| <<<Back 1 day (to 2014/05/12) | 2014/05/13 |
mvrhel_laptop | chrisl: if you have 64 bit windows, you could try to install gsview.. | 06:25.51 |
| I just sent an email, which should fix the issues that I was having earlier | 06:26.16 |
| good night | 06:27.27 |
| al | 06:27.29 |
| all | 06:27.31 |
ray_laptop | morning, all | 14:02.46 |
| kens: on bug 695216, re: comparefiles/Bug693711.pdf, Since this "fails" (with ERROR -15 closing pdfwrite device.), I suggest that this should not be part of our test suite. | 14:04.28 |
kens | i DON'T GET THAT RESULT | 14:05.06 |
ray_laptop | kens: since it's in there, then we attempt to render the pdfwrite output, which is what led me to open the 216 bug | 14:05.08 |
kens | #cpa sorry | 14:05.11 |
ray_laptop | kens: I just tried it again. A HEAD build on 32-bit Windows | 14:05.33 |
kens | that's what I used | 14:06.07 |
ray_laptop | kens: I assumed that you were SHOUTING because it was noisy in your house ;-) That or you were making sure that I heard you way over hear in CA | 14:06.31 |
| kens: that is strange. Let me fire it up with the debugger and see where the -15 is coming from... | 14:07.20 |
| hi, marcos | 14:09.35 |
| marcos: since the comparfiles/Bug693711.pdf has badly broken fonts, and pdfwrite produces a broken pdf, should it really be part of our test suite ? | 14:12.46 |
| kens: BTW, I just tried on peeved and get the -15 error. | 14:13.04 |
kens | I'll switcdfh back to master and retest. For me the pdfwrtite output is OK | 14:13.07 |
ray_laptop | I wonder if you have some different fonts or something | 14:13.24 |
kens | Doubtful... | 14:13.35 |
| It certainly did not previously cause a -15 error, and the resulting PDF< while missing fonts, worked OK In GS | 14:13.54 |
ray_laptop | my laptop has a LOT of fonts, which is why I thought the problem might have been on my end, but peeved is a vanilla ubuntu system | 14:14.37 |
henrys | gs/mupdf meeting in 15 minutes, SOT meeting in 45 | 14:16.07 |
ray_laptop | oh, joy ;-) | 14:16.21 |
| I can hardly wait ;-) | 14:16.31 |
| kens: are you able to see the -15, or do you want me to look into it | 14:18.54 |
kens | Not yet, but I'm rebuilding from master | 14:19.10 |
| as I was on mooscript before | 14:19.23 |
| so I need a new binary | 14:19.53 |
chrisl | tor8: there's a commit in my mupdf repo that's a punt at a common libjpeg memory management layer | 14:26.30 |
henrys | Robin_Watts: are the hosting problems resolved? | 14:30.40 |
Robin_Watts | yes. | 14:30.45 |
mvrhel_laptop | good morning | 14:30.48 |
henrys | Robin_Watts: I see it is working | 14:30.49 |
| Robin_Watts: do you want to keep maintaining that stuff? | 14:31.15 |
mvrhel_laptop | chrisl: thanks for the feedback. I agree with almost everything you had to say. I plan to add in drag-n-drop. def. need that | 14:31.33 |
Robin_Watts | henrys: For now, sure. If we ever need to investigate other hosting options etc, I'll pass it on. | 14:31.38 |
henrys | Robin_Watts: okay | 14:31.49 |
| so mvrhel_laptop is supposed to switch to working on SOT - can others look at his customer bugs? | 14:32.22 |
mvrhel_laptop | and I do need to speed up the first page display to match the old gsview, but as you go through the doc, esp at a moderate resolution the new version is much much faster | 14:32.43 |
chrisl | mvrhel_laptop: I'm possibly not the best person to give un-biased feedback - using Windows usually means start off angry! | 14:32.48 |
Robin_Watts | henrys: Can I? Please? | 14:33.04 |
mvrhel_laptop | chrisl: and I agree with the buttons needing the 3d feeback | 14:33.05 |
paulgardiner_lap | No me me | 14:33.16 |
henrys | Robin_Watts: I think that would nullify what we are trying to do. | 14:33.47 |
mvrhel_laptop | henrys: let me spend today looking at these | 14:34.05 |
| then at least I can get some comments if they look difficult | 14:34.29 |
henrys | mvrhel_laptop: okay if you can hand them off, I know ray_laptop can do the color index thing easily | 14:34.31 |
mvrhel_laptop | right | 14:34.38 |
| looks like one of them used to be assigned to chrisl.... | 14:35.19 |
| ok I will look at 695221 | 14:35.45 |
| that should be easy | 14:35.48 |
ray_laptop | mvrhel_laptop: I can take any that you want. Just assign them over | 14:35.58 |
chrisl | IIRC that comes off the back of a build change, but it's definitely *not* a build issue! | 14:36.00 |
mvrhel_laptop | ray_laptop: ok . I will likely give you 695074 | 14:36.25 |
kens | ray_laptop : 32-bit debug version, built from current master HEAD, "gswin32 -sDEVICE=pdfwrite -sOutputFile=\temp\out.pdf" works for me and does not exit with -15 | 14:36.26 |
chrisl | mvrhel_laptop: I thought the "flat" buttons were part of the Windows 8 "new world order" | 14:36.39 |
kens | I'm using the original test file, I'll go get the one from tests | 14:36.41 |
mvrhel_laptop | and 695118. I just need to sit down and do it | 14:36.47 |
| chrisl: I dont like it either though | 14:37.01 |
| I like feedback | 14:37.05 |
| i dont understand why kens cant install though | 14:37.25 |
chrisl | But I assume there's not a lot we can do about it..... | 14:37.25 |
mvrhel_laptop | :( | 14:37.27 |
henrys | I won't try to give away Robin_Watts bugs I know he covets them for fear there will be no escape from SOT | 14:37.33 |
kens | Well I cna *install*, it just doesn't work afterwards | 14:37.43 |
mvrhel_laptop | chrisl: we can create what ever we want in xaml. it is pretty easy | 14:37.45 |
pedro__ | there is no escape | 14:37.46 |
mvrhel_laptop | kens: still same error? | 14:37.51 |
kens | Yes, identical | 14:37.56 |
mvrhel_laptop | that makes no sense at all | 14:37.58 |
kens | I'll try on my laptop later,then if its still a problem, you cna diagnose it at the staff meeting :-) | 14:38.16 |
mvrhel_laptop | there are only 3 files in the installer for sure . you must be grabbing the wrong file | 14:38.17 |
chrisl | kens: I can reproduce the ERROR -15 exit with the file from comparefiles | 14:38.32 |
kens | mvrhel_laptop : I pulled the zip file and installed the files in that | 14:38.35 |
| chrisl I@ll get he file as soon as I can stop typing here | 14:38.52 |
henrys | it is kind of discouraging to see folks like Robin_Watts and paulgardiner struggling. smart people banging their heads against the wall is usually evidence something needs to be profoundly reconsidered. | 14:38.55 |
mvrhel_laptop | ok | 14:38.56 |
marcosw | sorry to be late, I'm not at home and the irc port is blocked. | 14:39.12 |
mvrhel_laptop | well I will join in the head banging shortly | 14:39.30 |
chrisl | henrys: I'd have thought going bust was a sign things needed to be reconsidered...... | 14:39.35 |
Robin_Watts | henrys: I'm not even at the stage of actively developing/debugging really. | 14:39.44 |
pedro__ | henrys: there's just a lot of complexity - ATS was maintained by a team of 5 on an ongoing basis | 14:39.47 |
kens | chrisl ray_laptop still works for me, pulled the file from peeves | 14:40.09 |
Robin_Watts | I'm still trying to get existing commits through testing and onto the trunk. | 14:40.17 |
chrisl | Weird..... | 14:40.19 |
ray_laptop | kens: you should be able to get it to fail on peeves (you have a log in there) | 14:40.53 |
Robin_Watts | I live with the optimistic hope that when we get ATS behaving properly and build machines up to date etc it'll all be much simpler... | 14:40.56 |
kens | Its byte for byte identical to the one I already have anyway | 14:40.57 |
| ray_laptop : I *really* don;'t want to debug it on peeves | 14:41.07 |
henrys | so the other thing on my gs list was language switch, how goes it chrisl ? | 14:41.07 |
chrisl | henrys: it's progressing. I'm wrestling with moving the interpreter initialisation out of the graphics lib just now | 14:41.51 |
paulgardiner_lap | We were expecting it to be difficult, but it's actually turned out a little worse than expected, at least to get started. We hadn't accounted for the state we'd be taking it over in, and how our own knowledge of the source is a little specialised. (There were different teams on different aspects so there is much we have never touched before) | 14:42.28 |
henrys | have you guys heard about visual studio now supportng cross platform stuff - cordova I don't know if that is of use to mupdf or SOT | 14:42.29 |
ray_laptop | kens: I get a -15 from pdf_finish_FontDescriptor (called from pdf_finish_resources from line 2401 in pdf_close) | 14:43.13 |
paulgardiner_lap | Both Pete and I are confused at how far from working the G version is when some how a version had been released in the past. | 14:43.22 |
kens | ray_laptop : well I can look at that. | 14:43.29 |
Robin_Watts | henrys: If either SOT or mupdf were written in javascript, maybe :) | 14:43.43 |
ray_laptop | kens: it fails on the first resource (j=0, pres->rname = R264 | 14:44.17 |
chrisl | Robin_Watts: for goodness sake, don't give tor8 any ideas! | 14:44.23 |
pedro__ | there's cerrtainly a lot of opportunity to streamline the SOT builds and UI code | 14:44.24 |
henrys | chrisl: do make that the high priority if you don't get interrupted by a top ten customer. I think not having that really hurts us in the "preview" stage and mvrhel_laptop may pull in a printer customer in Japan | 14:44.35 |
tor8 | chrisl: rust looks like an interesting language ;) | 14:44.43 |
henrys | Robin_Watts: oh I thought it was a "real" thing | 14:45.24 |
kens | ray_laptop : the first call into there has pres->rname=R174 | 14:45.31 |
chrisl | henrys: I can't really make it a higher priority that "top" - but it gets quite soul destroying..... | 14:45.49 |
kens | ray_laptop : it does. however,return a -15 | 14:46.52 |
henrys | chrisl: maybe scrumming each day with ray_laptop and I might help a bit. Sometime just explaining it to somebody helps. Whatever you think | 14:47.07 |
paulgardiner_lap | On the other hand, given the difficulties, progress on ATS has been huge, and we have some headway with the G version, in that we can authenticate now and we aren't having to wait for replies to our support request | 14:47.13 |
chrisl | henrys: there's not much to explain, that's why it's not the most inspirational of tasks - there are *very* few "eureka" moments..... | 14:48.07 |
henrys | the other meeting thing was to please test mvrhel_laptop's gsview as he sends out messages to tech, seems to be really nice work. | 14:48.16 |
paulgardiner_lap | Oh, and given we are supposedly still in the gs/mupdf meeting, I at last got around to submitting the new iOS MuPDF to the app store, and some how it got accepted! | 14:48.17 |
henrys | paulgardiner_lap: wow | 14:48.45 |
| very nice | 14:48.48 |
mvrhel_laptop | nice | 14:48.51 |
kens | ray_laptop : the -15 is overwritten later when we write out fonts. | 14:49.18 |
| ah no, that;s a local redefinition | 14:49.44 |
| Hmm, OK it does return a -15, but its bogus | 14:50.44 |
| and for me the resulting file renders 'correctly' given that the fonts are not embedded. | 14:51.10 |
ray_laptop | kens: is the xref broken in your resulting pdf (it is for me) ? | 14:51.53 |
kens | No, mine seems OK | 14:52.03 |
| ]Buti just overwrite it, so let me look in a minnute | 14:52.13 |
chrisl | ray_laptop: " **** Warning: File has an invalid xref entry: 387. Rebuilding xref table." ? | 14:52.30 |
henrys | Finally (for me) is everyone happy with workflowy? It is time to renew. Let me know if there is some other system you'd like to use. | 14:52.52 |
ray_laptop | chrisl: kens: me too: **** Warning: File has an invalid xref entry: 387. Rebuilding xref table. | 14:52.52 |
chrisl | henrys: workflowy seems okay to me..... | 14:53.21 |
kens | TO be honest, I don;t think this is new behaviour, and given that the input file is broken, I'm not hugely inclined to do anything with it | 14:53.31 |
chrisl | henrys: OTOH, now we're running wikis...... | 14:53.50 |
ray_laptop | I think it's interesting that rust went from 0.1 all the way up to 0.9, then 'reverted' to 0.10 :-) | 14:54.14 |
| maybe they should have started with 0.001 :-) | 14:54.41 |
| *way* back in 2012 | 14:54.59 |
henrys | chrisl: we also all have access to Google Drive and I could easily replace workflow with a git managed text file that would be more flexible for us, but difficult for Miles and Scott | 14:55.38 |
chrisl | henrys: it shouldn't difficult for Miles and Scott, it can still be accessed via a URL | 14:56.31 |
ray_laptop | presumably you'd only need git to make changes | 14:57.23 |
henrys | chrisl: I would certainly prefer that and I'll do it if others are okay with it. | 14:57.26 |
chrisl | ray_laptop: exactly | 14:57.44 |
henrys | ray_laptop: has expressed security concerns with giving data to google, I don't think that much of an issue though | 14:58.13 |
ray_laptop | henrys: that way we could ALL be pushing changes in real time during staff meetings ;-) | 14:58.16 |
chrisl | Google already have *all* our e-mail....... | 14:58.37 |
henrys | ray_laptop: I believe the permission can be set as they are now | 14:58.43 |
| ray_laptop: I think multiple drivers even with source control is going to be confusing, but I don't feel strongly about it | 14:59.33 |
ray_laptop | henrys: if I had concerns over google having our 'workflowy' type of data, I don't recall them, so I withdraw any objections. I *like* the idea of having a git repo text file | 15:00.01 |
| henrys: I was kidding about us all pushing changes. hence, the winky | 15:00.32 |
henrys | marcosw: anything on the quality front for gs or mupdf? | 15:00.47 |
| I need 2 minutes before the SOT meeting be right back | 15:01.28 |
ray_laptop | henrys: used "quality" and gs and mupdf all in the same sentence -- seems strange ;-) | 15:01.42 |
| kens: it isn't causing a problem AFAICT, but in stepping through to see where the -15 came from, I see that line 2359 in pdf_close calls pdf_write_page and ignores any return code (which can be errors) | 15:04.02 |
mvrhel_laptop | brb | 15:04.05 |
henrys | chrisl: most folks pull their email down - I realize it goes through google but clearly there are more security concerns with data that takes up residence at Google. | 15:04.34 |
kens | ray_laptop : yes, that way it finishes the PDF, even if there are errors. Otherwise you get a very badly broken PDF | 15:04.46 |
marcosw | henrys: I've run the gs banded/page mode comparison and there are some differences that I'll open bugs for in the next day or two. | 15:04.56 |
chrisl | henrys: my mail resides on gmail...... | 15:05.02 |
ray_laptop | henrys: even when you pull your mail down, it stays on google | 15:05.03 |
Robin_Watts | even having our email pass through google is bad. | 15:05.17 |
henrys | ray_laptop: if you tell it to it does | 15:05.27 |
Robin_Watts | We should run our own domain and bypass them entirely. | 15:05.37 |
kens | henrys I pull mail off google, and delete from server, and GMail doesn't delete it | 15:05.58 |
Robin_Watts | henrys: Once it's been through google, do you honestly believe we can trust them not to take a copy? | 15:06.02 |
henrys | Robin_Watts: we've discussed this before and the problem was how to admin that domain. | 15:06.28 |
Robin_Watts | That's not my memory of what the problem was :) | 15:06.56 |
| I have a (well, several) email domains. | 15:07.12 |
| Take wss.co.uk for example. | 15:07.24 |
henrys | the issue is I'd have to manage the domain because miles and I share secrets about all of you that you aren't supposed to know ;-) member? | 15:07.44 |
Robin_Watts | henrys: sure. | 15:07.57 |
| wss.co.uk takes precisely 0 hours a week admin time for me. | 15:08.11 |
chrisl | Well, that answers the question: henrys you need to admin the mail system - then there's no conflicts ;-) | 15:08.32 |
Robin_Watts | The big issue, as I recall, was that you wanted to keep the gmail interface. | 15:08.42 |
henrys | Robin_Watts: you're trying to avoid the SOT meeting , I get it ;-) | 15:08.45 |
| okay we'll discuss it after the SOT meeting. | 15:09.11 |
Robin_Watts | Well, the SOT meeting report for me is: still struggling to get stuff reapplied. | 15:09.15 |
| ATS is much better now (thanks to Joseph), and we can actually run tests etc. | 15:09.31 |
| So nothing should get pushed to golden/master without it passing the tests. | 15:09.52 |
| And we can expand the set of 'precommit' tests as we get more and more builds working etc. | 15:10.12 |
| Joseph ran back through all our commits to date and found that most of them are fine and cause no problems. | 15:10.48 |
| We have about 8 that DO cause problems, most of which are reapplications from the branch to the trunk. and some of the problems are bogus, I think. | 15:11.12 |
henrys | paulgardiner_lap, Robin_Watts, pedro__ : Miles was hoping to show before and after results on his upcoming Japan trip, I've said that didn't look like we were at that stage. Is there anything Marcos could put together to show some progress? | 15:11.17 |
Robin_Watts | so I'm trying to get those working. | 15:11.24 |
marcosw | do the tests do bitmap comparisons? If so does someone look at the differences? | 15:11.49 |
Robin_Watts | marcosw: Yes, the tests do bitmap comparisons. | 15:12.05 |
| If there are no changes, no need to look. | 15:12.18 |
| If there are changes you can look at them on a web based interface. | 15:12.30 |
marcosw | and that happens before the push to golden/master? | 15:13.26 |
pedro__ | yes | 15:13.41 |
Robin_Watts | henrys: I don't think my font changes will make enough of a difference to show anything up. | 15:13.52 |
ray_laptop | I think the progress has just been swallowing the SOT mess and finding out what was missing from what we got. | 15:14.04 |
pedro__ | in the original system it would be done via the review system once tests had passed | 15:14.06 |
mvrhel_laptop | it sounds like great progress to me | 15:14.20 |
Robin_Watts | I've got some PDF export fixes that are nice, but aren't going to show up well in a before/after thing as people will be interested in how things look when it's loaded the first time. | 15:14.38 |
ray_laptop | mvrhel_laptop: I agree that is great progress -- just not customer visible | 15:14.52 |
mvrhel_laptop | difficult set up stuff that simply had to get done | 15:14.55 |
pedro__ | henrys> there are a few changes in the agents (word/ppt) but they're mostly individual bug reports - not the big layout changes | 15:14.56 |
Robin_Watts | pedro__ might have had some fixes, but I think he's been swallowed in SOG now. | 15:15.03 |
ray_laptop | SOG ?? | 15:15.37 |
Robin_Watts | Smart Office for Good. | 15:15.59 |
pedro__ | displaying some images which were previously missing, supporting paragraph background color, supporting auto-text color etc | 15:16.02 |
mvrhel_laptop | those would all be nice things to show for before and after | 15:16.23 |
henrys | pedro__: If we have enough for marcos to create a before after picture show that would be great. It doesn't need to be on golden either, just some level of confidence that it can be eventually integrated | 15:17.11 |
paulgardiner_lap | pedro__: presumably those changes would show up rather nicely when someone views the documents that you were using to show the bug | 15:17.21 |
pedro__ | ok, that should be easy enough do | 15:17.30 |
| paulgardiner_lap - yes, probably the best way | 15:17.57 |
henrys | pedro__: have you been getting feedback from Good? | 15:18.03 |
| pedro__: seemed like a "run around" to me. | 15:18.33 |
pedro__ | henrys: they have a bit of a call-center approach to support - the ticket was passed around 3 different people getting nowhere, and Paul actually found the problem on Saturday | 15:18.48 |
| I talked to the original suport guy in the US on Monday | 15:19.06 |
| I'd certainly just do phone calls from now | 15:19.24 |
henrys | pedro__: noted | 15:19.46 |
| marcosw so it sounds like pedro__ is going to give you some stuff for miles' trip do you have time to organize it into something presentable, do you need help with it? | 15:21.09 |
pedro__ | at least now we're able to work through the real integration problems - there seems a lot of work on the secureFS we weren't anticipating | 15:21.20 |
| henrys: when is Miles trip? | 15:22.00 |
henrys | pedro__: oh I was just going to do a marketing piece on the virtues of secureFS for Miles ;-) | 15:22.07 |
mvrhel_laptop | we leave in 1 week | 15:22.14 |
| next week Wed. | 15:22.19 |
paulgardiner_lap | Yes, at least we aren't up against a brick wall now, but we are still very confused as to how secueFS ever worked. | 15:22.29 |
pedro__ | ok | 15:22.30 |
| henrys: excellent timing ;) | 15:23.30 |
henrys | so the other thing was sorting out SO vs SO+, are the feature lists correct on the app store? There seems to be little to distinguish the two release - pdf annotation and some language stuff? | 15:24.30 |
Robin_Watts | paulgardiner_lap: From what I've been hearing on skype, the problems with good/securefs are not problems with the securefs concept itself, just the particular good implementation of it, right? | 15:24.52 |
paulgardiner_lap | yes | 15:25.15 |
pedro__ | Robin_Watts: absolutely - it seems like missing implementation of some underlying file/directory creation | 15:25.27 |
paulgardiner_lap | It's just that we'd assumed that must have been used in the previous releases | 15:25.34 |
Robin_Watts | so henrys can write a marketing piece with a clear conscience. | 15:25.35 |
pedro__ | marketing always has a clear conscience ;) | 15:25.50 |
Robin_Watts | pedro__: marketing has NO conscience. | 15:26.10 |
henrys | paulgardiner_lap: I don't understand if Good is wrong all their "secure" apps are wrong? | 15:26.44 |
paulgardiner_lap | Not that Good is wrong. It's our integration with Good | 15:27.11 |
marcosw | henrys: if I get it in the next day or two I should be okay, time wise. | 15:28.16 |
Robin_Watts | The idea is that securefs is a filing system that is built on a load of alien functions. | 15:28.17 |
paulgardiner_lap | The secureFS concept is fine. AFAIK goods concept is fine. It's the strange bits of sticky tape and bent wires that join them together in the source that is the cause of confusion | 15:28.23 |
Robin_Watts | These alien functions get implemented for each possible provider (so Good would be one, BigTinCan presumably has another etc) | 15:28.53 |
henrys | paulgardiner_lap: anyway this is going to be "Good" soon, right? And customers are going to flock to Good and buy SOT? | 15:29.05 |
paulgardiner_lap | Absolutely. Not a doubt in my mind. Complete certainty | 15:30.11 |
paulgardiner_lap | crossed fingers behind back | 15:30.20 |
henrys | so other than auditing the feature list (see above) that's all I had. | 15:31.09 |
| is jogux back soon? | 15:31.14 |
pedro__ | he's away for roughly 2 weeks - left yesterday | 15:31.37 |
| but he'll be online most evenings | 15:32.13 |
henrys | pedro__: so what about ATS in his absence? | 15:32.16 |
Robin_Watts | he's available to fix problems. | 15:32.56 |
| and we can log in (though as proved yesterday, when things go wrong, we can burn hours hunting for what's wrong) | 15:33.28 |
pedro__ | we have to sort out the remaining ATS failures and we can use it for pre-commit testing before pushing to Golden | 15:33.33 |
| still have other build/test clients to sort out, but the real priority was getting the linux builds & app tests running | 15:34.36 |
henrys | Robin_Watts: I do think google has worked very well for us. I do use the apps and the drive - spam wise hard to beat. I can add it to the agenda and put it to a vote if you want. I"d probably vote to stay with it. | 15:36.02 |
Robin_Watts | henrys: It's been discussed before, and nothing fundamental has changed. I think their T&Cs are appalling, but I was voted down. | 15:36.40 |
| So this is just me grumbling. | 15:37.05 |
| If it was my decision, I'd have ghostscript.com handled by us. Those people that want to use gmail can then forward stuff out to a gmail address. | 15:37.52 |
henrys | Robin_Watts: I thought of you last friday evening, I went to a street fair and had chocolate covered bacon. | 15:39.09 |
Robin_Watts | I have a bacon chocolate bar downstairs. | 15:39.27 |
| Sadly, it's (shockingly) not very nice :( | 15:39.39 |
henrys | Robin_Watts: yea didn't care for it. It sounded right. | 15:39.58 |
paulgardiner_lap | Marmite flavoured chocolate is surprisingly nice (presumably only if you like Marmite). Not sure what the equivalent of Marmite is in the US. | 15:42.15 |
Robin_Watts | marmite chocolate is not nice if you don't like marmite. | 15:42.33 |
| Though oddly, marmite flavour crisps are. | 15:42.54 |
| paulgardiner_lap: Tar. | 15:43.03 |
paulgardiner_lap | :-) | 15:43.49 |
henrys | there must be some enzyme marmitetase some people have that break down that nastiness before they taste it. | 15:45.35 |
paulgardiner_lap | marmitetase. Like it. :-) | 15:46.27 |
kens | chrisl (for the logs) could you take a quick look at bug #695227 please, and see if the code goes through the bbox device when on Linux using the default device (whatever that is, X11 ?) it certainly doesn't on Windows. | 15:50.37 |
| I'm not unhappy with the proposed patch, but I don't really understand why we're even in gdevbbox | 15:51.16 |
| Night all | 16:08.56 |
mvrhel_laptop | ray_laptop: you there? | 16:18.50 |
| chrisl: well it def. was your commit on the 30th that resulted in Bug 695221. However, I need to understand a bit more about USE_LARGE_COLOR_INDEX=0 and the DeviceN devices like tiffsep. According to the comment in msvc.mak USE_LARGE_COLOR_INDEX has to be 1 for this device | 16:27.30 |
| and so it may be that the customer trying to set this to 0 is not going to work... | 16:27.59 |
| oh lovely I get a crash during a reclaim when I set USE_LARGE_COLOR_INDEX=0 | 16:29.21 |
| I may push this one off to ray_laptop too.... | 16:30.30 |
henrys | if I had a nickel for the engineering hours on color index... | 16:31.44 |
mvrhel_laptop | henrys: yes | 16:37.57 |
| I made some comments about what I found and passed this to ray. It may not really be a bug since according to the comment in the code you cant have USE_LARGE_COLOR_INDEX=0 for tiffsep. | 16:38.55 |
| that leaves me with 695118 which is one that I just need to get done. I think I know what I am going to do for this one. | 16:39.45 |
| it will take a solid day | 16:39.59 |
| oh and I need to answer the question to support from the 8th from customer 885 too | 16:42.43 |
Robin_Watts | tor8: Did you see the mupdf customer question this morning. | 16:58.03 |
| It didn't come to support, but I quoted it in it's entirety in my reply. | 16:58.16 |
| They were complaining that 1.4 is slower at loading pages. | 16:58.26 |
| ISTR you said you had ideas about how to fix that ? | 16:58.36 |
tor8 | Robin_Watts: is this related to bug 695183? | 17:54.29 |
Robin_Watts | I don't think so. | 17:55.00 |
| unless that bug happens to be to do with multiple page PDF loading. | 17:55.14 |
| I thought that bug was about rendering being slow, but maybe he was using multiple page PDFs and it's the PDF loading that's slower. | 17:55.48 |
| That might make more sense. | 17:55.51 |
| See my mail to support earlier. | 17:56.07 |
tor8 | the change to dynamic page trees should make initial startup faster, but individual page access slower | 17:56.18 |
| "page-loading speed" is awfully vague; some profiling data would be nice if they could provide it | 17:57.01 |
Robin_Watts | tor8: Do you want to take over communicating with them? I have my hands full with SOT at the moment. | 17:58.22 |
tor8 | Robin_Watts: okay. | 17:59.33 |
Robin_Watts | Ta. | 17:59.39 |
tor8 | Robin_Watts: I love bug reports like 695228... | 18:00.50 |
Robin_Watts | yeah. | 18:08.38 |
mvrhel_laptop | Robin_Watts: you there? | 20:07.30 |
Robin_Watts | mvrhel_laptop: I am. | 20:20.34 |
mvrhel_laptop | oh hi Robin_Watts | 20:20.44 |
| I was wondering if there was a way to turn off AA in mupdf? | 20:21.01 |
Robin_Watts | mvrhel_laptop: There is. | 20:21.09 |
mvrhel_laptop | ok. next question. is it in the docs where I can find it or can you point me to something? | 20:21.37 |
Robin_Watts | fz_set_aa_level | 20:21.47 |
mvrhel_laptop | ah. ok. cool | 20:21.54 |
| thanks | 20:21.56 |
Robin_Watts | (was just looking for it, not being awkward :) ) | 20:21.58 |
| alphabits = 0 turns off aa, except for text. | 20:22.23 |
| text is pretty much always aa'd cos we get a greyscale thing out of freetype. | 20:22.50 |
| Or maybe it's that non-zero means 8 for text regardless. | 20:23.19 |
mvrhel_laptop | ah. ok. chrisl was complaining to me about that AA text in gsview and if I could turn it off | 20:23.19 |
Robin_Watts | Let me do a quick test. | 20:23.42 |
mvrhel_laptop | so are we stuck with AA text? I was thinking of having the option to turn it off | 20:23.43 |
Robin_Watts | OK. 0 does disable AA for text. | 20:24.22 |
mvrhel_laptop | ok cool | 20:24.28 |
| thanks Robin_Watts | 20:25.22 |
Robin_Watts | np. | 20:25.26 |
| Forward 1 day (to 2014/05/14)>>> | |