.29 |
| R4v3n_ | thanx ray_laptop .. i'll try that and see if i can get the results i want | 05:42.10 |
| ray_laptop | R4v3n_: but if the .NET app sends PS sequences to use GS to perform specific functons (as I suspect) there really isn't a good debug facility that Ghostscript provides (at least yet) | 05:42.52 |
| | R4v3n_: good luck | 05:43.06 |
| R4v3n_ | thanx again, you've been a great help and pointed me into the right direction, i'll see if i can find where the bug is | 05:44.13 |
| ray_laptop | Robin_Watts: mvrhel_laptop: (others) please review my patches for the SeparationOrder issues with bug 694821 and other devices (tiffsep). Note that this does not address psdcmyk producing CMYK separations when they are not requested (29-07I.PS_ | 05:56.58 |
| | two commits at http://git.ghostscript.com/?p=user/ray/ghostpdl.git;a=summary | 05:57.50 |
| mvrhel_laptop | ray_laptop: looks reasonable to me | 06:00.00 |
| jhabjan | good morning | 06:40.50 |
| | I saw in the log that someone mentioned Ghostscript.NET | 06:41.36 |
| | there is no 'getimages' function in the Ghostscript.NET | 06:43.01 |
| | so, I believe 'R4v3n_' is using some other library or code sample... | 06:44.17 |
| | but.. I'm always willing to help with Ghostscript API usage from any .NET language | 06:45.17 |
| fasteque | hi all, any MuPDF developer online? | 09:30.52 |
| kens | fasteque : its a bit early at the moment, you can ask your question and they will see it in the channel logs | 09:31.20 |
| fasteque | yeah sure, timezone, my bad. | 09:31.56 |
| kens | Not a problem, they are European, but tend to start late and work late, in roder to communicate with our US colleagues | 09:32.25 |
| | Someone will be around in an hour or so, so please feel free to ask your question | 09:33.37 |
| | Its even possible (though unlikely) that someone else will be able to help | 09:33.52 |
| fasteque | ok thanks. I was just trying to understand when text annotation will be supported by the library. MuPDF is nearly perfect by it misses this important feature. I know there are two bugs opened for that, but I just wanted to know if they have any update. | 09:34.02 |
| kens | Annotations are a work in progress, I'm not sure of the exact state of play but I'm reasonably sure text annotations will be in at some point | 09:34.33 |
| fasteque | good, because the problem is also when you open a PDF with annonations and you can't get them. I just did a modification in the Android project to select them, but more then "delete" is not possible. Thanks for your info. Let's see if later they'll tell me something about that. | 09:36.54 |
| Robin_Watts | fasteque: Well, you've summed it up quite well. | 10:02.50 |
| | We have some support for annotations, but it's not as complete as we'd like. | 10:03.01 |
| | It's on our to-do list, but it's not at the top of the priority list. | 10:03.16 |
| fasteque | Hi Robin, thanks for your update | 10:10.01 |
| chrisl | kens: ping | 10:28.55 |
| kens | pong | 10:29.40 |
| chrisl | that customer with the barcode font problem - I have a thought (I know it's not really our problem....) | 10:30.09 |
| | I they add the barcode font | 10:30.19 |
| kens | OK what's the thought ? | 10:30.20 |
| chrisl | Does the PS drive have a "never embed" list? | 10:31.04 |
| kens | Not normally | 10:31.13 |
| chrisl | Oh well, that's that then..... | 10:31.36 |
| kens | But it shares hte pdfwrite code | 10:31.37 |
| | So it can do | 10:31.47 |
| chrisl | No meant the actual PS printer driver, not the backend | 10:32.05 |
| kens | Oh definitely not then | 10:32.16 |
| | The only way they are going to get reliable 'correct' PDF files is to embed the font. | 10:33.20 |
| | 2 seconds googling turned up half a dozen free replicas | 10:33.34 |
| | Given teh name, I suspect this is a freebie too | 10:33.52 |
| | So the forbid embedding is probably an oversight | 10:34.05 |
| chrisl | Yeh, I just figured if we could stop the PS driver embedding the font, then they'd get the PDF they are expecting. | 10:34.15 |
| | And I have seen a Windows PS driver that has equivalents of always embed and never embed lists | 10:34.58 |
| kens | Hmm, let me look at the WIndows driver again | 10:37.45 |
| | I mean it shouldn't be embedding the font anyway | 10:37.58 |
| chrisl | It look like it's not a part of the stock Windows PS driver, it must have been customised driver.... | 10:38.58 |
| kens | There are some potions, but not that one | 10:39.08 |
| | options* | 10:39.11 |
| | THey could send them as bitmaps -) | 10:39.20 |
| | Of course that's *all* TT fonts so not a good idea | 10:39.29 |
| | Fundamentally, they should go get a font that permits embedding | 10:39.44 |
| chrisl | True, but it's decidedly crap that the PS driver ignores the permissions flag like that :-( | 10:40.20 |
| kens | Given that its contra the spec, causes our XPS interpreter problems, and fails PDF validation, I'm not going back to the old method of 'write an Encoding' | 10:40.26 |
| | If we oculd detect that the font was embedded in the PostScript stream, I'd be inclined to allow it to be embedded in the PDF (just as we did before) | 10:40.54 |
| | But I couldn't see a way to tell easily | 10:41.02 |
| | Hmm the font nam eseems to belong to Elfring Fonts Inc, and they have a 'free trial version'. I wonder if someone just didn't bother to buy the retail one | 10:44.49 |
| chrisl | kens: I don't know if/how pdfwrite can see it, but fonts read from disk have a PATH entry in the font dict - that only works for PS/PDF of course | 10:45.18 |
| kens | Ah yes, from their web site: | 10:45.32 |
| | PDF Files: These fonts will not embed in a pdf file. Embeddable fonts for pdf files require a special license. | 10:45.32 |
| | chrisl well, I oculd look at that, but it looks to me like this is deliberate policy fomr the font foundry. I think I'll add that to the bug report. Especially since I commented on white listing the font. | 10:46.23 |
| chrisl | kens: yes, on that basis, I think you're right. | 10:47.03 |
| kens | oh and I should check the copyright notice in the font too, I did decode it | 10:47.15 |
| | Hmm maybe its a differne tfont, it doesn't mention that foundry | 10:47.47 |
| | Copyright 1997 :-) | 10:48.09 |
| | chrisl this is kind of bizarre | 11:05.48 |
| | The font embedded in teh PostScript file has a fsType of 2 | 11:06.08 |
| | THe font they sent as a TT font has an fsType of 0 | 11:06.20 |
| | I wonder if they even bothereed to send the same font | 11:06.32 |
| chrisl | That sounds dubious.... | 11:06.47 |
| kens | Yeah, I suspect they just grabbed a free bar code 39 font and sent it to us | 11:07.10 |
| | Nto that it really matters, but it now means I hjave to decode the actual font data in the PS file | 11:07.28 |
| chrisl | I wouldn't bother - I'd leave the ball in their court now | 11:07.47 |
| kens | Well I must admit I'm (lazily) inclined in that idrection | 11:08.04 |
| chrisl | You have answered based on the font they supplied - if that's not the font they used, that is not your problem | 11:09.13 |
| kens | Well the answer is correct no ,atter which font they used to make the PostScript file. After its been embedded and mangled by the PS driver there is no simple way to reconstruct the original and have it work | 11:10.05 |
| Robin_Watts | http://www.mbti123.com/iq/en/ | 11:17.56 |
| | Any answers (and reasons) for questions 15,18 and 19 gratefully received :) | 11:18.23 |
| kens | oops, I missed 15, what was htat question ? | 11:24.14 |
| | Oh that was another squares question | 11:26.08 |
| | COUldn't see that one | 11:26.14 |
| | DOn't see 18 either | 11:26.39 |
| | Well I got 160 | 11:29.44 |
| | COuldn't see the answers to those questions though | 11:30.03 |
| madmoose | Robin_Watts: on 15, reading vertically, I think the difference between the first two rows is equal to the sum of the digits in the last row | 12:43.07 |
| | Which would make -5 the answer | 12:43.26 |
| kens | Hmm I tried something like that, but couldn't see it, let me look again | 12:44.40 |
| Robin_Watts | madmoose: Cunning. | 12:45.42 |
| kens | Yeah that looks correct | 12:46.18 |
| Robin_Watts | And indeed, that boosts my iq :) | 12:46.19 |
| madmoose | I'll try the rest when I get home :) | 12:46.59 |
| tor8 | Robin_Watts: damn, I see no pattern at all in 18 and 19 :( | 13:17.56 |
| Robin_Watts | fears he has ruined productivity for today :( | 13:18.35 |
| kens | chrisl, mind if I assign #695031 to you ? | 14:14.09 |
| chrisl | kens: it should be me anyway | 14:14.32 |
| kens | (though I'm inclined to feel 'well don't do that' is the answer) | 14:14.47 |
| | CHanged ti to fonts so you get it | 14:15.18 |
| chrisl | I'll look at it at some point - I'm having to go into the bank this afternoon | 14:16.29 |
| kens | NP | 14:16.48 |
| henrys | meeting in 5 minutes | 15:24.15 |
| chrisl | Okay, so it turns out I am here for the meeting..... | 15:26.32 |
| Robin_Watts | chrisl: try harder next time. | 15:26.49 |
| chrisl | Robin_Watts: I actually prefer our IRC meetings to talking to bank people...... but not when they cancel when I'm halfway there :-( | 15:27.27 |
| henrys | next big event is mvrhel_laptop trip, do we have some demo files he can use yet? | 15:30.33 |
| | Robin_Watts: I added a GhostDocs section to the agenda, and font metrics is the first item. | 15:31.15 |
| | chrisl: so was your plan to do a release after the meeting? I hope so. | 15:32.27 |
| chrisl | henrys: yes, given that the meeting is early in the month. I may do an RC before the meeting, though | 15:33.01 |
| Robin_Watts | henrys: Nope, not yet. | 15:33.12 |
| | I'm trying to build a new ghostdocs demo now (with the new fonts and logo from Miles etc) and then once I have that, we can look for some good example docs. | 15:33.48 |
| mvrhel_laptop | good morning | 15:34.24 |
| | sounds great. Thanks Robin_Watts | 15:34.48 |
| Robin_Watts | no worries. | 15:35.00 |
| henrys | Robin_Watts: do you need help or do you want to do it alone? | 15:35.37 |
| Robin_Watts | building the new demo? I don't think there is much scope for parellisation. | 15:36.11 |
| | or parallelisation, even. | 15:36.20 |
| henrys | Robin_Watts: I was thinking about the font metrics | 15:36.32 |
| Robin_Watts | henrys: Oh, that's done. | 15:36.41 |
| | kerning isn't right, but the basic metrics are. | 15:36.53 |
| | Getting kerning to be right will be... interesting. | 15:37.06 |
| | because the MS fonts have both 'GPOS' and 'kern' tables (and the GPOS takes priority). | 15:37.44 |
| | Ours only have kern tables. | 15:37.52 |
| henrys | Robin_Watts: so once you have something marcosw should start testing again? | 15:38.41 |
| | or do you think we need to wait for a kerning solution | 15:38.53 |
| | ? | 15:38.54 |
| Robin_Watts | which leads into all sorts of questions, like do things like libreoffice manage to match layout when they use freetype, that ignores lots of the tables in the fonts that make a difference? | 15:39.16 |
| | henrys: No, I think marcosw should test again. | 15:39.24 |
| chrisl | Robin_Watts: they probably use a layout engine, which does use those tables | 15:39.48 |
| Robin_Watts | After the meeting, if he's about, I can talk to him about that. | 15:39.51 |
| tor8 | Robin_Watts: libreoffice might use ICU for opentype support (and get GPOS kerning from that) | 15:40.01 |
| marcosw | morning | 15:40.15 |
| Robin_Watts | ok. | 15:40.21 |
| marcosw | sorry I'm late | 15:40.23 |
| Robin_Watts | marcosw: Are you going to be about after the meeting ? | 15:40.28 |
| tor8 | Robin_Watts: or it might use pango (the gtk+ equivalent) which uses harfbuzz under the hood for opentype layout | 15:40.31 |
| marcosw | Robin_Watts: yes | 15:40.35 |
| Robin_Watts | tor8: fair enough. | 15:40.54 |
| paulgardiner | marcosw: Did the dll I sent fix the problem with the GhostDocs viewer? | 15:41.25 |
| chrisl | Robin_Watts: although Freetyepe doesn't use those tables, it *should* give access to the data in them, IIRC.... | 15:41.30 |
| tor8 | chrisl: yeah. they started using the tables, then ripped that code out and made it into another project, which eventually turned into harfbuzz | 15:41.59 |
| | IIRC | 15:42.01 |
| Robin_Watts | paulgardiner: If you run MuPDF on an android device, and try to open a password protected file, and then cancel out, does it crash for you? | 15:42.16 |
| henrys | ray_laptop: did you see the discussion about the chunk manager and memento improvements that could be made in non-memento mode? | 15:42.18 |
| chrisl | tor8: yes, you can still get the (dead) freetype 1 code that did all that | 15:42.28 |
| marcosw | paulgardiner: to be honest I haven't tried it yet. | 15:42.53 |
| paulgardiner | Robin_Watts: Don't know. I haven't tried passworded files for a while. I don;t think I have one to hand either | 15:42.57 |
| | marcosw: No worries. Just so long as your not waiting on a solution. | 15:43.21 |
| henrys | paulgardiner: how's ghostdocs + mupdf coming along out there on the bleeding edge of artifex technology ;-) | 15:43.39 |
| ray_laptop | henrys: no, when was that discussion ? | 15:43.51 |
| | henrys: I thought i've read the IRC logs | 15:44.03 |
| henrys | ray_laptop: couple days back. I'll look it up for you later and tell you the date. | 15:44.39 |
| paulgardiner | henrys: it's moving along, but I'm yet to get to a point where I can really evaluate all that's needed. I'm still working on building without the FBS. | 15:45.02 |
| ray_laptop | henrys: that's OK. I'll just search for 'memento' It isn't mentioned all that much | 15:45.10 |
| kens | There's a bug report isn't there ? I think its all confused with PCL fonts as well | 15:45.54 |
| paulgardiner | henrys: just today I swapped to looking at the signature creation for Raed. Shouldn't be a long job, and I thought it would be good to get out of the way. | 15:46.17 |
| henrys | paulgardiner: okay good. | 15:46.40 |
| | kens: anything on the pdfwrite or interpreter front? | 15:47.05 |
| kens | Not at present, except that I need to talk with Michael after the meeting | 15:47.21 |
| chrisl | Does anyone have anything in the works that might influence the next release? | 15:48.00 |
| kens | Still working on tje colour management, that's why I need to talk to Michael | 15:48.16 |
| henrys | I didn't have much else for this week. Next week I wanted to get the agenda ready for the meeting so if everyone could look at that in advance and weigh in next week, that would be good. We don't want to sit around in Texas staring at each other with nothing to say. | 15:48.25 |
| kens | And I'd like to get the PostScript configuration enhancement done (chrisl) | 15:48.40 |
| chrisl | kens: I don't remember that.... | 15:49.12 |
| kens | TO allow thinkgs like MediaPosition to be put in the ps2write output | 15:49.26 |
| henrys | anybody else have meeting stuff? how's torscript tor8? | 15:49.27 |
| chrisl | kens: oh, right, of course | 15:49.44 |
| kens | I'd like to wedge that into the next release | 15:49.56 |
| ray_laptop | henrys: I found the discussion back on 2/6 -- what you are looking for is a better cname ? | 15:50.15 |
| tor8 | henrys: going to start trying to integrate it into mupdf now | 15:50.23 |
| Robin_Watts | ray_laptop: I made a memento change that passed a better cname in. | 15:50.42 |
| | In retrospect I should have done that for normal builds too. | 15:50.55 |
| ray_laptop | Robin_Watts: right -- I saw that in the log | 15:51.00 |
| Robin_Watts | Either you can do that, or I can. | 15:51.09 |
| ray_laptop | Robin_Watts: I've just wrapped up my last customer bug, so I'll go ahead and do it. Your busy having fun with fonts | 15:51.51 |
| henrys | marcosw: we do have a release coming up if you see something that should be fixed make it a blocker. You look at the bugs most often | 15:53.09 |
| ray_laptop | Robin_Watts: BTW, thanks for the once over on the patches. After you saw it, I disabled the "example" code to get rid of a couple of compiler warnings in gdevcmykog.c | 15:53.16 |
| marcosw | henrys: will do. I don't think there are any open bugs that are blockers, but I'll double check. | 15:53.47 |
| Robin_Watts | disabled how ? | 15:53.51 |
| henrys | of course we all should look at the bugs for blockers ... | 15:53.54 |
| marcosw | I've try to mark regressions as blockers as I enter them, figuring there isn't any reason to wait until a month before the release. | 15:54.39 |
| Robin_Watts | ray_laptop: I see. That's fine. | 15:55.14 |
| henrys | let's adjourn 5 minutes early. | 15:55.25 |
| kens | :-) | 15:55.50 |
| mvrhel_laptop | sounds good. kens. I have to help get kids out the door for about 5 minutes then we can chat about icc profiles | 15:55.56 |
| kens | No problem, I'll go get a coffee, ping me when ready | 15:56.08 |
| Robin_Watts | OK, marcosw... so could you rerun the layout tests please? | 15:56.38 |
| marcosw | Robin_Watts: sure. should have results later today. | 15:57.27 |
| Robin_Watts | marcosw: You're pulling from which repo? | 16:01.08 |
| marcosw | /home/robin/sauce/epage.git | 16:01.26 |
| Robin_Watts | perfect. thanks. | 16:01.34 |
| | Oh, wait... | 16:01.42 |
| marcosw | git sure is slow... | 16:01.45 |
| Robin_Watts | You'll need to change the build and/or command you're using slightly. | 16:02.01 |
| | What build command are you using now ? | 16:02.16 |
| ray_laptop | Robin_Watts: so, the reason you didn't always pass the cname into chunk_mem_node_add was because you thought that the extra parameter would impact performance ?? Seriously ?? Or was there something else ? | 16:03.00 |
| marcosw | scripts/linuxbuild.py da/test-shell -vs2005 -nonet -romfss2 -define=ALL_DAS -define=PDF_EXPORT -debug | 16:03.02 |
| Robin_Watts | ray_laptop: I didn't want to change the performance of normal builds, yes. | 16:03.52 |
| | Silly in retrospect. | 16:04.01 |
| | marcosw: OK. You should be able to drop the -vs2005 :) | 16:04.21 |
| | And if you use -romfss2 the fonts will be pickled into the exe. | 16:04.39 |
| marcosw | the run command looks like this: | 16:04.43 |
| | da-test -save /home/marcos/artifex/epage/test -hq home/marcos/artifex/epage/test.docx out.txt ooxml rom:libraries/image/resource rom:libraries/font/library | 16:04.47 |
| Robin_Watts | Oh, you had -romfss2 already. | 16:05.17 |
| | OK. So you should omit the two rom:/ things now. | 16:05.34 |
| marcosw | sounds easy enough | 16:06.19 |
| Robin_Watts | assuming I've not screwed up, yeah :) | 16:09.11 |
| marcosw | chrisl: when I updated from ufst to ufst-6.x for the weekly 'make uproduct' regression test I found a bunch of differences that didn't appear with the normal build. The differences included missing text and incorrectly bolded text (the latter in lots of files). Did I do something wrong or is that possible? | 16:09.12 |
| mvrhel_laptop | kens: I am back | 16:11.03 |
| kens | OK | 16:11.10 |
| mvrhel_laptop | so I am a little confused | 16:11.24 |
| kens | Hmm, by what ? | 16:11.33 |
| mvrhel_laptop | I ran the original file through pdfwrite | 16:11.39 |
| kens | With -dUseCIEColor ? | 16:11.48 |
| mvrhel_laptop | yes | 16:11.57 |
| ray_laptop | henrys: do you want the cname for adding a chunk to be the one that causes the chunk to be added even if that chunk holds multiple objects ? It's easy enough to have multiple object chunks use the "chunk_mem_node_add" string, and single object chunks to use the supplied cname. | 16:12.00 |
| kens | OK | 16:12.02 |
| mvrhel_laptop | let me do it again real quick here | 16:12.02 |
| ray_laptop | henrys: Robin_Watts: I pushed the patch to my repo that always uses the cname. I'll change it ro use chunk_mem_node_add for multiple object allocations if you want. One of you please check the change. | 16:14.38 |
| marcosw | bbiab, getting breakfast while epage compiles | 16:15.03 |
| mvrhel_laptop | kens: hmmm ok now I am seeing what I would have expected. I must have been doing something wrong before | 16:15.08 |
| | sorry! let me go through and look at these profiles now | 16:15.18 |
| kens | Well its non-tricial I guess | 16:15.20 |
| Robin_Watts | ray_laptop: Looks good to me. | 16:15.41 |
| kens | THe profiles seem 'OK', but they don't have a rXYZ, bXYZ or gXYZ tag | 16:15.41 |
| mvrhel_laptop | let me look at them with the inspector | 16:15.59 |
| chrisl | marcosw: Well, although it wouldn't have the effects you're seeing, but you shouldn't be doing "make uproduct" because that uses the old code | 16:16.05 |
| kens | WHich causes them ot be unusable in a link, At least that's what seems to come out of LCMS | 16:16.10 |
| henrys | ray_laptop: ah yes that does complicate things, perhaps cname should only use the client name for objects allocated in their own chunk, I don't know. | 16:16.30 |
| Robin_Watts | ray_laptop: I think you're right. It would be nicer if chunks only got the "informative" name if they were single object chunks. | 16:16.53 |
| henrys | ray_laptop: if we are allocating a new chunk because we are out of memory the "client" is the memory manager. | 16:17.24 |
| Robin_Watts | The cname is still associated with the individual blocks within chunks, right ? | 16:17.25 |
| chrisl | marcosw: the missing text is a typo, trivial to fix. The seg fault is a little more complex - I have a fix, but I'm trying to find tests for a (it seems!) rarely used path through the code. | 16:17.45 |
| Robin_Watts | Interesting: http://www.indiegogo.com/projects/fin-wearable-ring-make-your-palm-as-numeric-keypad-and-gesture-interface | 16:17.52 |
| mvrhel_laptop | I don't like that it created 425 versions of a profile fromabc..... I hope it did not create the same profile that many times. | 16:17.53 |
| kens | mvrhel_laptop : What created the profile ? | 16:18.09 |
| mvrhel_laptop | ghostscript | 16:18.14 |
| kens | pdfwrite may use it a lot of tiems, but it should spot reuse | 16:18.22 |
| | When I was testing it it did spot that it was the same space | 16:18.46 |
| mvrhel_laptop | kens: it is likely not you (pdfwrite) that is the source of that issue | 16:18.46 |
| ray_laptop | henrys: Robin_Watts: OK. I'll change that. | 16:18.49 |
| mvrhel_laptop | a separate issue that I will need to look at | 16:19.04 |
| kens | OK | 16:19.17 |
| mvrhel_laptop | kens | 16:19.21 |
| | ok these have a table | 16:19.25 |
| marcosw | chrisl: these changes aren't ones I've entered bugs for, those are for the standard ufst build. the ones asked about are the changes that occur with the 'make uproduct' build. I guess I'm asking if the 'make uproduct' build should have differences that the standard ufst build or if I'm doing something wrong. | 16:19.35 |
| mvrhel_laptop | as opposed to the rXYZ etc type | 16:19.36 |
| ray_laptop | Robin_Watts: yes, the cname in the -ZA output is still that of the caller, whether or not the object was in it's own chunk | 16:19.41 |
| kens | mvrhel_laptop : I guess that makes some sense. | 16:19.49 |
| Robin_Watts | ray_laptop: At some point I should sprinkle some Memento_label's in this code. | 16:20.08 |
| mvrhel_laptop | likely due to the fact that the ICC format can't directly match up to the CIE based transform with its single matrix and curves | 16:20.14 |
| chrisl | marcosw: I've no idea - if you want to mail me a few examples, I'll try them, and let you know what I think | 16:20.25 |
| kens | These are created by gs_colorspace_set_icc_equivalent when I encouter a CIEBased space and there is no icc_equivalent | 16:20.27 |
| mvrhel_laptop | kens | 16:20.29 |
| | oh | 16:20.32 |
| | intersesting | 16:20.38 |
| | there may be an issue with one of the structures | 16:20.46 |
| kens | Its calling gx_cieabc_to_icc() | 16:20.56 |
| marcosw | Robin_Watts: I needed to add a 'static' to get the current epage code to compile, should I push that change to your repository (the previous source needed two changes, so things are improving :-) ). | 16:20.56 |
| | chrisl: will do. | 16:21.00 |
| mvrhel_laptop | the profile inspector crashed when I tried to view it | 16:21.00 |
| Robin_Watts | marcosw: If you can push, go for it :) | 16:21.14 |
| mvrhel_laptop | kens: that is likely our profile | 16:21.15 |
| kens | mvrhel_laptop : in gsciemap.c around line 848. Oh, that doens't sound good | 16:21.15 |
| mvrhel_laptop | problem | 16:21.18 |
| | kens: I will look into this | 16:21.37 |
| marcosw | chrisl: actually I'll just enter a bug, it's as easy and if you decide it's a problem it will save a step (if not you can close the bug and improve your weekly fixed bug count). | 16:21.48 |
| mvrhel_laptop | sorry that I did not find this earlier kens | 16:21.51 |
| marcosw | ^easy^issue | 16:21.59 |
| kens | OK I'll be here for an indeteminate but reasonable length of time waiting for Stella and Melanie, let me know if I can help | 16:22.02 |
| kens | dounts that I cna but still.... | 16:22.14 |
| mvrhel_laptop | kens: ok likely not much for you to do | 16:22.16 |
| chrisl | marcosw: okay - my answer will probably be "use the new code"! | 16:22.35 |
| marcosw | Robin_Watts: you're right, I don't have permissions to push to the repository. Here's the diff: --- a/foundation/libraries/io/fs/romfss2.c | 16:25.14 |
| | +++ b/foundation/libraries/io/fs/romfss2.c | 16:25.14 |
| | @@ -310,7 +310,7 @@ Error RomFss2_read(Fcb *cb, size_t required, const uint8_t **block, | 16:25.14 |
| | return EPAGE_NO_ERROR; | 16:25.14 |
| | } | 16:25.14 |
| | | 16:25.14 |
| | - *block = (uint8_t *)mine->file->data + mine->pos; | 16:25.15 |
| | + *block = (const uint8_t *)mine->file->data + mine->pos; | 16:25.15 |
| | *delivered = mine->file->size - mine->pos; | 16:25.16 |
| | | 16:25.16 |
| | return EPAGE_NO_ERROR; | 16:25.17 |
| Robin_Watts | Thanks. | 16:25.32 |
| marcosw | There is another if you building the pdf export code: | 16:25.41 |
| | --- a/agents/pdf-exporter/pdf-exporter.c | 16:25.47 |
| | +++ b/agents/pdf-exporter/pdf-exporter.c | 16:25.48 |
| | @@ -489,7 +489,7 @@ static Error urlFromFilename(const char *fname, | 16:25.48 |
| | ((fnameLen>=Pal_strlen("file")) && (Pal_strncmp(fname,"file", Pal_strlen("file"))==0)) ) | 16:25.48 |
| | { | 16:25.48 |
| | Unicode *uniUrl; | 16:25.48 |
| | - error = Uconv_toUnicode((uint8_t*)fname, &uniUrl, | 16:25.48 |
| | + error = Uconv_toUnicode((const uint8_t*)fname, &uniUrl, | 16:25.49 |
| | Uconv_Encoding_Latin1, context); | 16:25.49 |
| | if (error) | 16:25.50 |
| | return error; | 16:25.50 |
| | @@ -1029,7 +1029,7 @@ static Error loadDocument(Epage_Context *epage, | 16:25.51 |
| | /* get file mimetype */ | 16:25.51 |
| | if (mimeTypeChar) | 16:25.52 |
| ray_laptop | marcosw: can't be bothered with pastebin ;-) | 16:27.09 |
| marcosw | pastebin? we don't need no stinkin' pastbin! | 16:27.52 |
| Robin_Watts | marcosw: Well, pastebin might mean I got to see the whole diff :) | 16:28.53 |
| marcosw | oops, sorry. I'll try that. | 16:29.05 |
| henrys | we should fix the permissions, yes? | 16:29.48 |
| marcosw | http://pastebin.com/EC86wzMV | 16:29.50 |
| | and the first one again: http://pastebin.com/sMvWp9xk | 16:30.46 |
| Robin_Watts | henrys: Well, we should have a central repo somewhere really. | 16:33.12 |
| | We don't currently let people push to other peoples personal repos. | 16:33.34 |
| henrys | Robin_Watts: oh right of course | 16:34.11 |
| Robin_Watts | But we'd need a new sort of central repo that wasn't world visible. | 16:34.44 |
| marcosw | Robin_Watts: I'm getting another error, but I don't see what's wrong with the code: | 16:35.04 |
| | cc1: warnings being treated as errors | 16:35.07 |
| | /home/marcos/artifex/epage/agents/test-shell/test-shell.c: In function 'main': | 16:35.07 |
| | /home/marcos/artifex/epage/agents/test-shell/test-shell.c:2962: error: cast discards qualifiers from pointer target type | 16:35.07 |
| Robin_Watts | resource_url probably needs to be a const char * | 16:35.47 |
| marcosw | It's declared as a "char *" (also I tried the const char * cast) | 16:36.12 |
| Robin_Watts | Indeed, but a string literal such as "rom:/...." is probably a const char *. | 16:36.57 |
| | and hence assigning that into a char * is discarding the const qualifier | 16:37.17 |
| marcosw | true, but I tried casting the string to a char * and that didn't work and volatile char * didn't work either (is volatile the opposite of const? I haven't programmed in C much lately). | 16:39.07 |
| Robin_Watts | marcosw: casting the string won't work. | 16:39.29 |
| | because that's implicitly what we're doing now. | 16:39.42 |
| | and we'd get the same error. | 16:39.50 |
| | volatile is not the opposite of const. | 16:40.00 |
| | Can you not just change "char *resource_url;" to be "const char *resource_url;" ? | 16:40.28 |
| marcosw | I was presuming that it was not const for a reason, but apparently not, since that worked. | 16:41.42 |
| ray_laptop | AIUI, volatile tells the compiler to assume that a variable may change due to another thread touching it | 16:42.14 |
| marcosw | accoring to the internets: | 16:42.32 |
| | A common misconception is to imagine that somehow const is the opposite of volatile and vice versa. | 16:42.35 |
| Robin_Watts | specifically, volatile says that the compiler can't assume it can optimise anything about accesses through a volatile pointer. | 16:43.35 |
| | For instance if I have: volatile int *fred; *fred = 1; *fred = 2; *fred = 3; the compiler can't remove any of them. | 16:44.20 |
| | It also can't reorder them w.r.t. other things. The classic example is where fred points to a hardware register to do polled io. | 16:45.02 |
| ray_laptop | HURRAY! All my customer bugs are closed. FINALLY Still have some customer enhancements... | 16:47.30 |
| kens | Congrats ray_laptop | 16:48.32 |
| mvrhel_laptop | kens: blah. it looks to me like no one is following the spec. so this profile has a special structure. the structure can have a curve, followed by a matrix, followed by a curve, then an MLUT and then a curve. Any of the items can be not present to make various combinations. This is indicated by setting the offset address for that item to zero. Which I clearly do as this particular... | 16:49.34 |
| | ...profile only has a curve, matrix and then a curve. Unfortunately, it appears that this is a problem for the profile inspector and for lcms | 16:49.36 |
| kens | OK mvrhel_laptop | 16:49.50 |
| Robin_Watts | ray_laptop: How's the eye btw? | 16:49.53 |
| kens | mvrhel_laptop : if this is a problem I can punt it through the odl code for now | 16:50.23 |
| mvrhel_laptop | kens: so, what I need to do is either put in identity items for those parts, or use a different structure. I may go ahead and use a different structure | 16:50.34 |
| | kens: no let me fix it on this end | 16:50.40 |
| kens | OK no problem | 16:50.52 |
| mvrhel_laptop | I may send this profile off to someone to look at too | 16:51.01 |
| | to me it is valid | 16:51.07 |
| kens | :) | 16:51.13 |
| mvrhel_laptop | from my reading of the spec | 16:51.13 |
| | I am going to see what lcms is doing with it too | 16:51.35 |
| | this is straight out of the sepc | 16:52.16 |
| | spec | 16:52.18 |
| | "The offsets indicatethe number of bytes from the beginning of the tag to the desired data. If any of the offsets are zero, that is anindication that processing element is not present and the operation is not performed." | 16:52.24 |
| kens | well you're way p[ast my competence level | 16:52.31 |
| | But if I can help, do let me know | 16:52.48 |
| mvrhel_laptop | kens: no you can move on to bigger and better things. I should have this resolved this week | 16:53.07 |
| kens | Thanks mvrhel_laptop | 16:53.15 |
| ray_laptop | Robin_Watts: I still have the Bell's. Some people tell me they think it is improving, but I can't tell since I see myself every day. I think my eye is getting easier to close. That would be great | 16:59.54 |
| | Robin_Watts: but the eye exam yesterday went well. My cornea has a little bit of drying on the lower 20% but he said it's looking good and what I am doing is working | 17:01.03 |
| Robin_Watts | Yeah, I think you probably have to trust to external viewers to tell you how you look :) | 17:01.35 |
| | Good to hear it's getting better. | 17:01.48 |
| ray_laptop | it'll be nice to not have to tell people "no, I didn't have a stroke" | 17:02.51 |
| madmoose | Yay, I'm a contributor | 17:23.11 |
| Robin_Watts | madmoose: Indeed. Thanks again! | 17:24.04 |
| henrys | Robin_Watts: that scan build warning looks scary, didn't look at his code though | 17:26.16 |
| Robin_Watts | henrys: ? | 17:26.46 |
| henrys | Robin_Watts: there were compiler warning with your commit from Thomas | 17:27.21 |
| Robin_Watts | which one? | 17:27.49 |
| | There were 3 Thomas commits and I can't see any warnings... | 17:28.05 |
| | "Undefined or garbage value returned to caller" in mutiff.c? That's scanbuild being crap. | 17:29.52 |
| | henrys: Where are you seeing these compiler warnings reported? | 17:39.37 |
| henrys | in the email sent to you and tech with the subject New muPDF compiler warnings | 17:40.17 |
| Robin_Watts | Gah. Can't see for looking. | 17:41.23 |
| | sorry. | 17:41.26 |
| | Right, so just the one new warning, and it's scanbuild being crap. | 17:41.39 |
| | scanbuild gets confused by the try/catchery. | 17:42.00 |
| henrys | Robin_Watts: yes, I see | 17:43.00 |
| | I wonder if it could handle it if the jmp stuff was right in there or if it doesn't handle jumps at all | 17:44.55 |
| | not important | 17:45.05 |
| Robin_Watts | henrys: The setjmp is in the fz_try macro | 17:46.46 |
| | (i.e. not buried in a function) | 17:47.02 |
| henrys | Robin_Watts: I am convinced it is exceedingly difficult/impossible to have clean code without exception machinery | 17:47.03 |
| Robin_Watts | henrys: Yes. IMAO mupdf is much nicer than gs in this regard. | 17:47.29 |
| | but it's not a magic bullet. | 17:47.35 |
| | There are costs to working with exceptions, but overall it's a win, I feel. | 17:48.08 |
| henrys | no but having functions return values and not error codes is a huge win IMHO | 17:48.17 |
| Robin_Watts | yeah. | 17:49.48 |
| marcosw | Robin_Watts: the new epage/word comparison is done and on the web at the same location as before. Many documents showed great improvement, but tables are still a problem as are footnotes (epage seems not render footnotes at all). | 17:52.35 |
| Robin_Watts | marcosw: Thanks. | 17:52.54 |
| marcosw | some of the documents show no results for epage, I'm checking to see if there are error messages from epage or just a problem with my script. | 17:53.31 |
| Robin_Watts | Did they show results before? | 17:56.58 |
| jhabjan | in ghostscript, when using display device, it seems GraphicsAlphaBits and TextAlphaBits are ignored when gsapi_init_with_args is called | 18:00.47 |
| kens | And on the command line ? | 18:01.47 |
| jhabjan | on command line when using other devices it works | 18:02.21 |
| kens | Its entirely possible the device doesn't implement it | 18:02.21 |
| | using other devices isn't relevant, has ot be the sme device | 18:02.36 |
| jhabjan | hm.. how do I test 'display' device via command line? | 18:03.03 |
| kens | Its hte default device in WINdows | 18:03.21 |
| ray_laptop | jhabjan: you can either omit the -sDEVICE=____ or use -sDEVICE=display | 18:04.38 |
| | jhabjan: these parameters work with gsapi_init_with_args because gswin32.exe and gswin32c.exe use that and they work for me. | 18:07.44 |
| jhabjan | it works via command line | 18:08.15 |
| ray_laptop | jhabjan: the command line sends the args using gsapi_init_with_args | 18:09.08 |
| | jhabjan: you may have the argc wrong in your call ?? | 18:09.40 |
| jhabjan | args.Add("-sDEVICE=display"); | 18:10.09 |
| | if (Environment.Is64BitProcess) | 18:10.09 |
| | { | 18:10.09 |
| | args.Add("-sDisplayHandle=0"); | 18:10.09 |
| | } | 18:10.09 |
| | else | 18:10.10 |
| | { | 18:10.10 |
| | args.Add("-dDisplayHandle=0"); | 18:10.11 |
| | } | 18:10.11 |
| | args.Add("-dDisplayFormat=" + | 18:10.12 |
| | ((int)DISPLAY_FORMAT_COLOR.DISPLAY_COLORS_RGB | | 18:10.12 |
| | (int)DISPLAY_FORMAT_ALPHA.DISPLAY_ALPHA_NONE | | 18:10.13 |
| | (int)DISPLAY_FORMAT_DEPTH.DISPLAY_DEPTH_8 | | 18:10.13 |
| | (int)DISPLAY_FORMAT_ENDIAN.DISPLAY_LITTLEENDIAN | | 18:10.14 |
| ray_laptop | jhabjan: please use pastebin or the like rather than pasting large floods here | 18:12.55 |
| jhabjan | I will, sorry once again | 18:13.11 |
| madmoose | I've seen nice code with error return codes, and I've seen nice code with exceptions. And I've seen crap code with either too. mupdf is very nice and readable. | 18:13.49 |
| ray_laptop | jhabjan: you can try changing the -dTextAlphaBits=4 to -sTextAlphaBits=4 -- that should give you: Unrecoverable error: typecheck in .putdeviceprops | 18:14.44 |
| | (I make that mistake sometimes when typing a command line, that's how I know) | 18:15.24 |
| jhabjan | yep, I got typecheck (-20) error | 18:17.35 |
| | -sTextAlphaBits=4 throws error, -dTextAlphaBits does nothing, -sGraphicsAlphaBits=4 and -dGraphicsAlphaBits does nothing | 18:23.17 |
| mvrhel_laptop | kens: you still here? | 18:23.29 |
| kens | for a while yes | 18:23.35 |
| mvrhel_laptop | so, I now I don't think the problem is with the profiles that I am creating. | 18:23.54 |
| | there is something else going on | 18:23.57 |
| kens | I can believe that | 18:24.12 |
| mvrhel_laptop | when I run the original file with -dUseCIEColor | 18:24.12 |
| | to tiff32nc for example. it uses the profiles and all works fine | 18:24.28 |
| kens | Yes | 18:24.36 |
| | THe problem is I don;t use hte profile I embed it | 18:24.44 |
| mvrhel_laptop | when I run out to the pdfwrite device I only get a partial file | 18:24.45 |
| kens | I get a full file | 18:25.04 |
| mvrhel_laptop | I dont' let me rurun again | 18:25.11 |
| kens | LIke hte one I sent | 18:25.13 |
| | mvrhel_laptop : Are yo uusing current code ? | 18:25.34 |
| | I fixed a bug that caused truncated files a short while back | 18:25.47 |
| mvrhel_laptop | ok | 18:25.50 |
| | let me update | 18:25.52 |
| jhabjan | ray_laptop: never mind, I found a workaround, I'm setting '/GraphicsAlphaBits 4' after setting /HWResolution right before executing >>setpagedevice | 18:27.35 |
| ray_laptop | jhabjan: sorry -- I was away. Yes, you can set device parameters via setpagedevice or putdeviceprops | 18:29.43 |
| jhabjan | ray_laptop: no worries, I was just expecting this two parameters will be applied by default if they are used with gsapi_init_with_args | 18:31.14 |
| mvrhel_laptop | kens: so the file I get is 156K and when I try to open it with acrobat I get invalid transparency group | 18:32.15 |
| kens | yes, its the progfile | 18:32.29 |
| mvrhel_laptop | the file you sent me is larger | 18:32.44 |
| | 370K | 18:32.54 |
| | let me step through the file that pdfwrite is making and I will see what is going on with it | 18:33.22 |
| kens | compressed file michael | 18:34.02 |
| mvrhel_laptop | ah ok | 18:34.40 |
| ray_laptop | jhabjan: they _should_ be. Have you tried -Z: -- it echoes the arguments to stderr | 18:36.21 |
| | henrys: for the patch to the chunk allocator from norbetj I've put "Thanks to Norbert Janssen" in the log message. Should I mention the company as well ? | 18:37.40 |
| henrys | ray_laptop: I usually just mention Norbert | 18:38.36 |
| mvrhel_laptop | kens: oh I know what is going on now | 18:39.42 |
| | crap | 18:39.48 |
| kens | Hmm that doesn't sound good | 18:39.56 |
| mvrhel_laptop | so these profiles that I generate for replacing CIE color spaces are one way | 18:40.15 |
| | in that they are used for mapping from source to CIE. | 18:40.41 |
| | not from CIE to source | 18:40.44 |
| kens | I wondered if it might be something to do with that | 18:40.46 |
| ray_laptop | henrys: thanks | 18:40.58 |
| mvrhel_laptop | in this particular case, this color space is getting used for a transparency group | 18:41.04 |
| | which needs to be two way | 18:41.08 |
| kens | ah..... | 18:41.13 |
| | THat meybe explains why its only on *some* files | 18:41.21 |
| mvrhel_laptop | kens: let me do a little thinking about this | 18:42.21 |
| | today | 18:42.22 |
| kens | OK thanks mvrhel_laptop | 18:42.34 |
| kens | is still waiting for the family to come home | 18:46.47 |
| jhabjan | ray_laptop: i just tried (-Z), both parameters are listed via stderr | 18:47.56 |
| | ray_laptop: http://pastebin.com/5DXyQ6zH | 18:49.41 |
| ray_laptop | jhabjan: I can't imagine what's going on. | 18:52.59 |
| jhabjan | ray_laptop: no worries, I will use a workaround with setpagedevice, thanks | 18:54.50 |
| mvrhel_laptop | kens: part of me is thinking that in this odd case, that we simply should simply use the appropriate DeviceGray, DeviceRGB, or DeviceCMYK for the group color space when writing the pdf color space when we have a color space that was from a PS CIE color space | 18:56.39 |
| | there are plenty of warnings not do use -dUseCIEColor | 18:57.06 |
| | s/do/to/ | 18:57.11 |
| kens | :-) | 18:57.14 |
| | I odn't think I cna tell its for transparency | 18:57.33 |
| | Its just a colour spce when I see it | 18:57.54 |
| mvrhel_laptop | you don't know when you are pushing a group? | 18:58.00 |
| kens | Not at the level of the colour handling | 18:58.10 |
| | I could obviously carry some state around but that owuld be ugly | 18:58.38 |
| mvrhel_laptop | hmmm. the transparency code is handing you a group where one of the parameters is an ICC color space which was created from a PS color space | 18:58.52 |
| | I don't think you need a state | 18:59.00 |
| | kens: I will poke at it and see if I can make something and let you decide if you like it or not | 18:59.46 |
| kens | Well I know its a CIEBased space, and there is no ICC equivalent | 18:59.49 |
| | But the code that writes the colour space doesn't know that its being called form the transparency handling | 19:00.10 |
| mvrhel_laptop | kens: I am suggesting that you fix it up when you get the call of a group push | 19:00.29 |
| kens | Hmm you mean change teh colour space from CIEBased to 'soemthing else' ? | 19:00.50 |
| mvrhel_laptop | hmm. actually, let me make the fix in ztrans.c | 19:03.11 |
| | then you will never know | 19:03.14 |
| kens | :-) Well that's easier for me certainly.... | 19:03.30 |
| mvrhel_laptop | we really should not be using these spaces for group color spaces | 19:03.39 |
| kens | But I am not sure that will work | 19:03.41 |
| mvrhel_laptop | trust me... :) | 19:03.50 |
| kens | AH wait, it will be OK for CIEColr | 19:03.56 |
| | And I bet the cololur space push is fine with the UseDeviceIndependentColor flag too | 19:04.24 |
| | I shoudl check that though | 19:04.35 |
| mvrhel_laptop | if the group is a real (two way) ICC color space all will be fine. Its just this -dUseCIEColor case that will be different | 19:05.52 |
| kens | Yes, I was just thinking of the case where I convert the space in the pdfwrite code. There is a flag to have it emit ICCbased colour there which involves (obviously) changing the colour space. I just need to check that it works. I think it will but this has worried me slightly | 19:07.05 |
| mvrhel_laptop | hmm kens, let me look at this more. I actually can't tell in ztrans.c that the profile was from a PS CIE color space | 19:08.53 |
| kens | OK I may be called away at any moment now | 19:09.13 |
| | OK hte device independent colour flag is fine, it leaves the group colour spaces alone | 19:10.32 |
| | Well its been a long day, I think I'll go find out where the family is, goodnight all | 19:12.10 |
| mvrhel_laptop | kens: good night | 19:14.45 |
| | I think I see how to do this | 19:14.50 |
| | will have something for you to review tomorrow | 19:14.58 |
| Robin_Watts | mvrhel_laptop: Did you have stuff you needed me to look at? | 19:22.34 |
| mvrhel_laptop | Robin_Watts: there is a bunch of stuff in my repos. It is far from complete but if you want to start looking at it that would be great | 19:23.07 |
| Robin_Watts | so each commit is in a reasonable state, just the whole functionality isn't there? | 19:23.47 |
| mvrhel_laptop | it all is self contained in the winrt platform folder (even though this particular project is not winrt) | 19:23.51 |
| | Robin_Watts: I would say the last two commits are in a reasonable state, with the one before the most recent one being a large initial commit for the "real" thing | 19:24.50 |
| | Robin_Watts: from here on out, the commits will be smaller and more contained | 19:25.04 |
| | each hopefully adding some new functionality | 19:25.29 |
| Robin_Watts | So, will some of this stuff be refactored out of the winrt folder in the end? | 19:25.30 |
| mvrhel_laptop | Robin_Watts: yes. just for now, it is easier for me to do it in there | 19:25.46 |
| Robin_Watts | ok. I will have a look. | 19:26.00 |
| mvrhel_laptop | Robin_Watts: thank you | 19:26.05 |
| Robin_Watts | I have to have something to occupy me while I repeatedly rebuild the epage stuff :( | 19:26.28 |
| mvrhel_laptop | Robin_Watts: so it is the win_desktop branch | 19:27.22 |
| Robin_Watts | ok. | 19:27.35 |
| mvrhel_laptop | gsview is the c# project | 19:28.12 |
| | MainWindow.xaml.cs is obviously the main viewer | 19:28.37 |
| Robin_Watts | oh, C#. Joy. | 19:29.15 |
| mvrhel_laptop | and we have an interface to gshostscript through ghostsharp.cs and through mudpf with mudocument.cs | 19:29.30 |
| Robin_Watts | OK. I'll see what I can make of it. | 19:29.38 |
| mvrhel_laptop | the interface to mupdf through mudocument shares the c++ code that I wrote for winRT by going through an interface DLL called mupdfnet which is included in the solution | 19:30.26 |
| marcosw | Robin_Watts: one of the files that epage won't read is a bad file (it's an .html document with a .doc extension, which word seems to happy to open). The other file appears perfectly okay, but epage says it doesn't recognize the header. I've opened a bug: http://bugs.ghostscript.com/show_bug.cgi?id=695032 | 19:34.14 |
| Robin_Watts | marcosw: Thanks | 19:34.26 |
| marcosw | should I make you the default assignee for ghostdocs bugs? | 19:35.22 |
| Robin_Watts | marcosw: Either me or paulgardiner. | 19:43.01 |
| | could we have a ghostdocs-bugs address that goes to the pair of us, as we currently do for mupdf-bugs? | 19:43.38 |
| | I thought henrys plans was that this stuff would belong to paul. I'm happy to be copied in etc, but I don't want to be 'shoehorning' myself in. | 19:44.24 |
| mvrhel_laptop | kens: ok it works now | 19:44.42 |
| | oops | 19:45.01 |
| | hold on | 19:45.03 |
| henrys | I think it is fine going to paulgardiner marcosw but I don't feel strongly about it. | 19:45.05 |
| | Robin_Watts: ugh Miles' email | 19:45.17 |
| Robin_Watts | henrys: I'm replying now. | 19:46.01 |
| mvrhel_laptop | darn. no it does not work. wtf | 19:46.02 |
| marcosw | Robin_Watts, paulgardiner, and henrys: I'll set up a ghostdocs-bugs alias that forwards to robin and paul. | 19:54.35 |
| henrys | marcosw: probably a good thing if you make the testing high priority. We are trying to make business decisions and understanding test results would be useful. | 19:56.49 |
| | marcosw: especially if a pcl bug comes in, this testing has higher priority ;-) | 19:57.46 |
| marcosw | henrys: I've opened bugs for all of the issues that I've found (there are currently 15 open ghostdocs bugs). | 20:04.51 |
| | There are probably additional issues that are obscured by the major differences that line spacing/word wrap causes (some of which seem to be much better with the newly modified fonts). | 20:05.39 |
| mvrhel_laptop | kens: ok now it works | 20:07.09 |
| | whew | 20:07.11 |
| | so we now avoid sticking in CIE color spaces as trans. group color spaces | 20:07.31 |
| | kens: I will clusterpush and if it looks ok commit | 20:08.26 |
| marcosw | okay, the ghostdocs-bugs alias is live and the existing bugs have been assigned to it. | 20:10.55 |
| | henrys: we've received the quality logic office2013 tests. They aren't very big, so I'll check them into the tests-private repository. | 20:12.46 |
| | I'll also run them through word2013 and ghostsdocs and post the comparison pdfs. | 20:13.04 |
| henrys | marcosw: okay when you get to a point where you can give some sort of opinion about the overall quality sound off. | 20:13.25 |
| | my macbook crashes periodically when moved, nothing written to the logs sigh ... | 23:57.06 |
| mvrhel_laptop | henrys: I almost have my laptop back to normal | 23:58.19 |
| | back to gsview now that I got the icc problem fixed | 23:59.39 |
| henrys | mvrhel_laptop: I wish I had a clue what was going on I did a hardware diagnostic and it's clean | 23:59.45 |
| | Forward 1 day (to 2014/02/12)>>> | |
IRC Logs