| <<<Back 1 day (to 2012/06/07) | 2012/06/08 |
mvrhel | Robin_Watts: maybe. I think I only used it when I added the stack. From what you found I should have used it more though ;) | 00:17.12 |
| bbiaw | 00:23.29 |
Ch3rryC0ke | Hey there-- I'm wondering if anyone has tried and/or had success in using Ghostscript in a Win 8 metro app? | 01:08.25 |
mvrhel | good grief. I was planning to use gx_default_copy_alpha as a template for the devn color case but it has macros stacked upon macros | 05:41.37 |
kens | henrys ray_laptop (for the logs) Its entirely possible that HighLevelDevice isn't appropriate in an y given case. In fact pdfwrite and ps2write have a slew of parameters which are used to control (amongst other things) the behaviour of the PDF interpreter. For examples, see also AllowIncrementalCFF, IsDistiller, WantsToUnicode. | 07:20.13 |
| If you feel you need another paramter to control somethign then that seems sensible. However I will alter gdevvec.c so that all devices based on it are 'HighLevel'. | 07:21.00 |
| Currently 'HighLevelDevice' is used so that we don't set the default halftone if it is true. This is because if we do, for a file with no halftone in it, we end up with an inappropriate halftone being set (pdfwrite/ps2write can't tell its the default and emit it). | 07:23.25 |
Robin_Watts | just loves how $849 translates to £800 in consumer electronics. | 10:36.52 |
| I fear Marcosw may be getting a siny new camera delivered to him before the next meeting. | 10:39.35 |
| s/siny/shiny/ | 10:39.42 |
| chrisl: We use libpng 1.2.44 in gs. Latest version is 1.2.49 which addresses a "serious vulnerability" | 10:50.48 |
kens | must have missed that question | 10:51.49 |
Robin_Watts | kens: no question. I've just been to the libpng homepage (as part of fixing Gigs's bmpcmp bug) and spotted that there is a large red section of text about this vulnerability. | 10:53.01 |
| and thought it was worth reporting the new version to chrisl. | 10:53.12 |
kens | Ah... | 10:53.42 |
chrisl | Robin_Watts: OKay, thanks, I'll look into it - although, given how we use it, I think it would be hard to exploit such vulnerability | 11:00.01 |
Robin_Watts | chrisl: Probably true. | 11:00.11 |
chrisl | Actually, we should probably update to 1.5.x - I think we can use the newer releases, now...... | 11:01.02 |
| I'll have a proper look at it after I've fed the cat - and myself...... | 11:03.04 |
Gigs- | heh yeah don't do what MS did | 12:48.43 |
| IE had broken PNG rendering for years and years because they forked an ancient libpng (and closed the source, and didn't acknowledge their use of the code in true MS fashion) and never updated it | 12:49.32 |
Robin_Watts | Power cut central here. | 13:22.55 |
| git subtree is now an official part of git, apparently. | 13:30.45 |
chrisl | How long before we can rely on it, though...... | 13:31.23 |
Robin_Watts | probably about as long as it will take us to actually do anything about the stuff that might need it :) | 13:44.19 |
| chrisl: ping | 14:55.01 |
chrisl | Robin_Watts: pong | 14:55.11 |
Robin_Watts | I've just added memento to the standalone jbig2dec build. | 14:55.22 |
chrisl | Okay | 14:55.51 |
Robin_Watts | Shelly was wanting to use jbig2dec to look for leaks rather than having to make lots of PDF files. | 14:55.54 |
chrisl | I'll ping him to let him know | 14:56.09 |
Robin_Watts | Do you want to look over my changes before I commit them? | 14:56.10 |
| (As keeper of the build system etc) | 14:56.18 |
chrisl | No, just go ahead | 14:56.32 |
Robin_Watts | ok. | 14:56.37 |
chrisl | I trust you :-) | 14:56.43 |
Robin_Watts | The only change I've made to gs is that when it builds the jbig2 stuff if does -DGSBUILD | 14:57.07 |
| so the jbig2dec code doesn't include the wrong version of memento. | 14:57.31 |
chrisl | Sounds fine | 14:57.53 |
Robin_Watts | cool, thanks. | 14:57.58 |
chrisl | Robin_Watts: we're up to date with libpng now - turned out to be a little more involved than I'd hope, but it's done now..... | 14:59.05 |
Robin_Watts | With 1.5 or 1.2 ? | 14:59.14 |
chrisl | 1.5 | 14:59.19 |
Robin_Watts | Ah, excellent! | 14:59.24 |
chrisl | I'd forgotten XPS *reads* png files! | 14:59.48 |
| Oh, that's a thought..... | 15:00.33 |
| tor8: ping | 15:00.36 |
Robin_Watts | tor8: ping too :) | 15:26.17 |
tor8 | chrisl: Robin_Watts: pong. | 15:46.31 |
Robin_Watts | tor8: Just to say I'm going to have a quick look at bug 693097 so we don't both start looking at it. | 15:47.03 |
chrisl | tor8: I made some changes in the ghostpdl/xps code to support libpng 1.5.x - you might want them for muxps | 15:50.31 |
tor8 | chrisl: muxps has its own png parser, but thanks for the heads up | 16:09.23 |
henrys | hintak litigation sigh... | 16:10.45 |
Robin_Watts | litigation? | 16:11.39 |
henrys | he just seems to argue endlessly. | 16:12.04 |
Robin_Watts | ah, right, yes. | 16:12.10 |
| Are the pxl devices high level ? | 16:12.40 |
henrys | they are vector devices? | 16:13.03 |
Robin_Watts | that's what I'm asking. | 16:13.16 |
henrys | that was a statement not a question | 16:13.18 |
Robin_Watts | right. | 16:13.24 |
kens | henrys I'll bisect the problem from Stefan Leiterman tomorrow. Its OK in the current code but fails on GS versions up to 9.04 | 18:22.57 |
| Heading off now, goodnight all | 18:23.17 |
henrys | oh okay | 18:23.27 |
mvrhel | bbiaw. | 18:30.32 |
henrys | I guess on the windows build removing devices requires manually editing msvc.mak, is that correct or am I missing something. | 19:05.41 |
| ? | 19:05.48 |
Robin_Watts | henrys: Yes. | 19:05.52 |
| How else would you like it to be done? | 19:06.10 |
| (it would be nicer if both unix and msvc builds were controlled by a single list...) | 19:06.58 |
henrys | linux and mac can have a list... or they can take collections of devices (i.e. PRINTERS) or single devices all set up with with configure. | 19:09.18 |
Robin_Watts | but by default they get them from... devs.mak ? | 19:09.59 |
| no... | 19:10.25 |
henrys | see configure.ac | 19:10.41 |
| search for --with-drivers | 19:10.57 |
Robin_Watts | unix-gcc.mak, unixansi.mak, openvms.mak, macosx.mak, macos-mcp.mak all have huge lists | 19:11.56 |
| msvc.mak is in keeping with them. | 19:12.04 |
henrys | well the real list is generated by configure and placed in the Makefile each time configure is run. | 19:13.00 |
Robin_Watts | Right. So the split is configure'd platforms vs non-configured makefile platforms. | 19:13.41 |
henrys | I suppose chrisl_away is keeping that stuff around so folks can compile without configure. | 19:13.48 |
marcosw | sure is quiet around here. just thought i'd say hello from Panama City⦠| 19:48.13 |
Robin_Watts | marcosw: Hey. | 19:48.22 |
| Where in Panama City? | 19:48.28 |
marcosw | at the airport. our flight leaves in 45 minutes. | 19:48.48 |
Robin_Watts | Ah, right. No point in recommending a restaurant then :) | 19:49.01 |
marcosw | oops, got to run to the gate, flight leaves in 30 minutes. | 19:49.25 |
mvrhel_laptop | henrys you around? | 20:06.09 |
| henrys: never mind. just wanted to remind you that I am gone next week. I sent an email to tech | 20:52.02 |
Gigs- | Robin_Watts: I'm getting a lot of missing files in the libpng directory | 21:33.22 |
| Robin_Watts: trying to compile the new bmpcmp | 21:33.29 |
| mv gs/libpng/scripts/pnglibconf.h.prebuilt gs/libpng/pnglibconf.h | 21:36.58 |
| and two missing files on the suggested build command line at the top of bmpcmp | 21:37.16 |
| but it compiled now | 21:37.23 |
Robin_Watts | Gigs: Fab. Hopefully it's working for you now. | 23:18.10 |
| Forward 1 day (to 2012/06/09)>>> | |