| | <<<Back 1 day (to 2011/09/29) | 2011/09/30 |
| chrisl | kens: the problem with Bug 692561 is in pdf_ops.ps in the procedure "setshowstate" | 08:18.46 |
| kens | Yes, I figured that much | 08:28.37 |
| | Your fix might correct my problems with pdfwrite and Tr 3 :-) | 08:28.55 |
| chrisl | Oh, back again..... :-) | 08:45.45 |
| kens | For now..... | 08:45.53 |
| chrisl | So, that tr 3 thing is a funny one - I'm really surprised it hasn't come up before | 08:46.37 |
| kens | Well, as I said in the thread, I think its rare for a marking operation to rely on the currnepoint being moved by a non-marking (3 Tr) operation | 08:47.07 |
| | I think thre's still a different problem with the pdfwrite usage though :-( | 08:47.40 |
| | Still, good to get it fixed. | 08:47.53 |
| | Did you notice that it affected *all* the text ? It just wasn't obvious unless there were two characters together ;-) | 08:48.19 |
| | Hmm, this means I'm goign to end up having to redo all my changes to pdf_ops.ps when I come to commit them, oh well. | 08:48.56 |
| chrisl | I noticed the problem affected most of the text, I was comparing with mupdf | 08:49.12 |
| kens | Ah, I was comparing with Acrobat | 08:49.21 |
| chrisl | Do your changes clash with mine? | 08:49.38 |
| kens | No, but I am modifying pdf_ops.ps so I'll need a merger | 08:50.02 |
| | merge | 08:50.07 |
| chrisl | Yeh, but if you commit locally, then do a pull/rebase, if there isn't a clash, it should merge automatically | 08:50.40 |
| kens | Yeah, and I trust that ? ;-) | 08:50.49 |
| | I'll do it manually :-) | 08:50.59 |
| chrisl | git is *much* better at it than svn/perforce - I still check the end result, but it rarely fouls up for me. | 08:51.46 |
| kens | OK your changes are not affecting my pdfwrite problems, shame :-( | 08:54.23 |
| | If its pdfwrite we never enter that piece of code. | 08:54.42 |
| | So, back to figuring out what pdfwrite is up to.... | 08:54.53 |
| chrisl | I'd expect it to be doing something not too different - the "show" that I've sorted out the currentpoint after seems rather redundant if not going to a vector device! | 08:56.07 |
| kens | Its quite different, the current point is nonsense after pdfwrite executes a 3 Tr operation. | 08:57.08 |
| | But pdfwrite uses a totally different set of text rendering procs, and I've heavily modified those as well | 08:57.28 |
| chrisl | Hmm, so I wonder why it does a show there at all...... | 08:58.18 |
| kens | Wel, I think ps2write might quite like it ;-) | 08:58.40 |
| | Although I'm not certain about that, it may handle text rendering modes itself as well. I should check | 08:59.17 |
| chrisl | But ps2write only needs the currentpoint updated for Tr 3, doesn't it? | 08:59.50 |
| kens | I would think so, yes. | 09:00.04 |
| | Can't think of any reason for emitting the actual text | 09:00.14 |
| chrisl | Unless it spits out some pdfmark magic for hidden text | 09:00.29 |
| kens | Well it 'might' set Tr 3 and the prolog might handle that, which is waht I meant when I said I'd have to check./ | 09:01.02 |
| | It might be why I got so many diffs on my last run. | 09:01.12 |
| chrisl | and txtwrite I would expect to be treated as a vector device as well, so would use the pdfwrite procs | 09:01.43 |
| kens | It does. | 09:02.10 |
| | One of its outptu modes write the text rendering mode as part of the text attributes. | 09:02.36 |
| chrisl | Well, it does sound the "show" is pointless, but it's easy enough to fix, so I'm going to leave it there. | 09:03.17 |
| kens | It seems harmless for now, I'd be inclined to pass it to Alex and let him decide. | 09:03.37 |
| chrisl | At this point, my inclination is to leave it - it is remotely possible that there is code out there in the wild that expects it now. | 09:16.36 |
| kens | Seems reasonable. | 09:16.47 |
| chrisl | kens: thanks for cutting the file down, there, it made things much easier! | 09:37.45 |
| kens | Didn't realise you'd used the reduced file.... | 09:40.56 |
| chrisl | Yeh, I'd just got the first page alone, and decompressed and was about to start hacking when the bugzilla mail came through with your cut down file - great timing | 09:44.22 |
| kens | :-) | 09:44.33 |
| Robin_Watts | tor8: Did you see the mail from Company K ? | 11:06.06 |
| | It was about text extraction, so I left it to you to answer. | 11:06.26 |
| tor8 | Robin_Watts: any ideas for how to improve it? the current method uses a distance based on the font size as a trigger, in the absence of a space character. I believe kens used a distance based on the width of another character (was it 'i'?) for the textwrite device | 11:09.55 |
| Robin_Watts | tor8: Anything you do is going to be a heuristic. | 11:10.41 |
| tor8 | the only "hard" solution is to get the producer to output space characters | 11:11.02 |
| kens | tor8 I'm using the average of all the glyphs in the text fragment | 11:11.18 |
| Robin_Watts | I'd be tempted to take the minimum of whatever value you have now for the estimated space width, and the width of the 'i' character. | 11:11.27 |
| tor8 | if there is an 'i' character in the subset font :) | 11:11.59 |
| | but yeah, we could try that | 11:12.11 |
| Robin_Watts | I'd personally be tempted to think that a gap counts as a space if it's > 1/2 the average width of a char, and less than 2 average widths of a char. | 11:13.00 |
| kens | that's what I use ;-) | 11:13.09 |
| | But only for joiing text fragments | 11:13.31 |
| Robin_Watts | But ideally that logic should be kept separable in the code so that other people can alter it to do smarter stuff if they want. | 11:13.59 |
| | (I could imagine people using dictionaries for specific languages to decide whether stuff forms into words or not) | 11:14.47 |
| kens | Arabic and CJKV are easy | 11:15.19 |
| | Hmmm I can't use the swapcolors_quick method if the colour I'm swapping to is a pattern colour :-( | 13:36.36 |
| Robin_Watts | Why not ? | 13:36.58 |
| kens | It crashes pdfwrite | 13:37.05 |
| | Probably because pdfwrite captures patterns as high level spaces. | 13:37.28 |
| Robin_Watts | swapcolors_quick is intended to be used for you to swap the colors, do something, then swap back. | 13:37.37 |
| kens | It GPFs | 13:37.49 |
| | Or rather, pdfwrite GPFs trying to use the swapped colour | 13:37.59 |
| Robin_Watts | where "do something" is "do something simple" | 13:38.07 |
| kens | All I do is emit the colour, the colour emission is what barfs | 13:38.27 |
| | pdc->colors.pattern.p_tile is NULL | 13:39.23 |
| | pdc is of type gx_drawing_color | 13:39.39 |
| Robin_Watts | That's probably because the dev_color hasn't been set ? | 13:39.47 |
| | (Definitely, if you intend to do any drawing operations, you should use gs_swapcolors) | 13:40.19 |
| kens | I thought that was the point of gx_remap_color | 13:40.20 |
| Robin_Watts | Right. So you gs_swapcolors_quick, then call gs_set_dev_color (which calls gx_remap_color) ? | 13:40.44 |
| kens | Hmm, I was calling gx_remap_color directly. | 13:41.06 |
| | You think I should call gs_set_dev_color ? | 13:41.16 |
| Robin_Watts | I would personally have called gs_set_dev_color, cos it's what all the rest of the code does. | 13:41.31 |
| kens | Well that seems a reasonable reason | 13:41.46 |
| | I can do that easily enough its only 2 changes | 13:42.06 |
| Robin_Watts | gx_set_dev_color, rather. | 13:42.11 |
| | but... I can't see that that will solve your problem. | 13:42.20 |
| kens | Ah, that explains why I can't find gs_set_dev_color..... | 13:42.31 |
| Robin_Watts | unless the !color_is_set thing saves you. | 13:42.53 |
| kens | Might do, I'll try it. | 13:43.04 |
| | I don't understand the internal colour stuff at all. The PS end I can handle | 13:44.03 |
| | Well, it didn't crash | 13:44.25 |
| | And the output looks OK too :-) | 13:44.48 |
| Robin_Watts | Don't look directly into it's eyes! | 13:45.17 |
| | Just back away slowly... | 13:45.21 |
| kens | No, time for another cluster push. | 13:45.31 |
| | Will still (probably) leace 2 files with seg faults and a slew of other problems. | 13:45.46 |
| Robin_Watts | I'm in clist hell. | 13:46.00 |
| kens | On the plus side, a test file which never worked is much improved, and teh outptu is much more efficient | 13:46.03 |
| Robin_Watts | It's adjacent to rop hell. Less goo, but more sharp corners. | 13:46.16 |
| kens | In some ways its worse, some of it is written in PostScript.... | 13:46.43 |
| | Bizarrely the Tr 3 problem I was having was completely different from the Tr 3 problem Chris and I looked at this morning, but the effect was similar | 13:47.09 |
| | I now have so may windows open I can't find anything... :-( | 13:47.38 |
| Robin_Watts | kens: What editor do you use? | 13:47.58 |
| | kens: You need to buy another monitor. | 13:48.21 |
| kens | I think you're right about the monitor, or possibly a virtual desktop thing | 13:49.09 |
| Robin_Watts | I find 2x 1920x1200 works nicely. | 13:49.29 |
| kens | I sue lots of different editors from Notepad to VS. Depends what I'm doing | 13:49.29 |
| | err, use, not sue.... | 13:49.45 |
| Robin_Watts | 1920x1200 is about right for VS. And that leaves another one for chatzilla/console2 etc. | 13:49.57 |
| kens | Git, Firefox, Acrobat, news reader,..... | 13:50.37 |
| Robin_Watts | Git runs in Console 2. | 13:51.04 |
| | Firefox I pop up when I need it (often on the second monitor, so I can copy/paste into VS etc) | 13:51.28 |
| | Acrobat (on whichever monitor doesn't have the ghostscript output I'm comparing to). | 13:51.59 |
| kens | ponders second monitor | 13:51.59 |
| Robin_Watts | kens: Does your graphics card have a second output? If so, it's a no brainer. | 13:52.23 |
| kens | crawls on the floor to find out. | 13:52.43 |
| | Looks like it doesn't, the other plug looks like VGA | 13:53.20 |
| Robin_Watts | Lots have 1 VGA, 1 DVI. | 13:53.33 |
| kens | Yes, that looks like it | 13:53.47 |
| Robin_Watts | Both of my monitors are connected via VGA. | 13:53.55 |
| | (through DVI -> VGA adapters). | 13:54.13 |
| kens | I wonder if I could stick a second one on the VGA output, I have an old CRT here somewhere.... | 13:54.22 |
| Robin_Watts | I have a KVM switch on one that works on VGA. | 13:54.26 |
| | kens: Almost certainly. Got to be worth a try. | 13:54.39 |
| | What graphics card is it? | 13:54.43 |
| kens | ATI something or other. | 13:54.54 |
| Robin_Watts | Ah, right. Well the ATIs are very good at handling dual monitors and have been for ages (I've had many ATI cards, and done dual head with all of them) | 13:55.22 |
| kens | 256 Mb ATI Radeon X1200PRO | 13:55.23 |
| | 1300 | 13:55.33 |
| | Or possibly its just time to invest in a new PC | 13:56.19 |
| Robin_Watts | Time for Uncle Miles to buy you an i7? | 13:57.12 |
| kens | Maybe, but I hate changing computers, and it would come with Windows 7.... | 14:01.21 |
| Robin_Watts | Windows 7 seems like a brilliant update to vista. | 14:01.36 |
| | I'm reluctant to move from Windows XP :) | 14:01.49 |
| kens | I'm unconvinced, it hides even more stuff | 14:01.50 |
| | And lies. | 14:01.56 |
| Robin_Watts | kens: I think a 2nd monitor will 'just work' with what you have. | 14:03.20 |
| kens | I may well try it. Of course I don't want to use a CRT long term so I'd have to invest in a second monitor. Do you kow if the two can be different resolutions ? | 14:04.01 |
| Robin_Watts | Yes. | 14:04.11 |
| | different resolutions/timings/color depths. | 14:04.21 |
| kens | OK, then I guess I'd better go lug that old one out. | 14:04.27 |
| Robin_Watts | http://www.amazon.co.uk/Samsung-2443BW-24-Inch-Monitor/dp/tech-data/B001EX01YM/ref=de_a_smtd | 14:06.53 |
| kens | Hmm, my screen resolution changed. | 14:07.26 |
| Robin_Watts | kens: The driver may currently be set to 'clone' the output. | 14:08.34 |
| | so it might have resized down to a res that both can cope with. | 14:08.48 |
| kens | I don't think so it didn't have teh scond monitor then | 14:09.39 |
| | But it does now. | 14:09.44 |
| | OK now my main is back to 1920x1200 | 14:10.26 |
| | second is at 1280x1024 | 14:10.44 |
| | Not sure what it will support | 14:10.51 |
| Robin_Watts | and you can drag windows between the two OK? (i.e. it's not cloning) | 14:11.09 |
| kens | not sure yet | 14:11.22 |
| | Yes I can drag across. Second monitoir at 1600x1200 which is tops. Obviously it reconfigured the primary to be the same | 14:12.00 |
| Robin_Watts | Cool, sounds like you're sorted then. | 14:12.54 |
| | I warn you, this will ruin ever working on a laptop for you :) | 14:13.05 |
| kens | I already feel that way, VS is well-nigh impossible on a laptop display | 14:13.41 |
| | I guess I should pick this monotir up off the floor and find some desk space. | 14:13.56 |
| | Not looking forward to lifting it up. | 14:14.09 |
| Robin_Watts | Mind your belt buckle :) | 14:14.48 |
| kens | I'd forgotten how much these things weigh... | 14:15.09 |
| Robin_Watts | My big iiyama monitor came with a large sign on the screen warning that people often scratch the screen with the buckle as they lift 'em. | 14:15.22 |
| kens | Hmm, now I have trouble remembering where I left the mouse ;-) | 14:22.31 |
| Robin_Watts | kens: Windows has various things for that. | 14:22.49 |
| | One will give the mouse pointer a trail to make it easier to spot. | 14:22.59 |
| | Another will locate it with rings when you hit a key. | 14:23.10 |
| kens | yes, I think the first thing is to make the text bigger.... | 14:23.29 |
| | Ore esolution smaller.... | 14:23.43 |
| | Its a bit flickery. | 14:23.49 |
| Robin_Watts | Probably best to lower the res on the CRT then, cos you'll get larger text and faster refresh. | 14:24.13 |
| kens | Yes, that's what I'm working on | 14:24.31 |
| Robin_Watts | You'll probably find that it's been set up as 'default monitor'. | 14:24.31 |
| | IF you change that to the right monitor definition you'll get better results. | 14:24.46 |
| kens | Its already correct, auto detected it. | 14:25.20 |
| Robin_Watts | oh, ok. DCC is working then. | 14:25.33 |
| kens | OK 85 Hz, now I can look at it | 14:26.06 |
| | Without my eyes hurting | 14:26.30 |
| | Low resolution though, to think I once used this as my primary display ;-) | 14:26.48 |
| | That's better, now the screen is on the left and so is the desktop | 14:27.24 |
| Robin_Watts | I lugged a large CRT monitor up onto the desk the other day to use with the board from company I, and it failed to cope with 1024x768. | 14:27.48 |
| kens | !!! | 14:28.06 |
| | THis ancient one is running at that | 14:28.13 |
| Robin_Watts | Time to throw that one away, and reclaim several cubic feet of office space methinks. | 14:28.20 |
| kens | Well yes, except that it wasn't previously occupying any space in my office, its my emergency backup display | 14:28.43 |
| | And was stored away | 14:28.51 |
| Robin_Watts | I meant my one. I have 2 such backup CRTs, and that's more than I need. | 14:29.30 |
| kens | Cool, using gx_set_dev_color seems to have got rid of all my seg faults. | 14:29.43 |
| Robin_Watts | woo hoo. | 14:30.06 |
| kens | Its a good start. | 14:30.16 |
| | Now I have to look at the 298 diffs and then figure out why they have the wrong colours. | 14:30.48 |
| | especially ps2write, which I expetced to be unchanged. | 14:31.18 |
| Robin_Watts | This may be a really dumb question, but.... | 14:45.53 |
| | when using the clist, should I expect the offsets given by -ZL during the writing phase to match up with those given during the reading phase ? | 14:46.26 |
| kens | No idea, sorry. | 14:50.55 |
| | But it seems likely to me | 14:51.06 |
| tkamppeter | chrisl, I am seeing your LCMS patch. You are also deactivating the option to use LCMS2 shared. Do you also have very many fixes there? | 15:06.11 |
| | chrisl, and is upstream ignoring or refusing them? | 15:06.27 |
| Robin_Watts | tkamppeter: Our lcms2 is currently 'stock', I believe. | 15:06.59 |
| chrisl | tkamppeter: We have problems with lcms2 that haven't been addressed at all yet. | 15:07.04 |
| Robin_Watts | But using lcms2 at all is experimental at this stage. | 15:07.55 |
| | For at least one file, it goes wrong. | 15:08.02 |
| | For others it will be slower than our lcms1 (significantly so). | 15:08.21 |
| | When we get the lcms2 problem fixed, we will port our lcms1 changes to lcms2, and post them upstream. | 15:08.52 |
| tkamppeter | Robin_Watts, so then I will use shared lcms2 during the development cycle of Ubuntu 12.04 to get a broader testing. If it performs too badly and produces too many reported bugs which do not get fixed in time I will pull back to built in lcms1 shortly before Ubuntu 12.04 release. | 15:10.16 |
| Robin_Watts | What happens during the development of 12.04 is of less concern to me than what happens during the imminent release (which is 11, right?) | 15:11.37 |
| tkamppeter | Robin_Watts, the current release is 11.10 to go out on Oct 13. | 15:15.34 |
| Robin_Watts | Right. And in that, you are shipping gs built against a stock lcms1, right ? | 15:15.54 |
| tkamppeter | Robin_Watts, I would like to use the embedded lcms1 for that release, but they do not let me. | 15:16.03 |
| Robin_Watts | As long as they realise that that means any bugs they report will be ignored. | 15:17.04 |
| tkamppeter | Right, that's it, and they do not let me switch to the embedded one, which I wanted to do to address the Apple-generated figures bug, and which I also want to do now to fix other possible segfaults and color quality shortcomings. | 15:18.05 |
| | Robin_Watts, or are the problem only segfaults? | 15:18.25 |
| Robin_Watts | both segfaults and, I believe, color correctness. | 15:18.44 |
| tkamppeter | Robin_Watts, are the colors generally off with the stock liblcms1, defeating CM completely or ar these glitches here and there? | 15:28.02 |
| Robin_Watts | tkamppeter: mvrhel2 is the authority on the color fixes that have been done to lcms, but I *believe* it's more than some files will use profiles in a way that lcms1 doesn't support properly, and radical colour differences will be given for those files. | 15:29.04 |
| | The larger problem with lcms1 (in my eyes) is the lack of checking for allocation failures. | 15:29.30 |
| | If you run gs in a position where you are running low on free memory, and an lcms allocation fails, it will SEGV. | 15:30.06 |
| | (because it will assume the allocation succeeded and then indirect through NULL). | 15:30.34 |
| | We've fixed at least some of those problems in our version. | 15:31.02 |
| tkamppeter | Robin_Watts, trying to convice them to use your fixed lcms1 in GS at least for an update package for 11.10. | 15:56.21 |
| Robin_Watts | tkamppeter: Cool. | 15:56.54 |
| henrys | tkamppeter:just curious how this is organized, when you say "convince them" who are you referring to? | 16:15.43 |
| tkamppeter | henrys, once the technical lead of the Ubuntu Desktop team and also a person of the release team. | 16:26.24 |
| henrys | kens did NoSu talk to you about pdfmark and modifying the cropbox? he was looking for you. | 16:44.57 |
| tkamppeter | henrys, the Ubuntu guys left for the weekend now. I will try to move it on next week. | 16:48.30 |
| chrisl | tkamppeter: is there anything *we* could do to help you make the case? | 16:50.12 |
| kens | henrys NoSu didn't, but someone else did | 16:50.40 |
| | Heading off now, have a good weekend all | 16:51.36 |
| henrys | take care kens | 16:51.45 |
| Robin_Watts | you too kens | 16:51.47 |
| tkamppeter | chrisl, perhaps you could directly e-mail some people. I can give you the mail addresses in a private dialog. | 16:52.34 |
| | henrys, ^^ | 16:52.44 |
| Robin_Watts | Would it help for us to produce a definitive statement of what the situation is with lcms etc? | 16:54.02 |
| mvrhel2 | good morning | 16:54.34 |
| Robin_Watts | i.e. describe the kind of fixes we've done, why we haven't moved to lcms 2, what we hope to do etc, what the problem with shared libs in this instance is ? | 16:55.12 |
| | morning mvrhel2, ray_laptop | 16:55.27 |
| | ray_laptop: I'm stuck in clist hell. | 16:55.34 |
| henrys | tkamppeter:this does seem rather odd to me. What if we made the ubuntu packages would the same, I think we used to make an rpm package - would you use it as is? | 16:56.30 |
| | s/would the same/would we have the same issues/ | 16:56.59 |
| mvrhel2 | maybe I should write my own cmm | 16:57.00 |
| chrisl | mvrhel2: you must have a couple of hours spare, then..... ;-) | 16:57.49 |
| ray_laptop | the entire business of preferring (unknown version) shared libraries seems at odds with having a quality supported product | 16:58.02 |
| mvrhel2 | :) | 16:58.04 |
| tkamppeter | mvrhel2, if you write your own cmm, it has no upstream outside Ghostscript, then we can package it with Ghostscript without any problems. Then your cmm in GS is the original upstream. | 16:58.35 |
| chrisl | So we could just rename our lcms, and hope no-one notices...... | 16:59.21 |
| mvrhel2 | hehe | 16:59.55 |
| ray_laptop | chrisl: GPL allows people to fork software -- we can change the name of our version | 16:59.56 |
| | gscms | 17:00.14 |
| tkamppeter | henrys, Ubuntu would not accept third-party-made packages as core part of the distro. | 17:00.35 |
| ray_laptop | and we can strip out all of the stuff we don't need | 17:00.40 |
| mvrhel2 | yes | 17:00.44 |
| tkamppeter | chrisl, you have my message? | 17:00.47 |
| henrys | lcms is not GPL but yes | 17:00.48 |
| chrisl | tkamppeter: yes, not typing fast enough to reply..... | 17:01.07 |
| Robin_Watts | I suspect the issue for the Ubuntu guys is that if updates appear (security fixes etc) then they want to apply them in one place, and everything gets the benefits. | 17:01.10 |
| tkamppeter | but lcms is at least free software? | 17:01.10 |
| mvrhel2 | yes. it has an MIT license which is rather liberal | 17:01.36 |
| | if I recall the Mozilla folks used lcms for the basis for their cmm | 17:01.58 |
| tkamppeter | Robin_Watts, that's it. Security updates are probably still accepted for lcms1 by LCMS upstream. | 17:02.02 |
| henrys | tkamppeter:we are happy to write a letter than, let chrisl know the recipients. | 17:02.05 |
| Robin_Watts | But in this case, lcms1 is not being updated any more. Effectively we have our own fork. We are therefore the "upstream" for it. So I don't see the problem with Ubuntu accepting our 'fork' as being ok. | 17:02.13 |
| tkamppeter | henrys, chrisl got the list and confirmed. | 17:02.31 |
| henrys | mvrhel2:what is the status of informing marti or our changes, are you keeping up with that? | 17:02.44 |
| | maybe by the next release this won't be an issue. | 17:03.01 |
| mvrhel2 | I need to get the most recent changes to him. | 17:03.01 |
| ray_laptop | Robin_Watts: that sort of makes sense. Ubuntu would have to accept a shared lib built from our fork as an update in that case | 17:03.13 |
| mvrhel2 | There is a color conference in November that I am thinking about going to. He may be there | 17:03.28 |
| Robin_Watts | ray_laptop: I'm not proposing that we become the maintainers of the lcms used in general by Ubuntu. | 17:03.51 |
| ray_laptop | Robin_Watts: you mean lcms1 or lcms2 as well ? | 17:04.12 |
| tkamppeter | chrisl, henrys, tell that the built-in lcms1 of GS has once a systematic checking of memory allocation, to eliminate segfaults and second, that the colors of the shared lcms are not correct, defeating our inclusion of color management in Ubuntu. | 17:04.34 |
| ray_laptop | the upstream for lcms1 bugs wouldn't be too bad | 17:04.36 |
| henrys | I can understand the distros position - suppose there is an important security fix for lcms and marti fixes it... | 17:04.42 |
| Robin_Watts | I mean that we should be allowed to have our own internal, non-shared, lcms1 fork as part of our build. | 17:04.43 |
| chrisl | tkamppeter: Okay, I will try to get a mail out before I finish today. | 17:05.26 |
| ray_laptop | chrisl: thanks (also to ken) for the fast response on cust 532's Tr bug | 17:06.41 |
| chrisl | ray_laptop: no problem - and thanks for explaining to Len how broken the file was! Nice one to get fixed for us as well. | 17:07.23 |
| ray_laptop | that file is sure severely flawed ! | 17:07.29 |
| chrisl | That's an understatement, IMHO! | 17:07.43 |
| tkamppeter | For the integration in the distro my plan is to not touch the lcms package which is part of the distro. There are too many dependencies on it. What I would do is to switch GS from using the shared LCMS to using the built-in lcms1. | 17:07.44 |
| Robin_Watts | tkamppeter: That would seem an ideal solution to us. | 17:08.13 |
| ray_laptop | isn't that what chrisl 's commit did ? | 17:08.53 |
| chrisl | ray_laptop: yes, but Ubuntu is getting, essentially, 9.04, so still had the option of using the shared lib | 17:09.32 |
| henrys | tor8:hows the iOS stuff coming? I have miles' ipad now so I can try stuff out if you need. | 17:19.33 |
| tor8 | henrys: I got the cross compilation working with the makefiles, so I don't have to maintain the stupid Xcode project file for all the libraries at least | 17:21.31 |
| henrys | sigh the pcl changes to do device gray for text have cascaded into a huge change | 17:26.00 |
| | if anybody knows of a good tool to find included but unneeded include files let me know. | 17:35.49 |
| chrisl | Okay, calling it a day there - enjoy your weekends, everyone! | 17:41.15 |
| henrys | bye chrisl you too! | 17:41.31 |
| tkamppeter | chrisl, can you resend the mail | 17:41.38 |
| | chrisl, you mis-typed one address. | 17:41.48 |
| chrisl | tkamppeter: just going to now - sorry about that | 17:41.50 |
| tkamppeter | chrisl, thanks. | 17:45.59 |
| chrisl | tkamppeter: done - was weird, I deleted the full stop, and it kept coming back! | 17:46.10 |
| Robin_Watts | Aargh. | 17:56.39 |
| | (c<<3) + 8 != c << 3 + 8 | 17:56.52 |
| sebras | Robin_Watts: time to read up on the precedence table? ;) | 17:57.34 |
| Robin_Watts | sebras: No, annoyingly, I am really well aware of that one. | 17:58.06 |
| | I just missed it. | 17:58.10 |
| sebras | Robin_Watts: I realize, maybe it's more a matter of having your fingers automagically know the table than being aware of it. | 17:59.48 |
| Robin_Watts | Was buried in a macro, so too many parens already. | 18:00.09 |
| | Forward 1 day (to 2011/10/01)>>> | |
IRC Logs