| | <<<Back 1 day (to 2014/06/09) | 2014/06/10 |
| mstefanko | with all that know how i'm suprised you guys haven't created a mupdf conversion api saas service.. :) | 05:15.48 |
| pedro_pc | thinks my IRC client went a bit nuts | 08:20.32 |
| | sorry for the connection closed spam :( | 08:21.26 |
| kens | I only saw one disconnect/reconnect | 08:21.49 |
| pedro_pc | hmm - maybe just connection attempts at this side then. Nothing in the logs anyway. I came back to see connection attempts every minute for an hour ;) | 08:24.24 |
| kens | I've been here for > 90 minutes and I only see one disconnect and reconnect | 08:24.47 |
| chrisl | pedro_pc: don't worry, if it did get noisy, we'd tell you! | 08:25.18 |
| kens | ghostbot doesn't log disconnects and reconnects though | 08:25.21 |
| pedro_pc | chrisl: ta :) | 08:26.00 |
| chrisl | Or just ban ya! ;-) | 08:26.14 |
| pedro_pc | sounds like a plan | 08:28.33 |
| henrys | oh it is meeting time | 14:29.30 |
| kens | Yep :-) | 14:29.39 |
| henrys | letâs start with some mupdf stuff | 14:31.14 |
| | I donât see any hope miles is going to let Robin_Watts or paulgardiner get back to mupdf ui stuff, does anyone know somebody who would do it on a consulting basis, zeniko sebras ? | 14:32.25 |
| | tor8 ^^^ | 14:32.42 |
| kens | ....tumbleweeds..... | 14:33.59 |
| henrys | kens:this business of scaling the entire coordinate system by 10 in pdfwrite (by default), I wonder if that should always be done - I can see if you fall back to rendering it is needed. | 14:34.22 |
| mvrhel_laptop_ | morning. | 14:34.31 |
| kens | henrys it would be 'difficult' to get rid of now. I wish we didn't do it though. | 14:34.55 |
| Robin_Watts | mvrhel_laptop_, henrys: I just had Martin Portmann skype me about ETS. | 14:35.03 |
| henrys | Robin_Watts: refer him to support | 14:35.21 |
| Robin_Watts | I've told him I'll ring him back in 30 mins after the meeting to see what he wants. | 14:35.28 |
| | but I will try to move it to email asap. | 14:35.39 |
| kens | assume she wants what he wanted last time, someone to come anbd tell him how to use it | 14:35.57 |
| mvrhel_laptop | I thought we answered everything | 14:36.24 |
| Robin_Watts | mvrhel_laptop: me too. | 14:36.34 |
| kens | He wanted someone to come onsite and do training | 14:36.35 |
| jogux | henrys: what sort of ui stuff? emobix may well have someone ;-) | 14:36.36 |
| henrys | Robin_Watts: tell him you are on high priority work and to send his request to support. | 14:36.44 |
| mvrhel_laptop | I think you guys are the closest | 14:36.50 |
| Robin_Watts | henrys: ok. | 14:36.54 |
| mvrhel_laptop | ;) | 14:37.11 |
| henrys | Robin_Watts: you are working on that high priority SOT stuff right? ;-) | 14:37.20 |
| Robin_Watts | henrys: Today, yes :) | 14:37.36 |
| henrys | jogux: bug fixing and enhancements - do you have the agenda link? | 14:38.10 |
| Robin_Watts | I need to talk to mvrhel_laptop about the crash he was seeing in mupdf. | 14:38.15 |
| jogux | henrys : I do, just a sec. | 14:38.22 |
| mvrhel_laptop | Robin_Watts: did you replicate the issue? | 14:38.30 |
| | or we can talk after the meeting | 14:38.44 |
| Robin_Watts | mvrhel_laptop: I could not replicate, but lets talk after the meeting. | 14:39.16 |
| mvrhel_laptop | ok | 14:39.22 |
| henrys | jogux: under mupdf - âlots of enhancements and bug fixes â¦â I guess the agenda doesnât have a good description | 14:39.24 |
| jogux | henrys : yeah, I was hoping for slightly more :) | 14:39.39 |
| | if Robin/Paul know I can talk to them directly. | 14:40.05 |
| tor8 | henrys: only person I can think of who might be available would be glenn | 14:40.08 |
| henrys | jogux I can associate some bug #âs with that after the meeting. | 14:40.26 |
| jogux | henrys: ok | 14:40.33 |
| paulgardiner_lap | jogux: MuPDF bugs that are assigned to me would be a good start. | 14:41.05 |
| tor8 | jogux: search bugzilla for mupdf component and OS as iOS and android. | 14:41.11 |
| henrys | chrisl: outlines for devices? howâs that. Itâll really help mvrhel_laptopâs xps printing. | 14:41.20 |
| tor8 | most of those are assigned to paul or robin, but there's a dozen unassigned | 14:41.32 |
| chrisl | henrys: I've had a brief look - it should be pretty straight forward. I want to finish up my customer bug first | 14:42.03 |
| mvrhel_laptop | chrisl: that will be great | 14:42.32 |
| henrys | tor8: do you know if sebras would be interested in doing some of this stuff? | 14:42.39 |
| tor8 | henrys: interested, yes. have time to? doubtful. | 14:42.52 |
| henrys | tor8: I hate to let that stuff fester | 14:43.34 |
| jogux | henrys : well anything iOS or Android specific sounds right up our street (do you guys say that?) | 14:43.57 |
| tor8 | henrys: yeah. | 14:44.10 |
| henrys | we say alley | 14:44.10 |
| jogux | tor8 / paul : thanks :) | 14:44.18 |
| tor8 | jogux: a lot of stuff currently assigned to paul is forms work, which is more core mupdf library and less in risk of festering | 14:44.43 |
| henrys | tor8: maybe you should take that I thought we agreed to that at the meeting. | 14:45.19 |
| tor8 | but the ui stuff in the ios and android apps is more outwardly visible when it festers, and I'm busy enough with the core library to get really involved in gritty UI framework details | 14:45.33 |
| | henrys: yeah. I think we agreed that I'd take over the core forms work. | 14:45.57 |
| jogux | well android/ios are our speciality, mupdf core I know very little about just now :-) | 14:46.12 |
| Robin_Watts | Matt @ emobix would be well placed to do android/ios UI stuff, I would have thought. (but I don't know how the emobix rates stack up for this stuff - that would be a business decision) | 14:46.18 |
| henrys | Iâll talk to miles about using emobix resources. | 14:46.31 |
| tor8 | although I'd ideally see the android app restarted as a pure Java application that uses the JNI stuff that Robin has been working on in his spare time | 14:46.58 |
| pedro_pc | feels the same about SO... | 14:47.15 |
| henrys | pedro_pc: you mean doing native stuff⦠we had a discussion along those lines at the meeting | 14:48.17 |
| mvrhel_laptop | yes | 14:48.34 |
| paulgardiner_lap | tor8: restarting wouldn't necessarily make sense as very little of it is coloured by going through JNI | 14:48.35 |
| pedro_pc | the build is pretty convoluted at the moment, and a pain to debug | 14:48.42 |
| jogux | there are some nice helper utils to generate the jni layers these days (albeit most of them are targetting C++ rather than pure C) | 14:48.58 |
| pedro_pc | I've been looking at that a bit doing the Good stuff | 14:48.58 |
| Robin_Watts | pedro_pc: MuPDF is at heart a C library, with a published API. | 14:49.09 |
| | but for the android app, rather than expose that API in all its generality, we just exposed the potted sections we needed. | 14:49.33 |
| tor8 | paulgardiner_lap: I'm thinking in terms of providing a reflection of a subset of the MuPDF C API in Java | 14:49.40 |
| Robin_Watts | We should really have exposed the entire API to java and then called that. | 14:49.54 |
| pedro_pc | nods - more like the SOL stuff - I think we could get a cleaner build doing the apps as std android studio builds using SOL | 14:49.59 |
| Robin_Watts | But that was more work, and we wanted something that works. | 14:50.05 |
| tor8 | and then build UI components on top of that, and then build the app on those. and then all android stackoverflow questions that aren't about how to compile we could answer with: use the API instead of hack mupdf.c | 14:50.20 |
| Robin_Watts | s/works/worked quickly/ | 14:50.20 |
| henrys | mvrhel_laptop: it does seem you should support kens first and we can get that dreaded project done. | 14:50.48 |
| mvrhel_laptop | henrys: yes. I agree. I will see if I can get that wrapped up this week | 14:51.10 |
| henrys | mvrhel_laptop: a clean slate would be nice before diving into SOT | 14:51.28 |
| mvrhel_laptop | yes. | 14:51.36 |
| henrys | kens: with the conversions you would be just about finished right? | 14:52.11 |
| kens | henrys I discovered some new areas of unpleasantness today working on 16 bit images | 14:52.40 |
| henrys | ugh | 14:52.54 |
| kens | But the code is already far better than the old code, so I'm in no great hurry | 14:52.57 |
| paulgardiner_lap | tor8: yeah I like that idea too, but you'd still want an Android View class that displays pages and an AdapterView for controlling the positioning of those on screen. I'm not sure you'd write those much differently because of having the whole API in Java. You'd be able to clean them up, but I doubt if it would magically simplify what is fairly awkward stuff | 14:53.55 |
| mvrhel_laptop | This is a good time for me to get my part done though as gsview is winding down significantly | 14:54.18 |
| | before I get into SOT | 14:54.36 |
| henrys | not everyone is here but enough for a majority - does anyone object if there vacation is posted to an Artifex calendar (google calendar shared)? | 14:54.45 |
| tor8 | paulgardiner_lap: yeah, I don't expect the actual GUI code we have to change much but it would open up for customers to be able to do their own much more easily | 14:55.02 |
| mvrhel_laptop | That would be a good idea | 14:55.02 |
| Robin_Watts | henrys: Sounds good to me. | 14:55.12 |
| henrys | I ask because I know some have commented on google and privacy. | 14:55.26 |
| kens | NP for me | 14:55.29 |
| tor8 | paulgardiner_lap: and we could do a much simpler demo example | 14:55.34 |
| Robin_Watts | Presumably shared between artifex employees rather than globally public. | 14:55.37 |
| henrys | artifex and google employees ;-) right. | 14:56.02 |
| Robin_Watts | Google knowing when I go on my holidays doesn't bother me too much :) | 14:56.20 |
| jogux | robin: Iâd bet they already know anyway ;) | 14:56.45 |
| Robin_Watts | I think they'll be able to guess *where* I'm going from helens search logs... | 14:57.10 |
| paulgardiner_lap | henrys: On the subject of the SO library, we have a potential customer who has reported problems. jogux had a look to see if any are easy fixes, but it seems not. | 14:57.17 |
| henrys | Robin_Watts: next trip to africa youâll start getting web banners for binoculars | 14:57.20 |
| Robin_Watts | nah. expensive camera lenses :) | 14:57.47 |
| paulgardiner_lap | henrys: I'm not sure how fast they are going to move and whether they are likely to start to pressure us before Sept. | 14:57.55 |
| henrys | paulgardiner_lap: I was going to start SOT in 3 minutes | 14:58.10 |
| paulgardiner_lap | Oh right. Okay | 14:58.19 |
| henrys | any other mupdf or gs stuff? | 14:58.24 |
| | if not, adjourned! | 14:58.48 |
| Robin_Watts | henrys: mvrhel has a threading problem with mupdf that is showing up under gsview. | 14:59.02 |
| kens | coffees | 14:59.11 |
| jogux | paul: I donât think Iâd quite reached that conclusion for all of them yet. at least two I need to check into the code a bit more. | 14:59.16 |
| Robin_Watts | That's the only thing I know of that I'd like to fix before we do a 1.5 release - and even that isn't crucial. | 14:59.25 |
| henrys | Robin_Watts: I did forget to ask about the release, that looks pretty important to get underway. | 14:59.59 |
| | tor8 ^^^ | 15:00.22 |
| paulgardiner_lap | jogux: okay good | 15:00.50 |
| henrys | Robin_Watts: Iâd get the release out now gsview doesnât have to go out with an integral release does it mvrhel_laptop ? | 15:01.47 |
| mvrhel_laptop | henrys: no not at all | 15:02.03 |
| Robin_Watts | henrys: Fair enough. No one else has reported the problem :) | 15:02.17 |
| henrys | pedro_pc: wanted to ask about android for âGoodâ? Do you think we are close? | 15:04.25 |
| | paulgardiner_lap, pedro_pc : I wonder if Richard speaks like he writes, with the missing words and all. Very strange. | 15:05.08 |
| pedro_pc | I think so, but I was saying that at the end of last week too - chasing down a couple of issues with the file transfer call, then I think its primarily testing | 15:05.09 |
| | henrys> yeah, he suddenly became very terse | 15:05.26 |
| henrys | pedro_pc, paulgardiner_lap : also Miles is going after other MDMâs do you guys have any experience with others? Is it likely to be similar changes? | 15:06.44 |
| tor8 | Robin_Watts: henrys: alright, I take it that is a go for wrapping up the 1.5 release now then? | 15:07.16 |
| pedro_pc | henrys> debugging the SOT android/native stuff is a bit tedious for Good due to the 'interesting' way we build it in SOT - basically end up using Log.d() calls most of the time which is less than useful. I spent some time trying to get an eclipse or android studio project together for it, but that's a backburner thing. Probably useful to do for future work though | 15:07.24 |
| Robin_Watts | tor8: I think so. | 15:07.26 |
| tor8 | Robin_Watts: there are two commits on tor/master in preparation for the release | 15:07.30 |
| Robin_Watts | tor8: looking now. | 15:07.41 |
| pedro_pc | henrys> there are a few supported in SO at present, all with SecureFS variants and some custom authentication code/UI - most seem broadly similar | 15:08.45 |
| Robin_Watts | tor8: lgtm | 15:09.34 |
| henrys | pedro_pc: thatâs good I was dreading some entirely new scheme weâd have to support | 15:09.43 |
| Robin_Watts | tor8: possibly the platform/android/res/values*/strings.xml should be updated too. | 15:10.03 |
| jogux | henrys : SOL / securefs is really setup so that the MDM vendor could do the integration. Itâs an âSDKâ. | 15:10.24 |
| tor8 | Robin_Watts: ah, yes. | 15:10.53 |
| henrys | jogux: you must know very different MDM vendors than I do. I donât know how these clowns get to work. | 15:11.10 |
| paulgardiner_lap | henrys: :-) | 15:11.28 |
| jogux | henrys : okay, bad phrasing on my part ;-) | 15:11.28 |
| Robin_Watts | henrys: in the shortbus. | 15:11.31 |
| tor8 | Robin_Watts: updated version on tor/master now | 15:12.36 |
| paulgardiner_lap | Richard probably doesn't really work for GD. He just turns up there by accident every day. And Brad has yet to realise this. | 15:12.40 |
| henrys | jogux: wonder about when everything will transfer to marcosw1 place. | 15:12.51 |
| Robin_Watts | tor8: great. | 15:13.06 |
| henrys | jogux: Iâve been holding off gettting an account until then but I suppose that is silly, it will all just transfer. | 15:13.21 |
| jogux | henrys : me too. I need to finish mac / windows hosted build, and there are a couple of bugs, then I want to setup a linux build VM from scratch. Itâs on the back burner a little at the moment as it is working for us. | 15:14.00 |
| | henrys : yeah, accounts should be easy to transfer accross | 15:14.45 |
| henrys | paulgardiner_lap: I think the mobile industry is like it was during the dot.com bubble with web programming, where they just hire anyone. | 15:15.17 |
| | pedro_pc, jogux : so if you have any questions about the agenda let me know, we have the accompanying discussion at the meeting which help us glue the mess of items into something coherent. I think reading it without that context can be difficult | 15:18.10 |
| jogux | Iâve heard getting good iOS/Android people in the US at the moment is extremely tricky. | 15:18.42 |
| Robin_Watts | yeah, everyone who has written html5 now counts themselves as an app programmer. | 15:19.31 |
| | apps are the new flash. | 15:19.53 |
| henrys | anyway thatâs all I have for now. Other SOT stuff, anything I should pass onto Miles? | 15:20.51 |
| pedro_pc | don't think so | 15:21.37 |
| Robin_Watts | henrys: nope. | 15:21.49 |
| jogux | canât think of anything else. | 15:21.52 |
| mvrhel_laptop | Robin_Watts: so you are not able to replicate the issue | 15:33.56 |
| | grabbing some breakfast bbiab | 15:38.54 |
| jogux | henrys: oh, I vaguely forgot something. We probably want to think about support for iOS 8 cloud drive in SOT, and possibly also SOL. (as well as iOS 8 SDK support in general) | 15:39.44 |
| henrys | jogux:Iâll add it to the agenda list | 15:40.58 |
| jogux | ios 8 probably ships mid-september, so we have a while. | 15:41.40 |
| | miles should probably enable SO2 for family sharing in itunes connect too (apple will have emailed him) | 15:42.11 |
| henrys | jogux: Iâll pass that on. | 15:42.41 |
| jogux | thanks | 15:42.52 |
| | ios cloud drive thing applies to mupdf too really | 15:43.32 |
| henrys | jogux: can you easily create a reproducible bug and enter it? | 15:46.17 |
| | bbiab | 15:57.04 |
| kens | goodnight all | 16:21.47 |
| Robin_Watts | mvrhel_laptop: OK, back now. Yell when you're back from breakfast. | 16:22.53 |
| jogux | henrys: itâs not so much a bug as âyou canât access files in the ios8 icloud drive as weâve not implemented itâ :) but Iâll log it into bugzilla. | 16:25.25 |
| Robin_Watts | jogux: An "enhancement" bug :) | 16:27.31 |
| mvrhel_laptop | Robin_Watts: back | 16:38.03 |
| Robin_Watts | mvrhel_laptop: OK. I tweaked the multi-threaded.c example for windows, and I cannot make it go wrong with that file. | 16:38.42 |
| mvrhel_laptop | Robin_Watts: ok . so is there anything I can do on my end | 16:39.19 |
| | actually let me see if the android app has issues with it too | 16:39.31 |
| Robin_Watts | mvrhel_laptop: Can I rebuild gsview here and run it? | 16:39.44 |
| | I have Windows 7 and VS2010 | 16:39.49 |
| mvrhel_laptop | if you have visual studio 2013 | 16:39.56 |
| Robin_Watts | I do not. | 16:40.06 |
| mvrhel_laptop | hold on | 16:40.15 |
| jogux | robin / henrys : opened as http://bugs.ghostscript.com/show_bug.cgi?id=695303 | 16:41.15 |
| Robin_Watts | Otherwise, I could connect into your laptop using the google chrome thing? But that occupies your machine... | 16:41.29 |
| henrys | jogux: thanks | 16:41.51 |
| jogux | oh, should I open a mupdf one too? | 16:43.05 |
| Robin_Watts | jogux: Probably. | 16:43.29 |
| mvrhel_laptop | Robin_Watts: ok. so that android app also has no problems with it. | 16:43.42 |
| | Robin_Watts: let me poke a bit at it here first and see if I can figure out what the issue might be | 16:44.04 |
| Robin_Watts | mvrhel_laptop: If you have 2 threads accessing the same image, it would be interesting to know whether they are both on the same page or not. | 16:44.38 |
| mvrhel_laptop | ok | 16:44.45 |
| Robin_Watts | i.e. is the problem with an image that's shared across 2 different pages, or is it to do with a single page being redrawn at two different resolutions, and the image being evicted during that time. | 16:45.21 |
| | mvrhel_laptop: I suspect that it requires the store to be set to a particular size for this to occur. | 16:45.46 |
| jogux | ok - http://bugs.ghostscript.com/show_bug.cgi?id=695304 for mupdf / icloud drive. | 16:45.47 |
| mvrhel_laptop | Robin_Watts: I found the problem. It was not in mupdf | 17:20.42 |
| | I had a bug in the viewer code which was on occasion rendering the same page on different threads due to how I was testing if annotation display was on or off | 17:21.23 |
| | essentially the first time the page was encountered it was unknown if it had annotations and I was letting that be a case to have the page rendered. with the render page +/- 1 range there were cases where the first encounter had not finished and it thought annotations are still unknown so go ahead and render this page again. | 17:23.16 |
| | sorry that I had you waste you time looking at the file. :( | 17:23.48 |
| | so, essentially rendering the display list is thread safe with out locking as long as you don't do the same page which seems reasonable | 17:25.22 |
| | glad to have this thing fixed | 17:26.28 |
| Robin_Watts | mvrhel_laptop: Ah, cool. | 17:30.47 |
| mvrhel_laptop | I hope you did not spend too much time on this Robin_Watts | 17:31.04 |
| Robin_Watts | So the problem was that you were rendering the same display list on 2 different threads without any locking? | 17:31.24 |
| mvrhel_laptop | yes | 17:31.29 |
| Robin_Watts | Hmm. I'd vaguely have hoped that wpuld be safe. | 17:31.42 |
| mvrhel_laptop | and the crash was in the tile stuff | 17:31.42 |
| | try page 10 of the doc I sent you | 17:31.54 |
| Robin_Watts | it was page 10 that was failing? | 17:32.07 |
| mvrhel_laptop | that might be zero based page 10 | 17:32.12 |
| | so page 11 | 17:32.23 |
| Robin_Watts | mvrhel_laptop: OK. I'll have a look later. Want to find this bloody WMF bug. | 17:32.33 |
| | thanks. | 17:32.38 |
| mvrhel_laptop | Robin_Watts: no. problem do you want me to open an enhancment and cut down the file? | 17:32.55 |
| Robin_Watts | mvrhel_laptop: No, leave it for now. | 17:33.22 |
| mvrhel_laptop | ok | 17:33.25 |
| rayjj | hi, all | 18:15.39 |
| | sorry I missed the meeting. My mother in law had a doctor's appt this AM that my wife had to get her ready for, so I had to take the kids to school, then I had my eye doctor appt. | 18:19.48 |
| Robin_Watts | rayjj: I don't think you missed much. | 18:28.47 |
| | Who wrote the original ETS SSE code? | 18:28.57 |
| henrys | Robin_Watts: I thought raph did | 18:31.35 |
| Robin_Watts | ah, ok. | 18:31.57 |
| rayjj | Robin_Watts: I also think Raph wrote the original SSE code for ETS | 18:45.59 |
| | ISTR that the ETS code and the C code diverged and gave different results | 18:46.40 |
| | even back when Raph was doing it | 18:46.53 |
| henrys | for the logs kens you and I have support june 10 - 20. | 19:00.44 |
| mvrhel_laptop | grabbing some lunch now... | 19:17.28 |
| | out for a bit. | 20:12.24 |
| henrys | jogux: miles has agreed to 30 hours a week for mupdf ios and android, current smart office resources should not be used. | 20:44.38 |
| | current smart office resources == you and pete | 20:45.04 |
| | paulgardiner, Robin_Watts are you good with that? | 20:45.35 |
| jogux | damn, you saw through my clever plan to escape SO? ;-) | 20:45.41 |
| | thanks Henry - Iâll try figure out where to fit that in; is there any particular âhighest priorityâ? There seems to be one P2 bug and everything else is P4 | 20:48.36 |
| | [Iâm using http://bugs.ghostscript.com/buglist.cgi?columnlist=product%2Ccomponent%2Cassigned_to%2Cbug_status%2Cresolution%2Cshort_desc%2Cchangeddate%2Cpriority&list_id=15649&op_sys=iOS%204&op_sys=iOS%205&op_sys=iOS%206&op_sys=iOS%207&op_sys=Android&order=priority%2Cbug_severity&product=MuPDF&query_based_on=&query_format=advanced&resolution=--- as a search ] | 20:49.30 |
| henrys | jogux: start with any crash. looks like several of these are not really bugs at all, like 693719 | 21:04.40 |
| | jogux: anything that isnât a UI problem should probably have the Hardware field changed to âallâ just to keep your list simple | 21:06.52 |
| | Hardware and OS | 21:07.30 |
| jogux | yeah. I guess anything that can be reproduced on a desktop build can go through onto someone elseâs list :) | 21:07.39 |
| henrys | yes probably tor will field those. | 21:08.11 |
| marcosw_ | henrys: are you looking at the VI2816_2px.ps blank pdfwrite issue? | 22:08.50 |
| henrys | marcosw_: no kens agreed to look at it when he came in. I sent mail to support and ken so heâll follow up when he gets in. | 22:12.59 |
| marcosw_ | okay, I ran a bisect and found output pdf is blank starting with c561232cf26e060b89fc4f3bd4bf5c679731d4db | 22:37.08 |
| Robin_Watts | henrys: re: emobix on mupdf. Sounds good. | 23:11.20 |
| henrys | marcosw_: oh cool kens will see that in the logs, itâs his commit | 23:11.36 |
| | Robin_Watts: okay good | 23:11.45 |
| paulgardiner | henrys: yes emobix on MuPDF sounds good to me. | 23:52.03 |
| | Forward 1 day (to 2014/06/11)>>> | |
IRC Logs