| | <<<Back 1 day (to 2014/07/08) | 2014/07/09 |
| rayjj | how much do I hate DNS/BIND... Let me count the ways... | 01:24.26 |
| jogux | morning all. looks like http://smartoffice-forum.artifex.com is working now. | 09:12.47 |
| | ah, I see Robin has discovered that too :) | 09:13.08 |
| Robin_Watts | jogux: Yeah. I've posted 3 things, and am now out of ideas... | 09:15.08 |
| jogux | hehe | 09:15.25 |
| | a â2.3 releaseâ thread would work too I guess? | 09:15.35 |
| | already out for iOS, coming soon for Android, etc. | 09:15.44 |
| Robin_Watts | jogux: Oh, good plan. With a list of changes/fixes? | 09:16.44 |
| jogux | nods | 09:16.53 |
| mattchz | âThis topic is now closed. New replies are no longer allowed.â | 09:17.26 |
| | do we already have special accounts? or should I just create one? | 09:18.46 |
| mattchz | just creates one. | 09:19.44 |
| jogux | mattchz: yeah, thatâs the right thing to do | 09:19.57 |
| mattchz | Hmm, email comes from ron@rlicht.com. Probably want to change that. | 09:20.12 |
| jogux | haha :) | 09:20.54 |
| | Iâll make a list to email to Ron. but where /should/ they come from? sos@ ? | 09:21.30 |
| mattchz | sos maybe creates the wrong connotations :) | 09:22.05 |
| | (whatâs the s for anyway) | 09:22.10 |
| Robin_Watts | Smart Office Support | 09:25.36 |
| | I don't think we should give that address publicly though. | 09:26.39 |
| mattchz | smartoffice-forum@artifex.com ? | 09:27.31 |
| jogux | mattchz: seems reasonable to me. and I guess that work forward to sos? | 09:28.26 |
| | s/work/would/ | 09:28.29 |
| mattchz | or, mrcubee@picsel.com | 09:28.35 |
| Robin_Watts | smartoffice-forum@artifex.com seems smart. | 09:31.57 |
| jogux | is it marcosw for things like that? | 09:33.30 |
| Robin_Watts | yeah. | 09:35.09 |
| mattchz | back to banging my head against android crashlytics. | 09:37.55 |
| | (or rather eclipse) | 09:38.12 |
| jogux | okay, email send about email address. | 09:50.34 |
| | http://smartoffice-forum.artifex.com/t/welcome-to-artifex-software-forums/6 needs a quick edit to remove the version number I think. | 10:02.57 |
| Robin_Watts | jogux: Are you an admin? | 10:08.31 |
| | You are... | 10:09.05 |
| paulgardiner | is confused by crashlytics. Seems to be showing #6 as the most common which is not the shutdown but I was expecting. | 10:13.35 |
| | s/but/bug/ | 10:13.57 |
| jogux | robin_watts: yep | 10:14.38 |
| paulgardiner | Ah. Someone closed the big one. Did they fix it? | 10:14.43 |
| jogux | I thought I reopened it | 10:14.49 |
| | I had closed it because I thought weâd fixed it :( | 10:15.03 |
| jogux | has really reopened it now | 10:16.03 |
| paulgardiner | No problem. Just had problems finding it. | 10:16.18 |
| | Strangely, doesn't look like the one you pointed me at the other day though | 10:17.12 |
| jogux | Iâll take a look at #6 | 10:17.13 |
| | paulgardiner: odd. | 10:17.24 |
| paulgardiner | I remember commenting on it involving Dashboard_isVisible | 10:18.01 |
| | which this one doesn't | 10:18.10 |
| jogux | I remember you commenting like that | 10:19.19 |
| paulgardiner | On the other hand it wouldn't surprise me if there were a number of ways that querySaveViewStateOnShutDn can end up going down to the IControl level | 10:19.51 |
| jogux | I canât find one involving the dashboard in a similar crash. odd. | 10:21.23 |
| paulgardiner | And going back a few days in irc logs, clicking on the link you posted, I see #2. | 10:22.45 |
| | Not sure where I got Dashboard_isVisible from | 10:23.11 |
| | Still I'm looking at the right one now. :-) | 10:23.35 |
| mattchz | re: crashlytics. Iâm thinking itâs probably ok to commit our API key to the public repo. | 10:28.27 |
| | just not our API secret. | 10:28.52 |
| jogux | mattchz: we wouldnât want random builds other people do to report into our crashlytics | 10:31.40 |
| mattchz | I donât think they would end up doing that, as presumably the build couldnât be uploaded withbout our secret, and thus wouldnât have a valid build ID. | 10:32.10 |
| jogux | hrm. possibly. though we might get lots of warnings about missing dsyms :) | 10:32.32 |
| mattchz | Yeah, Iâm not sure. | 10:32.41 |
| | where do those warnings go? | 10:32.49 |
| jogux | they popped up in the web interface for me. but then seemed to go away again :) | 10:33.04 |
| mattchz | On android, it seems like a file is created like this: | 10:34.29 |
| | #This file is automatically generated by Crashlytics to uniquely | 10:34.42 |
| | #identify individual builds of your Android application. | 10:34.43 |
| | # | 10:34.44 |
| | #Do NOT modify, delete, or commit to source control! | 10:34.46 |
| | # | 10:34.47 |
| | #Wed Jul 09 11:23:51 BST 2014 | 10:34.48 |
| | version_name=@string/version | 10:34.49 |
| | package_name=com.artifex.mupdfdemo | 10:34.50 |
| | build_id=<redacted> | 10:34.51 |
| | version_code=50 | 10:34.56 |
| | app_name=MuPDF | 10:34.56 |
| | It looks like the build_id is autogenerated by the build system, and presumably then used by the Crashyltics library to try and upload. | 10:35.36 |
| | thereâs another file crashlytics.properties which contains: | 10:40.06 |
| | #Contains API Secret used to validate your application. Commit to internal source control; avoid making secret public. | 10:40.08 |
| paulgardiner | jogux: there's a test in here if (UE2FvExplorer isActive() || Dashboad_isVisible()), so I guess both need fixing. Back to the three choices I had before. 1) Make preShutdownFn run under the scheduler (which it's prelude says it should). 2) Try to cache all the state that querySaveViewStateOnShutDn relies on. 3) Take out more of the tests and have the view state potentially saved unnecessarily (and | 10:43.55 |
| | maybe wrongly - not sure) | 10:43.57 |
| | Any intuition on this? | 10:44.30 |
| jogux | paulgardiner: I think we decided 3) last time? | 10:50.11 |
| paulgardiner | Yes, which should work provided this whole thing is just an optimisation. | 10:52.48 |
| sebras | tor8: god morgon. | 11:39.49 |
| tor8 | hi sebras | 11:42.15 |
| mattchz | did anyone find out if that Chris guy had a contributor agreement? | 12:49.55 |
| Robin_Watts | Chris? Not Michael Cadilhac? | 12:50.19 |
| mattchz | oh, erm, yeah, him. | 12:50.35 |
| Robin_Watts | mattchz: We need to ask henrys. | 12:50.46 |
| mattchz | yep. henrys was going to ask Miles I think. | 12:50.57 |
| Robin_Watts | mattchz: image commit for review on robin/master if you wouldn't mind. | 12:51.15 |
| mattchz | sure, np. | 12:51.42 |
| | git is so slow in sot | 12:52.15 |
| | I guess itâs a much bigger repo than Iâve used for other projects :) | 12:52.25 |
| Robin_Watts | mattchz: It's a big repo :( | 12:52.29 |
| mattchz | indeed | 12:52.46 |
| Robin_Watts | Took several attempts to convert. | 12:52.53 |
| mattchz | I used git at picsel, I donât remember it being too bad. But I guess Iâve blanked all that out now | 12:53.16 |
| jogux | I think ghostscript.com isnât the fastest git server, particularly from the wrong continent | 13:03.56 |
| mattchz | this was local stuff | 13:04.42 |
| jogux | oh. odd. Iâll fit your ssd one day, honest :) | 13:06.48 |
| Robin_Watts | And another couple of commits. | 13:09.00 |
| | This afternoon I will look at adding jbig2 support to smart office. | 13:09.13 |
| | Seems something that should be easy to do, and we can do some nice before/after shots for Miles. | 13:09.36 |
| | Gawd. Windows update day. | 13:11.18 |
| mattchz | robin_watts/paul_gardiner/tor8: do you guys tend to build via ant debug for android? | 14:54.03 |
| | (or the equivalent in Eclipse). | 14:54.11 |
| | Iâm thinking we should only enable Crashlytics for release builds. | 14:54.18 |
| tor8 | mattchz: I only build with "ant debug". I never use Eclipse. | 14:55.56 |
| | but then, I rarely build the android version at all, so best to ignore me | 14:56.09 |
| mattchz | I presume googleplay builds are done using âant release'? | 14:56.22 |
| paulgardiner | I only build via Eclipse. I never use ant debug. | 14:56.25 |
| Robin_Watts | mattchz: Yes, I build using ant debug. | 14:57.34 |
| | yes, googleplay builds are done using ant release. | 14:57.57 |
| | specifically using the instructions here: http://twiki.ghostscript.com/do/view/MuPDF/AndroidReleases | 14:58.30 |
| mattchz | having played around with Crashlytics, I think we certainly want to remove all of the API keys from the public version, despite what I saying earlier to jogux. | 14:58.47 |
| | I donât think the Android version of Crashlytics actually uploads anything to do the servers unless youâve used proguard. | 14:59.06 |
| Robin_Watts | tor8: We have a request for the apks to be put on the mupdf site. | 14:59.15 |
| mattchz | so if we didnât remove the API keys from the public version, we would get random bug reports from others builds. | 14:59.31 |
| | thanks guys | 14:59.39 |
| Robin_Watts | tor8: Can you sort that out? http://ghostscript.com/~robin has the latest apks (MuPDF-5{0,1,2,3}.apk) | 15:00.24 |
| tor8 | Robin_Watts: okay. you can just scp them to /var/www/mupdf.com/downloads/archive on casper and symlink the latest release to /var/www/mupdf.com/downloads | 15:00.25 |
| | Robin_Watts: hm, what are the 0,1,2,3 variants? | 15:00.44 |
| Robin_Watts | tor8: 0 = armeabi, 1 = armeabi-v7a, 2 = x86, 3 = mips | 15:01.01 |
| tor8 | Robin_Watts: I'll copy them over | 15:01.02 |
| | mupdf-1.5-andoird-(armeabi|armeabi-v7a|x86|mips)? | 15:01.23 |
| Robin_Watts | I figure it might need a bit of html on the main mupdf page too. | 15:01.23 |
| tor8 | Robin_Watts: yeah, I'll sort that out | 15:01.48 |
| Robin_Watts | I'd be tempted to leave them numbered 50,51,52,53 as they match the versions on google play, and otherwise we *will* get confused. | 15:02.02 |
| | but possibly mupdf-1.5-android-50-armeabi.apk or something ? | 15:02.31 |
| tor8 | yeah, I can keep the 50 in there | 15:02.47 |
| | Robin_Watts: http://mupdf.com/downloads/ look reasonable | 15:05.03 |
| Robin_Watts | lovely. | 15:05.55 |
| | We might want to move v8 into archive now? | 15:06.11 |
| mattchz | excellent, thanks tor8 - I can close that bug now :) | 15:06.12 |
| tor8 | "Android viewer APK installer files" a good link title for the main page? | 15:06.15 |
| mattchz | is it worth adding the instructions to that wiki page? | 15:06.30 |
| tor8 | Robin_Watts: sure. | 15:06.34 |
| Robin_Watts | tor8: "Prebuilt Android Viewer downloads"? | 15:07.38 |
| | bah, I don't care. Perhaps we'd like to restrict this to technically competent people :) | 15:08.13 |
| mattchz | :) | 15:08.34 |
| | The only reason I really care is that Iâm trying to figure out how to do this crashlytics stuff. | 15:08.58 |
| tor8 | Robin_Watts: fair enough :) | 15:09.12 |
| Robin_Watts | mattchz: What instructions? | 15:09.19 |
| mattchz | http://twiki.ghostscript.com/do/view/MuPDF/AndroidReleases | 15:09.35 |
| Robin_Watts | Right, but add what instructions to that wiki page? | 15:09.55 |
| mattchz | a reminder to upload the builds to the website | 15:10.07 |
| | otherwise weâll probably forget ;) | 15:10.14 |
| Robin_Watts | Oh! Yes, bargain. | 15:10.17 |
| | Go for it. | 15:10.25 |
| | You probably need to register for the twiki. Use MattHolgate | 15:10.44 |
| | and then I'll make you staff so you can see the GhostDocs web. | 15:10.59 |
| mattchz | Iâm just looking at how to do a release with Crashlytics. I think if we restrict it to just ant release builds and leave it out of all eclipse builds, we should be fine. | 15:11.15 |
| | ok, will do, ta. | 15:11.23 |
| Robin_Watts | rebooty. brb. | 15:11.25 |
| | gawd. this takes ages when you have VMs to shut down etc. | 15:13.08 |
| mattchz | aww | 15:13.24 |
| | oh, iâve registered btw. | 15:13.28 |
| henrys | chrisl: wow froomeâs out | 16:09.29 |
| chrisl | henrys: must have been a pretty nasty fall :-( | 16:10.36 |
| Robin_Watts | Did I mention that my niece started day 2 of the Tour De France? | 16:11.10 |
| | (waved flag, not wore lycra etc) | 16:11.25 |
| henrys | Robin_Watts: did she finish the stage ;-) | 16:11.30 |
| Robin_Watts | She was wearing a yellow jumper :) | 16:11.45 |
| chrisl | Robin_Watts: I saw your request of evidence on fb..... | 16:11.54 |
| henrys | Robin_Watts: did you find a pic? | 16:12.09 |
| Robin_Watts | chrisl: evidence has turned up. | 16:12.12 |
| | I will try and repost it. | 16:12.18 |
| chrisl | Ah, cool! | 16:12.19 |
| Robin_Watts | Just in case anyone hasn't been paying attention: smartoffice-forum.artifex.com should be live now. Any ideas, suggestions etc, gratefully received. | 16:24.19 |
| mattchz | having been hacking around with Crashlytics for a while | 18:03.46 |
| | on Android, it seems to be very fiddly to make Crashlytics an optional part of the build process. | 18:04.03 |
| | Iâm wondering about an alternative solution | 18:04.25 |
| | 1) Change the android project into a library project | 18:04.37 |
| | 2) Create a new project which is the opensource version of MuPDF and simply depends on the library project. This is the one weâd build normally. | 18:05.00 |
| | 3) Create a third project which is the private version of MuPDF, and also depends on the library project but also contains the Crashlytics JAR and initialisation. | 18:05.28 |
| | This would result in the a bit more complexity, but would probably the cleanest/easiest way of doing it. | 18:06.10 |
| | Is this something weâd like to try, and is it worth the effort? | 18:06.17 |
| Robin_Watts | mattchz: I would say, that unless we get to use crashlytics that can track the NDK, it's not worth doing. | 18:08.34 |
| mattchz | ok, fair enough. Iâll contact Crashlytics to ask them, and otherwise park this | 18:09.12 |
| Robin_Watts | even if we can get access to the NDK, I'm not sure that a major restructuring is a great idea, unless it can literally be a wrapper. | 18:10.37 |
| mattchz | nods. | 18:11.16 |
| | We did something similar on another project, to support different variants of builds. | 18:11.35 |
| | I presume we still want to go with crashlytics on ios though? | 18:13.02 |
| | robin_watts: btw, I assume you were happy with Micahelâs patch? As henrys has said that it is ok to commit? | 18:15.20 |
| | ok, home time all. nn. | 18:18.03 |
| Robin_Watts | mattchz: (For the logs). I am happy with that patch, yes. | 18:21.00 |
| | and yes, we should go for crashlytics on ios. | 18:21.22 |
| | cos you have that working already, yes? | 18:21.36 |
| cryptopsy | thanks 4 make mupdf able to open images | 20:14.36 |
| | now i can remove that fucking shit 'feh' image viewer | 20:14.46 |
| | that's all for today, g2g harass some bug lists | 20:20.31 |
| henrys | mvrhel_laptop: do you want some help with the slides? Iâm sure we can reel in a few helpers | 21:22.18 |
| mvrhel_laptop | henrys: Let me see what I can throw together. If I am going to do the call, I would prefer to put the slides together myself. It is mainly going to be an overview of what has been added over the past year | 21:23.57 |
| | It should not take too long. Having fun now though figuring out edr stuff... | 21:24.28 |
| henrys | mvrhel_laptop: I would like to pitch any ps2write fixes we have made for the community. we should ask kens what would be good to add. | 21:26.05 |
| mvrhel_laptop | yes. definitiely | 21:26.18 |
| henrys | mvrhel_laptop: and of course all the high level color work | 21:26.40 |
| mvrhel_laptop | henrys; have tried to include mupdf in the slides also | 21:27.14 |
| | I try and pitch that for print preview each year as an option over poppler | 21:27.33 |
| henrys | mvrhel_laptop: oh right yes a slide on the new gsview | 21:27.57 |
| mvrhel_laptop | oh yes that too | 21:28.08 |
| henrys | if we ever have resource to port it to a platform they care about | 21:28.38 |
| | Forward 1 day (to 2014/07/10)>>> | |
IRC Logs