Log of #mupdf at irc.freenode.net.

Search:
 <<<Back 1 day (to 2019/09/11)Fwd 1 day (to 2019/09/13)>>>20190912 
quite the widgets of mupdf-gl remind me of Motif :)09:32.50 
ator quite: close, but they're actually modeled off of win95 :)09:41.03 
  motif has sharper edges on the bevels, and darker colors09:41.18 
quite heh :) how does that work, you use some widget engine for gl? or have to do it yourself?09:42.07 
ator quite: wrote my own very simple one09:43.18 
  it's an IMGUI so very easy to use and extend, but somewhat limited in fancy layout etc09:43.57 
  quite: http://git.ghostscript.com/?p=mupdf.git;a=blob;f=platform/gl/gl-ui.c#l744 that's the entire code for the button widget09:44.56 
quite immediate mode gui, ok, have heard the term but not quite understood the main difference from other gui engines. fewer layers of abstraction, maybe09:45.14 
ator main diff is in how you use it. with an IMGUI it's functional code. you never build up the gui with datastructures, no widget objects and event listeners and callbacks etc.09:46.27 
quite ator: aah, and ui_button is called as part of render-loop?09:54.38 
ator yeah. each frame "builds" the gui from scratch, and handles event during the "building"09:55.22 
  so the code calls if (ui_button("foo")) { do_foo() }09:55.34 
quite got it :)09:55.34 
ator Robin_Watts: ugh. this is annoying ... we can't really inline forms because then all transparency stuff get botched :(10:37.45 
Robin_Watts ator: Hmm.12:13.11 
  We could clone the xobject, and redact that, but that would leave the old xobject object still in the file, so not really redacted.12:14.06 
  Knowing whether an xobject is reused or not is hard...12:14.24 
  Need to (potentially) walk every page/annotation etc in the document.12:14.56 
  (and the content streams thereof)12:15.18 
ator Robin_Watts: or make a "clone" of each xobject every time it is invoked, and filter that instance for the invocation12:15.54 
Robin_Watts ator: Yes, but how do you know that the original xobject should be removed.12:16.54 
ator Robin_Watts: yeah. that's the tricky part :(12:17.18 
Robin_Watts ator: Yes, but how do you know whether the original xobject should be removed or not. (learn to speak english, robin)12:17.20 
  I am tempted to suggest that just removing the data from every instance would be the safest thing to do for now.12:18.01 
ator I think it may solve itself if we just don't copy the old XObject resource12:19.10 
  garbage collection would reclaim it when saving the redacted file12:19.18 
  and all the instances would be redacted (if they were part of the redaction annotation)12:19.38 
  if you redact from an instance on one page and not on another, then that's not our problem :)12:19.51 
Robin_Watts Suppose I have a 10 page file.12:19.53 
  every page uses the same xobject for the word 'ass'.12:20.18 
  No, wait, let me start again.12:20.41 
ator now if the file were crappy and each page has a resource dictionary containing all the xobjects in the file, whether or not they were used12:20.50 
  we could run into trouble12:20.53 
Robin_Watts 10 page file. All the pages share a resource dictionary that includes an xobject.12:21.00 
  only page 1 actually used the xobject.12:21.19 
ator I suspect we're talking about the same issue :)12:21.32 
Robin_Watts I think we are.12:21.36 
ator a page referencing an xobject that was not used, but redacted on another page, would leave the data in the file12:21.52 
Robin_Watts I think, for now, we should just redact the xobject and EVERY instance should be affected.12:21.52 
ator Robin_Watts: we could redact the same XObject multiple times, once per invocatien site (we just need to remember a list of matrixes where it was invoked on a page)12:22.34 
Robin_Watts Right, but that wouldn't help us with the removal of unused ones.12:29.34 
  I think, for redaction, we want to err on the side of caution.12:29.48 
  so for now, redact all of them.12:29.54 
  People will complain a lot more if we fail to redact something than if we redact the same information in several places.12:30.37 
ator Robin_Watts: yeah, we'd clean the XObject in-place just multiple times if invoked in multiple places12:40.11 
Robin_Watts right. nice.12:40.28 
ator at least that's my understanding of the filtering code12:40.38 
  gah, that API has grown into a wildvine...12:40.52 
dabeegmon Hello there - - -- not sure if anyone is out there14:09.19 
  I'm one of the weird types where I can have 40+ pdfs open across many of 20 desktops (running debian testing and love being able to set up things so they work for me!) just ran into issues with okular and am looking for a pdf viewed that is up to being loaded down like a supertanker and still functions well14:11.20 
sebras vtorri: ping?15:34.45 
  pink_mist: ping?15:35.08 
  you guys seem to have been lurking here for some time. does any of you know the maintainers of zathura?15:35.39 
vtorri sebras pong15:35.39 
sebras one of the devs used to be here, but I forget who it was. :)15:36.06 
vtorri sebras no, but my pet project is similar :)15:36.07 
sebras vtorri: do you have a website?15:36.20 
  vtorri: zathura's website pwmt.org appears to have vanished. :(15:36.42 
vtorri https://github.com/vtorri/etui15:36.48 
  sebras ha right, i've remarked that last week15:37.01 
  sebras maybe : https://github.com/pwmt/zathura15:37.44 
  or https://git.pwmt.org/pwmt/zathura15:38.09 
sebras vtorri: aha, https://pwmt.org/projects/zathura/ appeared to be dead, so I was utterly confused.15:39.24 
vtorri sebras i've updated mupdf from 1.12 to 1.16 and it's not working anymore :)15:40.56 
  no compilation pb but i may have forgotten something15:41.16 
sebras vtorri: thanks for providing impetus for including jbig2dec into oss-fuzz: https://github.com/google/oss-fuzz/pull/2816 ;)15:50.02 
pink_mist sebras: hi, sorry, no idea ... but seems vtorri was able to help =)15:54.19 
sebras yeah, thanks.15:54.59 
vtorri sebras https://github.com/moosotc/llpp/blob/master/LICENSE17:08.56 
  is it possible to have such license ?17:09.11 
  in Etui, i do like ffmpeg, i ensure that people configure with agplv3 flag to compile mupdf module17:09.57 
sebras vtorri: public domain is listed as a gplv3 compatible license: https://www.gnu.org/licenses/license-list.en.html#PublicDomain17:56.02 
  vtorri: if you link code which is public domain with code that is agplv3, the resulting binary is licensed under agplv3, which means that if a user asks for source code you need to provide the source code to not just the agplv3-part, but to the public domain part as well.17:57.12 
  at least that I how I, not being a lawyer, understand the license.17:57.21 
vtorri ok18:00.13 
 <<<Back 1 day (to 2019/09/11)Forward 1 day (to 2019/09/13)>>> 
ghostscript.com #ghostscript
Search: