| <<<Back 1 day (to 2016/05/24) | 20160525 |
Dylan | Hi | 08:34.52 |
ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 08:34.52 |
Dylan | any news about the license of ICC profiles ? | 08:35.38 |
kens | Why would there be ? | 08:36.07 |
Dylan | I asked last night some clarifications about the license | 08:37.12 |
kens | Hmm I didn't see that in the logs | 08:37.23 |
jogux | Henry asked chris & mvrhel to look at it | 08:37.39 |
kens | Yeah I see, well its much too early to coem back, really. Its going to be a few days | 08:37.58 |
Dylan | ok thank you | 08:38.32 |
kens | TBH as long as you don't pull the profile out and use it elsewhere, I can't see any legal problems anyway | 08:39.18 |
Dylan | In fact, I have packaged a software for Debian which contain an image with an integred ICC profile and they ask me to clarify the license for the ICC profile. | 08:45.28 |
kens | Seems a bit obsessive, but it'll get looked at. | 08:45.53 |
Dylan | The original discussion : https://lists.alioth.debian.org/pipermail/debian-med-packaging/2016-May/042197.html | 08:46.33 |
kens | If its urgent, you could try -dUseFastColor to produce the image, or render to something whcih doesn't contain ICC profiles, and use ImageMagick to convert the file format | 08:47.14 |
Dylan | it is not urgent, I prefere to not modify this image myself | 08:48.56 |
kens | Well, the most likely change is that we will modify the notice on the ICC profile, but that isn't going to help you for an existing image. | 08:49.33 |
| I'm only guessing mind you, since its not my field | 08:49.54 |
Dylan | It is not a problem, I will just add a comment to clarify the license in another file of my package | 08:52.27 |
kens | OK | 08:52.47 |
chrisl | Dylan: I don't understand what the problem is with the icc profiles | 09:18.03 |
Dylan | Debian will redistribute an image which contain an ICC profile with a copyright for the ICC profile but we don't know what is the license of this ICC profile. So, we don't know if we can redistribute it under a license compatible with Debian Free Software Guidelines. If the license of this ICC profile is not compatible with Debian Free Software Guidelines, I should remove it from the image before to redistribute. | 09:26.38 |
chrisl | Dylan: the files are shipped with Ghostscript, as part of Ghostscript. Are Debian now insisting we include a complete license for every file we ship? | 09:27.56 |
Dylan | it is just to be sure that the ICC profiles are under the same license that ghostscript | 09:31.03 |
chrisl | Dylan: there's been no query about using them and shipping them with Ghostscript for the last five years - why is it an issue now? | 09:31.48 |
jogux | chrisl: btw, the problem Dylan has is the software he is packaging (which is not ghostscript) contains a jpg file generated by ghostscript that has a Copyright Artfiex message in the embedded icc profile (aiui) | 09:31.58 |
chrisl | jogux: I know, I read the logs. I don't think we want to embed the entire license in each ICC profile, though | 09:32.45 |
Robin_Watts | I vote we do nothing, and let debian stew in their own dogmaticness. | 09:33.09 |
chrisl | Well, I can't see what we can do to satisfy them, short of bloat the files by including the license in each one | 09:33.56 |
jogux | is it intended that the resulting jpg file is AGPL licensed? (I never really understood what that actually means for a jpg file.) | 09:34.48 |
Robin_Watts | Fonts don't typically include the whole license file in. | 09:34.54 |
chrisl | The fonts don't mention the license at all | 09:35.23 |
Robin_Watts | Right. So how is this a problem, where that is not? | 09:35.51 |
| It's debian being asshat. | 09:35.57 |
| +s | 09:36.00 |
chrisl | Which is why I said I don't understand the problem | 09:36.08 |
Robin_Watts | chrisl: Indeed, I'm agreeing with you. | 09:36.34 |
jogux | I think one answer that works for Dylan is "the ICC profile is AGPL licensed, you can see this because it's the file <foo> in ghostscript.git and the AGPL license applies to it all". | 09:37.03 |
chrisl | At the moment we don't explicitly mention the iccprofiles directory in LICENSE, and I could add that - but I don't see how that solves Dylan's problem | 09:37.41 |
jogux | I think Dylan is okay so long as he can see/show/know that the file is AGPL licensed and hence under a debian compatible license. | 09:38.59 |
Dylan | jogux: this answer is enough for me | 09:39.00 |
| jogux: thank you | 09:39.31 |
jogux | dylan: just a sec | 09:39.35 |
| dylan: that wasn't an official answer :-) | 09:39.41 |
chrisl | Dylan: Sorry, I'm just getting highly p*ssed off with Debian doing this kind of thing | 09:39.55 |
Dylan | jogux: ok | 09:39.56 |
chrisl | Dylan: I've added the iccprofiles directory to the list in LICENSE - I still don't see how this changes anything from Debian's point of view, but...... | 09:44.03 |
| http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=9308777e | 09:44.06 |
| Dylan: We may do something further in the future (like grant permission for any use at all of the profiles), but for now....... | 09:45.35 |
Dylan | chrisl: thank you, it is ok for me | 09:47.43 |
chrisl | Dylan: To be clear: I still don't understand how that improves things. And to be doubly clear, if they start asking for the "source" of the ICC profiles, they're definitely out of lucj | 09:50.23 |
| luck | 09:50.30 |
jogux | robin: I'm curious about the #include <memory.h> in memento.h (the one in mupdf). xcode seems unhappy with it (mupdf/memento.h:164:10: error: include of non-modular header inside framework module 'MuPDF.mupdf.memento'). memory.h isn't part of C iirc? | 10:17.29 |
jogux | seems to have plenty of other issues too. cocoapods have mucked around with things again, breaking stuff that was working :-( | 10:18.34 |
Robin_Watts | jogux: Yeah, that's a bit crap isn't it? (memory.h). Let me check what I was on that day... | 10:38.31 |
| OK, so that line has been there since the initial import of memento into mupdf from gs. | 10:39.59 |
jogux | yeah, it's not a new problem. and builds fine in a normal iOS build. | 10:40.52 |
| it may be a symptom of other issues I have, but it seemed odd anyway. | 10:41.41 |
Robin_Watts | Ha. The commit message tells me I pulled it in from mupdf. | 10:45.01 |
| I suspect that #include <stdlib.h> would do. | 10:45.45 |
| tor8: I've run into an "interesting" issue with the alpha removal. | 10:47.28 |
| When I interpolate bitmaps, the calculations are carefully done so that sub pixel positioning is allowed for. | 10:48.13 |
| i.e. if I scale a bitmap that half overlaps a pixel along the top edge, the resultant scaled bitmap will have half the contributions used for the top edge. | 10:49.15 |
| This has historically worked nicely because we scale the alpha values too. | 10:50.12 |
tor8 | Robin_Watts: hmm. | 10:50.53 |
Robin_Watts | Now I'm ditching the alpha values where possible, scaled images are getting darker edges cos we're still getting half the contributions, but with full alpha. | 10:51.07 |
| I am tempted to think that I should reintroduce alpha when I scale. | 10:51.25 |
tor8 | Robin_Watts: alternatively, always scale to integer sizes | 10:53.35 |
Robin_Watts | tor8: That's gridfitting. | 10:53.51 |
tor8 | yes. and then do supersampling for type 3 fonts. | 10:54.01 |
Robin_Watts | and whether we gridfit or not has been tuned over a while and I'd rather not upset it. | 10:54.12 |
| I am going to try reintroducing alpha for the non-gridfitted cases. | 10:54.29 |
| That way we should get the best of both worlds. | 10:54.48 |
tor8 | okay. | 11:00.29 |
klingo | Hi, does anyone here happen to know if the android version of muPDF has an 'invert colours' mode? | 11:55.16 |
| If it does, I can't figure out how to enable it! | 11:55.28 |
| I'm guessing it does have one, but I don't know how to get the keyboard to pop up to press 'i' and choose invert! | 12:03.50 |
Robin_Watts | it does not. | 12:04.20 |
| MuPDF is a portable C library, with various tools (such as viewers) built around it. | 12:04.45 |
| The Android viewer is one such thing. | 12:04.50 |
| The C lib is capable of doing inversion. | 12:04.57 |
| The Android viewer doesn't use that option though. | 12:05.07 |
klingo | Ah bugger, thanks for the info :) | 12:07.11 |
jogux | Robin_Watts: So xcode really was getting upset over memory.h, even after I fixed all my other issues. It's only when building mupdf as a llvm module it whinges. Trivial commit on my master. I've only tested it in xcode. | 13:06.37 |
| Robin_Watts: http://git.ghostscript.com/?p=user/joseph/mupdf.git;a=commitdiff;h=855227465e8c0fe054f0de662968be558baaf840;hp=6d3c81d9465b24d3233188dc335c226950c06539 | 13:06.55 |
tor8 | jogux: LGTM, but Robin will have to weigh in on any potential windows issues | 13:16.21 |
jogux | thanks tor8. I think it should be okay for Windows but will wait on Robin. | 13:17.14 |
Robin_Watts | jogux: lgtm, but I'll test it here before we push. | 13:30.00 |
jogux | thanks Robin_Watts | 13:30.10 |
Robin_Watts | and that'll have to wait until I get stuff back together :) | 13:30.17 |
jogux | :-) | 13:30.27 |
| HenryStiles: mupdf podspec is updated for 1.9 release now btw (and also for the cocoapods 1.0 release, which was an absolute PITA) | 15:13.30 |
| I believe it won't work for projects that use swift, but that's really a wider issue. | 15:13.59 |
sh4rm4^bnc | how can i get a hyperlink url (or copy it to clipboard) with mupdf-x11 ? | 20:40.14 |
Robin_Watts | Drag with the right mouse button? | 20:45.45 |
benbro | can I open wbd files with ghostscript? | 21:03.47 |
| example: | 21:04.14 |
| http://www.virtualhomeschoolgroup.com/file.php/1/03_Science/02_Middle_School_Courses/Physical_Science/M02/Study_Tools/Game_Day/Physical%20Science%20M02%20Game%20Day.wbd | 21:04.16 |
| it's a whiteboard xml | 21:04.25 |
Robin_Watts | no | 21:18.00 |
benbro | thanks | 21:22.49 |
| the format seems simple to convert with a script | 21:23.18 |
| I just don't understand this element | 21:23.25 |
| <Path len="42">,,1,1,2,1,3,2,3,3,3,3,3,4,4,4,4,5,5,5,6,6,6,7,7,8,9,9,9,10,9,10,9,10,9,11,10,11,10,12,10,11</Path> | 21:23.28 |
| what the missing comma means... | 21:23.36 |
Robin_Watts | benbro: Possibly it's a format that could be added to MuPDF. | 23:24.30 |
benbro | Robin_Watts: what do you mean by added to MuPDF/ | 23:25.31 |
| ? | 23:25.32 |
Robin_Watts | benbro: MuPDF supports "document handlers". | 23:25.45 |
| PDF is handled by one, XPS by another, SVG another. | 23:26.02 |
benbro | there seem to be only few element types | 23:26.06 |
Robin_Watts | epub, images, etc. | 23:26.08 |
benbro | nice | 23:26.10 |
| but I don't understand how to convert this element | 23:26.21 |
| <Path len="42">,,1,1,2,1,3,2,3,3,3,3,3,4,4,4,4,5,5,5,6,6,6,7,7,8,9,9,9,10,9,10,9,10,9,11,10,11,10,12,10,11</Path> | 23:26.25 |
Robin_Watts | Which file are you looking at? | 23:26.30 |
| I suspect a missing number counts as zero ? | 23:27.00 |
benbro | http://www.virtualhomeschoolgroup.com/file.php/1/03_Science/03_Biology/01_General_Biology/M01_Study_of_Life/L3_Biological_Classification/Published/wbd/Lecture_3_Biological_Classification.wbd | 23:27.03 |
| http://pastebin.com/Wv4a38cU | 23:27.48 |
| you have a pentool, the bounding box (rect), stroke width, color | 23:28.06 |
| and the path | 23:28.09 |
Robin_Watts | I'd bet that list of numbers translates as 0,0,1,1,2,1,3,.... | 23:34.44 |
| but I can't figure out what the digits mean. | 23:34.54 |
benbro | missing could mean 0. didn't think of that | 23:35.26 |
| I don't know what the digits are | 23:35.35 |
| doesn't look like svg when you have move-to, line-to... | 23:35.51 |
Robin_Watts | Do you have software to read these files? | 23:36.57 |
benbro | I'm trying to register for blackboard collaborate trial | 23:37.21 |
| it requires java on windows or mac | 23:37.27 |
| http://try.bbcollaborate.com/trial/register.go | 23:37.40 |
| I think 'blackboard plan' can read it too | 23:38.00 |
Robin_Watts | If you get something that can read the files we can hand craft some and see what it does. | 23:38.11 |
benbro | https://en-us.help.blackboard.com/Collaborate/v12/Moderator/Blackboard_Collaborate_Plan | 23:38.41 |
| great. I'll try installing it in a windows 7 machine | 23:38.54 |
| need to install some stuff first | 23:38.58 |
| Forward 1 day (to 2016/05/26)>>> | |