| | <<<Back 1 day (to 2012/03/29) | 2012/03/30 |
| henrys | mvrhel:I'll be watching my mail but will be away from IRC. | 01:43.40 |
| mvrhel | henrys: no worries. I think I figured out the issue. | 05:28.28 |
| | I will be out tomorrow morning | 05:29.06 |
| ManDay | Can you read this http://www.opengl.org/documentation/implementations/intro.pdf ? | 09:15.02 |
| kens | Yes. | 09:15.31 |
| ManDay | thanks | 09:18.18 |
| sebras | tor8 robin: why dont we include a way to synch out the correct versions of the thirdparty source from the respective gits in our mupdf.git? | 09:21.04 |
| kens | chrisl ping | 09:47.03 |
| chrisl | kens: pong | 09:47.16 |
| kens | Do you mind if I change the parameters to .definettcidfont and .loadttcidfotn in gs_ttf.ps ? | 09:47.45 |
| | I want to pass in the existing CIDSystemInfo rather than reaplcing it with Adobe Japan1 | 09:48.07 |
| | WHich is what happens now | 09:48.14 |
| sebras | tor8: did you read the logs? | 09:48.31 |
| chrisl | kens: fine by me - looking at what's there now, it looks to me like it's just working around that info not being available previously - so it *should* be changed. | 09:49.57 |
| kens | chrisl, I think the information was always there, but somebody never bothered to pass it on. | 09:50.21 |
| | The functions are only called form the PDF interpreter | 09:50.33 |
| | Alex has done half the work I need, in that the CMap now contains the original CIDSystemInfo, I just need the font to have ti too :-) | 09:51.01 |
| chrisl | kens: actually, one thought: could you simply add the CIDSystemInfo to the dictionary that's already used as the parameter? I don't suppose it makes *much* difference, though.... | 09:51.20 |
| kens | I *think* this fixes my PDF/A bug report where teh CMap and font CIDSsystemInfo don't match | 09:51.21 |
| | chrisl I probably could, but it kept failing, and the PDF interpreter insists on running it in a stopped context (even with PDFSTOPONERROR) so it was too hard to debug.... | 09:52.05 |
| chrisl | Fair enough, it just seems a pointless complication to pass two dictionaries with the sole purpose of one being put into the other | 09:53.00 |
| kens | yeah I know, but like I said I couldn't figure out what it was complaining about.... | 09:53.27 |
| | This way worked trivially easily ;-) | 09:53.38 |
| | I've started a cluster push, but I don't expect any problems. | 09:54.13 |
| chrisl | I suppose it makes *some* sense to all the CIDFont dictionary "stuff" in the one place. | 09:54.39 |
| kens | :-) | 09:54.48 |
| | As long as you don't mind.... | 09:54.54 |
| | I'll be altering the PDF itnerpreter slightly too of course. | 09:55.14 |
| kammerer | hi all, Does fz_context cache limit work with mucbz? | 09:55.23 |
| chrisl | Nope - as long as you update the comments - I always end up deciphering this stuff when I have to look at anyway :-( | 09:55.58 |
| | kens: ^^ | 09:56.05 |
| kens | I added the new parameters to the comments describing the input. | 09:56.15 |
| tor8 | sebras: because not all third party repositories use git or have official git mirrors ;) | 10:00.14 |
| | kammerer: it should, but there may be bugs. we did receive a bug report about it recently. | 10:00.38 |
| kammerer | tor8: thx, will wait for fix) | 10:03.02 |
| sebras | tor8: I meant the thirdparty git repos on casper. | 10:03.03 |
| | tor8: or are they only to be used for gs? | 10:03.10 |
| | tor8: I thinking that it make more sense to be able to locate the old versions than trying to find old .zip-files in a download directory... | 10:03.51 |
| tor8 | sebras: the plan was to use them for mupdf with submodules | 10:03.55 |
| | but I still don't know which is best, to track the upstream git/svn/cvs repos or just make a thirdparty gits with tarball snapshots | 10:04.41 |
| sebras | tor8: right, that makes a lot of sense. | 10:04.42 |
| | tor8: tarball snapshots (with potentially cherry-picked upstream patches) sounds good enough. | 10:05.09 |
| siyb | any MuPDF developers present | 11:01.52 |
| | ? | 11:01.53 |
| sebras | siyb: yes, several. if you just stay online for a few minutes... | 11:17.36 |
| robin_watts_mac | mvrhel: (For the logs) Yes, all bets about efficiency are off when rops are involved :) | 12:22.11 |
| | paulgardiner: Are you using chatzilla? That should show the topic at the top. That includes a link to the irclogs. | 12:22.56 |
| | So (if like me) you have a crap memory, you can hunt for stuff in there. | 12:23.13 |
| paulgardiner | robin_watts_mac: Yeah ta. Found that. I did look this morning. Anything specific I missed, or just a useful tip? | 12:27.38 |
| robin_watts_mac | I was reading logs. | 12:28.01 |
| | You were asking about tor8's comments about function naming. | 12:28.11 |
| paulgardiner | Ah. I could have looked back further into the logs. Yes of course. Thanks | 12:28.42 |
| | On thing I can't get sorted with Chatzilla is how to get a specific sound to play when someone types my name. I've set it up seemingly correctly, but it doesn't work. | 12:30.13 |
| kens | Personally I use Miranda, which plays a snippet of Penfold when mu name is mentioned ;-) | 12:31.06 |
| | "Crumbs chief, what are we going to do now ?" | 12:31.24 |
| paulgardiner | :-) Like it. | 12:32.39 |
| robin_watts_mac | paulgardiner: Oh, urm... that just worked for me | 12:35.37 |
| paulgardiner | What without any setup? | 12:36.11 |
| | I have paulgardiner and paul as "stalk words" | 12:36.34 |
| robin_watts_mac | Preferences -> Global Settings -> Lists -> Stalk words ? | 12:36.40 |
| paulgardiner | Yep | 12:36.48 |
| robin_watts_mac | And there should be an 'aggressive notify' option somewhere. | 12:37.05 |
| paulgardiner | Under Global I have a file url for Channel: Stalk match | 12:37.20 |
| | I also have the aggressive notify setting checked. | 12:37.45 |
| robin_watts_mac | Right. | 12:38.06 |
| paulgardiner | file://C:/Program Files blah blah etc | 12:38.15 |
| robin_watts_mac | That should be enough for it to beep whenenever someone says paulgardiner | 12:38.22 |
| | Channel: Stalk match ? | 12:38.55 |
| paulgardiner | Yep | 12:39.02 |
| robin_watts_mac | Mine is set to "beep" | 12:39.06 |
| paulgardiner | I've just set mine back to beep. So let's see paulgardiner | 12:39.50 |
| | Can you call me. Doesn't work when I say it. | 12:40.15 |
| robin_watts_mac | paulgardiner: Did that work? | 12:40.20 |
| | First off, is the line shown as 'important' in chatzilla ? | 12:40.35 |
| paulgardiner | Yes that worked. My file url must be malformed | 12:40.43 |
| robin_watts_mac | Windows 7, right? So anything in Program Files is suspect :) | 12:41.13 |
| paulgardiner | Can you try again now? | 12:41.42 |
| | Got it I think : s/ /%20/g | 12:47.52 |
| robin_watts_mac | paulgardiner: beep? | 13:04.19 |
| paulgardiner | Have 3 beeps and then the url setup. Only heard the beeps. Ah well, three beeps should be enough to attract my attention. | 13:05.41 |
| robin_watts_mac | I'm off. Night all. | 13:26.00 |
| henrys | down to 4 customer bugs | 15:05.53 |
| kens | Oooooh | 15:06.00 |
| sebras | have a nice weekend artifexians! | 15:43.56 |
| ray_laptop | looks like I fixed the TextAlphaBits>1 + transparency + band_boundary issue just in time -- a new case just came in from Gemma (RBR1105565.pdf) | 15:55.06 |
| henrys | ray_laptop:it's obvious you are under less strain from 532 - nice set of fixes recently! | 15:56.16 |
| ray_laptop | henrys: I probably need to call them to find out what's going on. We never got a confirmation of the work-around from chrisl for Len's crash, and I haven't heard if they've tracked down the mysterious 2-3 second sporadic delays | 15:57.37 |
| henrys | good to be down to 4 bugs but now we are so close I wish we could get them done. | 15:58.31 |
| kens | won't happen, mine is well-nigh impossible | 15:59.05 |
| henrys | kens:what's the schedule on yours like if you actually sat down and did nothing else? | 15:59.56 |
| ray_laptop | kens: guess we just have to classify it as an enhancement ;-) | 16:00.31 |
| henrys | well if it is impossible, it's wontfix, by definition. | 16:02.36 |
| kens | henrys, months | 16:02.47 |
| | It needs implementation of pdfwrite with the clist to generate the 'transparentcy flattened' image in bands | 16:03.16 |
| | I don't even know where to start | 16:03.26 |
| henrys | then it is an enhancement or project it shouldn't be a bug. | 16:03.38 |
| kens | Well, I didn't classify it that way :-) | 16:04.06 |
| | It happens if you convert a transparent PDF where very large portions of the page are trnasparent (or a very large page, or very high resolution) and have set the COmpatibilityLevel to 1.3 (or idsabled transparency). | 16:04.51 |
| | SO we need to convert to an image. | 16:05.06 |
| | We do this in memory, if there is insufficient memory we generate a VM error, but it exhibits as a problem in discardtransparency group | 16:05.30 |
| | Arguably its a limitaion of transparency flattening. | 16:05.47 |
| | Now that PDF/A-2 is here, it should be less of a problem, because we don't need to flatten transparency in PDF/A-2 but we do for PSDF/A-1 | 16:06.10 |
| henrys | you can assign it to michael then he'd have all 4 bugs. | 16:06.16 |
| | ;-) | 16:06.21 |
| kens | Mean :-) | 16:06.27 |
| henrys | off to read the bug carefully, I looked at a while back but don't remember everything. | 16:07.17 |
| kens | I hope I remembered it correctly | 16:07.57 |
| henrys | your comment 8 is odd, I wouldn't expect the memory error to only happen with -dPDFDEBUG | 16:10.10 |
| kens | Let me look... | 16:10.19 |
| henrys | they're unrelated asfaik | 16:10.26 |
| kens | The memory error happens anyway | 16:11.10 |
| henrys | I didn't mean to interrupt you if you are in the middle of something just looking at marcos bug report. | 16:11.25 |
| kens | Basically we don't have enough memory to make a bitmap big enough to flatten the transparency | 16:11.28 |
| | Not a problem, I can easily be interrupted. | 16:11.39 |
| | The issue exhibits with or without -DPDFDEBUG as far as I know, I may have used the wrong switch in that comment, I think that shold actually be one of the -Z switches | 16:12.28 |
| henrys | and it is pswrite which is obsolete I guess ps2write doesn't matter. | 16:12.47 |
| kens | ps2write is equally affected. | 16:13.02 |
| | No transparency in PostScript | 16:13.07 |
| ray_laptop | kens: can we render it using an internal clist (a'la large pattern bitmaps) ? | 16:14.14 |
| kens | The comment #8 really refers to the fact that with 'some' debug switch (I don't think it can have been -dDPFDEBUG) I get a message form the memory manager about a failed allocation. Note the size of the requested allocation ;-) | 16:14.22 |
| | ray_laptop : This is beyond my competence ;-) | 16:14.34 |
| | We need 'some kind' of clist certainly. | 16:14.43 |
| henrys | of course, so it works on raster devices because we have the clist pdf-14 special compository gizmo and the high level devices can't make use of that. | 16:14.47 |
| kens | henrys yes I blieve this is the point | 16:15.00 |
| henrys | interesting. | 16:15.12 |
| ray_laptop | kens: well, maybe I should take it. We can then "playback" the clist in bands or lines in order to put the image into the target device (that doesn't support clist) | 16:15.34 |
| kens | pdfwrite at low PDF levels, or ps2write, require to render transparency to an image | 16:15.34 |
| malc_ | tor8: is it common for ttf fonts to have god awful kern for ea pair? | 16:15.38 |
| kens | ray_laptop : If you have some idea how to proceed, that's fine by me. | 16:15.56 |
| henrys | kens:does it work on a 64 bit machine? | 16:16.06 |
| kens | If you turn the transparency into an image I'll be happy. | 16:16.06 |
| | henrys it will work anywhere if yo uhave sufficient memory available | 16:16.18 |
| | WHich may require a 64-bit machine of course. | 16:16.28 |
| henrys | ha not a bug ;-) | 16:16.29 |
| ray_laptop | kens: If you want, assign it to me and I'll do the clist thingy. | 16:16.48 |
| kens | I'm open to redefining it as an enhanecemnet | 16:16.53 |
| | ray_laptop : I will do so, thank you. | 16:17.01 |
| | If henrys lieks I will also reclassify it as an enhancement | 16:17.15 |
| | I think I argued that way before, but Igor demurred | 16:17.31 |
| henrys | ray_laptop:I still think it is an enhancement but whatever you think. | 16:17.37 |
| kens | Like I said, I think I argued for it as an enhancement before. | 16:17.52 |
| | So I will make it one, ray can change it bakc if he'd rather it was a bug :-) | 16:18.09 |
| henrys | great now to find mvrhel | 16:18.26 |
| ray_laptop | kens: it shouldn't be all that hard since it is very similar to pattern-transparency-clist (without the complication of needing to step and repeat to tile the area) | 16:19.51 |
| kens | OK I have to run, but I'll stay logged in and will be back later. | 16:19.52 |
| ray_laptop | bye, kens | 16:19.59 |
| kens | ray_laptop : I'll remind you of that later :-) | 16:20.04 |
| henrys | marcosw:meeting today? | 16:31.41 |
| marcosw | yes, we should meet. | 16:32.21 |
| | do you know what the starting bug number should be? | 16:32.29 |
| henrys | no I really can't even remember the last time we had a meeting. | 16:33.08 |
| | 3-16 report starts with 692910 | 16:34.20 |
| marcosw | I think we met two weeks ago, but it might have been three... | 16:34.29 |
| henrys | I don't remember doing 910 but I haven't looked back farther. | 16:36.11 |
| | while you do that I have someone at the door - brb start without me if you like. | 16:37.28 |
| marcosw | okay, will start with 910 | 16:37.36 |
| henrys | baffling what does it mean? | 16:40.45 |
| marcosw | are you asking about 910? | 16:41.50 |
| henrys | yes | 16:41.56 |
| | I am sure cat.ps has fallen into my forgotten cabinet. | 16:43.08 |
| marcosw | i gather we are swallowing quotes in the cat.ps file. since it's a regression I figured I'd bisect and the it would be clear. Probably shouldn't have been assigned to ken. | 16:43.18 |
| henrys | so back to you for a bisect. | 16:43.36 |
| marcosw | already done. In fact I've already looked at bugs through 947 :-) | 16:43.53 |
| henrys | always good to have clear directions how to reproduce something also, this bug seems to lack that. | 16:43.58 |
| marcosw | most are mu bugs and assigned to Tor, I figure he or Robin can reassign. | 16:44.21 |
| henrys | 947 is actually p2 | 16:44.24 |
| | and tor should tend to it soon. | 16:44.46 |
| | tor8:you around? | 16:44.57 |
| marcosw | why is 947 P2? | 16:45.00 |
| henrys | just someone we want to please. | 16:45.13 |
| | I can talk to you about it offline. | 16:45.32 |
| marcosw | that's okay, I'm happy to make it P2 (and just did). | 16:45.57 |
| henrys | so we should put chrisl_away on the 949 cc: still unclear how to handle a bug in both interpreters - closing and all | 16:47.39 |
| marcosw | I've added chrisl_away, but it should be a simple matter of incorporation the 950 change into mu. | 16:50.04 |
| | I've closed 953 as wontfix. | 16:50.22 |
| ray_laptop | wow -- this has been around since 2008 (bug 689805) | 16:50.48 |
| marcosw | I've reassigned 944 to Robin since he said he'd get back to the user when he's back from vacation. | 16:51.34 |
| ray_laptop | marcosw: BTW, I closed a couple of bugs that relate to garbled glyphs. The RBR1105565.pdf from Gemma looks like another instance. If you can check their open bugs against current HEAD some may now be cleared up. | 16:52.25 |
| henrys | are you okay hanging onto 954 marcosw? | 16:52.31 |
| marcosw | yes. | 16:52.41 |
| ray_laptop | as soon as mvrhel wraps up the planar spot color changes, it'll clear a few more for cust 393 | 16:52.59 |
| marcosw | ray_laptop: I can check the open bugs from Gemma. | 16:53.37 |
| | I think there are only a few. | 16:53.46 |
| | I'm looking at 958, which appears to be a pdfwrite issue and so should go to kens. | 16:54.07 |
| henrys | yes | 16:54.40 |
| | that's it? | 16:55.06 |
| ray_laptop | marcosw: kens will hate you for that one -- "astonishingly complicated" as he says | 16:55.12 |
| henrys | can't wait for michael to see he has the only 3 customer bugs. | 16:55.47 |
| | one support bug from arthur | 16:56.57 |
| marcosw | henrys: yup, and I'm off to the airport. I'll check email after I check in and go to the lounge but won't be around after 1:30. I don't expect to need help with support, so other than being on european time everything will be as usual. | 16:57.24 |
| henrys | it should go to chrisl_away | 16:57.30 |
| marcosw | henrys: are you talking bout 958? | 16:57.47 |
| henrys | 942 | 16:57.54 |
| | I searched for support bugs | 16:58.07 |
| marcosw | not to tor8? He maintains the mu website, as far as I know. | 16:58.42 |
| | bugzilla just gave me the message: You are not allowed to change the bug status from UNCONFIRMED to UNCONFIRMED. | 16:59.06 |
| | I can't seem to modify bug 942, keep getting the same nonsensical error. | 17:00.11 |
| henrys | oh okay | 17:00.13 |
| marcosw | i.e. can't move it to Product: MuPDF. | 17:00.28 |
| henrys | weird | 17:00.33 |
| marcosw | can you try? | 17:00.43 |
| henrys | yes | 17:00.52 |
| | works fine for me. | 17:01.12 |
| | oops I gave it to chrisl_away but it will find its way to tor8 and chrisl_away should know about it. | 17:01.37 |
| ray_laptop | marcosw: I'm looking for other bugs that mention TextAlphaBits -- bug 692759 will probably be fixed now. Also bug 692671. | 17:01.45 |
| | marcosw: do you want me to check those, or can you do it ? | 17:03.02 |
| tor8 | marcosw, chrisl_away: the link mentioned in bug 92942 is from chris's web page. we host the mupdf downloads primarily on the google code project page, so it could redirect there. | 17:04.08 |
| ray_laptop | grabbing coffee -- bbiab | 17:04.45 |
| marcosw | tor8: okay, I'll assign it to Chris to look into. | 17:06.06 |
| henrys | tor8:692947 should be reviewed and integrated. | 17:06.51 |
| marcosw | someone beat me to it :-) | 17:07.02 |
| henrys | i assigned it to chrisl_away | 17:07.22 |
| marcosw | ray_laptop: I'll check the TextAlphaBits related bugs later today. | 17:07.38 |
| henrys | tor8:for business reasons, not technical. | 17:07.57 |
| marcosw | I'm offline for a couple of hours. be back around 12:30. | 17:08.26 |
| tor8 | henrys: yeah, I discussed it briefly with Robin. the main problem is who will maintain such patches. | 17:08.50 |
| henrys | I am asking for special treatment for that particular person, not integrate everybody port... | 17:10.36 |
| tor8 | we still need someone to maintain that special port. phonon/qnx is not an environment any of us know much about. | 17:13.29 |
| | and it'd look bad to take it on and then let it bitrot | 17:13.52 |
| henrys | can we ask him if he'll be keeping up with it and if so we'll (you'll) integrate it? | 17:15.44 |
| | you could also just take it and call it bitrot after some period of time. | 17:16.54 |
| | something we really should consider for ghostscript | 17:17.07 |
| | mvrhel:now you've got the only 3 customer bugs ;-) | 17:17.44 |
| mvrhel | hi henrys: I just sent marcos an email on where I am iwth 692618 | 17:18.07 |
| henrys | mvrhel:if you think others on the staff can help please ask, you have gotten stuck with some really hard ones. | 17:22.14 |
| mvrhel | henrys: when I get back from the airport I will look over things. leaving right now. back in a while | 17:36.51 |
| henrys | okay mvrhel | 17:42.05 |
| | bbiaw | 18:07.09 |
| kens | I'm off for the night, I think the ps2epsi.bat problem needs to come back to me, and I already briefly looked at the ridiculous Cairo PDF file, but forgot to re-assign it to me, sorry for the oversight. | 18:37.13 |
| mvrhel | ok. now back to fun with planar patterns | 20:23.37 |
| | Forward 1 day (to 2012/03/31)>>> | |
IRC Logs