| <<<Back 1 day (to 2015/12/10) | 20151211 |
jogux | fredross-perry: your mupdf commit looks fine. nice piece of bash scripting :-) | 10:04.55 |
| very minor thing but I'd usually write the if then all on one line using a ';' (ie. if [ "$MAJOR" -ge "&" ]; then ) rather than splitting it over 2 lines | 10:05.43 |
tor8 | fredross-perry: jogux: I'd just have used 'sed' to extract the version string, but bash magic works too I guess :) | 10:06.16 |
jogux | fredross-perry: I think we may want to disable bitcode in the xcode project for our appstore submission. (because it's not required, and if we enable it we have to manually download the final dSYM files from itunesconnect and manually upload them into crashlytics) | 10:07.08 |
| tor8: yeah. I probably might've tried sed too, mainly as I'm not great with the bash syntax. then probably I'd have got annoyed at the bsd sed on OS X and ended up doing it in perl :-) | 10:16.17 |
| fredross-perry: and very minor, but commit message doesn't really need to begin 'mupdf' as this is the mupdf repository :-) | 10:17.42 |
| fredross-perry: (and to be clear, even though I suggest switching bitcode off about, it's still a good idea to include bitcode in the library build, as some of our customers may need bitcode in the library) | 10:18.53 |
tor8 | fredross-perry: the prefix we usually use for the iOS app is just "ios: " | 10:28.24 |
| for commit messages | 10:28.35 |
jogux | tor8: Fred was asking about what to put in the appstore 'what's changed' for the 1.8 release too - I presume there must be some generic statement about 1.8 vs 1.7 that's already been written? | 10:31.58 |
tor8 | jogux: the CHANGES file in the mupdf project, and also the mupdf.com/news web page | 10:37.01 |
jogux | tor8: that'll do it. thanks! | 10:37.51 |
tor8 | Robin_Watts: a handful of fairly large commits on tor/master for review | 12:52.18 |
Robin_Watts | tor8: will look in a bit. Just in the middle of some hairy JNI. | 12:52.41 |
| I'm pulling the MuPDF jni stuff I did up to date. | 12:52.55 |
tor8 | Robin_Watts: a question about the JNI stuff then ... how closely do we match the C api? | 13:39.10 |
| do we match it function for function, and make most of the functions that "public static" thing | 13:39.40 |
| or do we make a new object oriented interface? | 13:39.49 |
| or do we do both, in two steps ... a core JNI library that just exposes the C functions as-is and then wrap those calls in a separate Java library | 13:41.24 |
| which adds the expected object oriented fluff | 13:41.36 |
| somehow I think the two-step approach is going to be the best in the long run | 13:43.18 |
jogux | tor8: the two step approach is certainly what we're doing for the NUI project. | 13:49.46 |
| JNI exactly matching the C makes a lot of sense to me. | 13:50.35 |
Robin_Watts | tor8: the JNI broadly matches the C. | 15:18.36 |
| There is some magic in there to hide the fz_context stuff. | 15:18.57 |
| Certainly I'm attempting to keep the JNI stuff as small as possible. | 15:19.41 |
| See robin/jni2 for what I have so far. | 15:22.05 |
tor8 | Robin_Watts: we should write up the OO interface in some doc somewhere | 15:26.18 |
| I'm thinking of replicating it in JavaScript (so we can have a "mutool js") | 15:26.38 |
Robin_Watts | tor8: We should write up the java OO interface using javadoc. | 15:26.52 |
tor8 | using mujs bindings, which ought to look very much like the java ones | 15:27.15 |
Robin_Watts | at the moment, I'm trying to make the existing app use the new OO interface. | 15:27.25 |
| which is making me extend the interface as I find bits I've missed. | 15:27.41 |
tor8 | sounds like a good approach | 15:27.51 |
Robin_Watts | and also making me comment out lots of the app :) | 15:27.52 |
tor8 | tor/script has the beginnings of a "mutool js script.js" command | 15:28.22 |
| but only direct mirroring of the fz_ apis, not wrapping things in object like objects | 15:28.39 |
| I ran into headaches of where things like "fz_new_pixmap_from_page" and "fz_write_png" would live | 15:28.59 |
Robin_Watts | OK, all your commits except the last 3 look good. | 15:31.19 |
| Number 3 makes sense, but it's going to break gsview, I suspect. | 15:31.51 |
| We should warn mvrhel. | 15:32.00 |
| Number 2: OK, but I reserve the right to sigh every time I have to deal with: fz_text_span *head, *tail; | 15:35.34 |
| I prefer: fz_text_span *head, **tail; | 15:35.56 |
| as it normally gives simpler code. | 15:36.13 |
| but in this code, so far at least, it's not too bad. | 15:36.32 |
| Can't see anything wrong with the top one. | 15:38.04 |
| So, lgtm. | 15:38.11 |
tor8 | Robin_Watts: the tail should only be used in the one place | 15:49.49 |
| Robin_Watts: thanks. | 15:51.08 |
| mvrhel_laptop: breaking changes inbound to mupdf that will possibly affect gsview. I'm renaming the structs dealing with extracted text to reduce potential confusion with the structs used by the graphics library. | 15:52.20 |
mvrhel_laptop | tor8: ok that is fine. let me know when you commit | 16:39.30 |
tor8 | mvrhel_laptop: already done :) | 17:14.52 |
mvrhel_laptop | tor8: ok :) I will pull it in this weekend and get it working with gsview | 17:15.28 |
fredross-perry | does everyone agree that we donât want bitcode in the appstore build? If not, then thereâs no reason (as I see it) to use bitcode in the Xcode project at all. Thoughts? | 18:29.33 |
Robin_Watts | fredross-perry: I have to confess to not following Josephs argument. | 18:35.44 |
fredross-perry | Me too. I think it has to do with how Apple handles crash reports and symbols. | 18:36.27 |
| can someone please (Joseph) email me details of how to get the provisioning profiles? Thanks. | 18:54.59 |
| see what it says here | 18:59.08 |
| https://developer.apple.com/library/tvos/documentation/IDEs/Conceptual/AppDistributionGuide/AnalyzingCrashReports/AnalyzingCrashReports.html | 18:59.09 |
| about bitcode under "Viewing and Importing Crashes in the Devices Window" | 18:59.09 |
Robin_Watts | The provisioning profiles are probably on casper somewhere? | 18:59.54 |
mvrhel_laptop | tor8: you still around or done for the evening | 19:41.09 |
| ah. nevermind found what I was looking for (ft_kind) | 19:48.27 |
| bbiaw | 19:59.30 |
| lunch time | 19:59.33 |
jogux | fredross-perry: it's best to mention 'jogux' when you're talking about me btw, then my irc client does something that might bring it to my attention | 20:09.04 |
fredross-perry | ok, swell. sorry, thanks. | 20:09.38 |
jogux | actually maybe I can get it to do that for 'joseph' too | 20:09.50 |
| so: provisioning profiles, you get those from the apple provisioning portal | 20:10.07 |
| what you /might/ be missing is the appstore certificate | 20:10.13 |
mlq | Hi! I am trying to render a single annotation in mupdf, but I think I am missing something since it remains empty: http://pastie.org/private/qp6cz7cbg1ga2teuq1vw - Any input would be great :) Thanks! | 21:00.30 |
rayjj | I ordered my new laptop (a Lenovo W541) but it may not get here fast enough. My Dell is doing *really* strange things. The touchpad was dead, then it started working, then stopped again. The battery charge indicator said 36% remaining, I plugged it in, and after 5 minutes it says 100% charging (and says 100% even after I unplug it | 22:52.31 |
| oops. Now it says 39% -- so that's more believable | 22:53.19 |
mlq | I think I got it, I needed to render the complete page dimension for the annotation and then extract the area of the annotation | 22:53.38 |
| Forward 1 day (to 2015/12/12)>>> | |