| | <<<Back 1 day (to 2014/04/22) | 2014/04/23 |
| Robin_Watts_ | jogux, pedro, paulgardiner: 2 commits on robin/master. The python workaround for mkromfss2.py that michael needed and an updated solution. | 08:37.19 |
| pedro__ | nods | 08:38.54 |
| | Robin_Watts_: did you figure out what was actually going on with the retval? | 08:43.20 |
| | Robin_Watts_: in the android-sot.vcxproj (the nmakerebuildcommandline) you call solution_build.py but dont set the -debug flag (in the Debug|win32 config) | 08:50.59 |
| Robin_Watts_ | pedro__: Indeed. | 08:51.16 |
| pedro__ | is that intentional? | 08:51.29 |
| Robin_Watts_ | Currently -debug doesn't do much, certainly it doesn't set the debuggable attribute. | 08:51.46 |
| | Making -debug set debuggable is on my list of things to do, and I'll update the solution then. | 08:52.09 |
| pedro__ | cool | 08:52.21 |
| Robin_Watts_ | actually, you're right, I should probably set it now. | 08:52.35 |
| | No, never figured out what was going on with the retval. | 08:52.47 |
| | but the change there at least makes it work. | 08:53.05 |
| pedro__ | it looks like a pretty woory-free change for retval, so if it works for michael I'm hapy | 08:53.22 |
| | s/woory/worry | 08:53.29 |
| jhabjan | gs 9.14 behaves strange....when TextAlphaBits parameter is used, within a single pdf, for some pdf pages it applies this value and for some not.. | 08:53.54 |
| | sample PDF: https://www.dropbox.com/s/v7cw8q72ct8itkk/test.pdf | 08:54.40 |
| | command line: gswin64 -dTextAlphaBits=4 -dFirstPage=2 test.pdf | 08:54.56 |
| | gs 9.10 works normal | 08:55.18 |
| | any idea or workaround? | 08:57.02 |
| kens | jhabjan : are you sure the text is 'text' ? | 08:57.32 |
| Robin_Watts_ | gotta rebuild gs to look. | 08:57.37 |
| kens | Give me a minute, I have a current build | 08:57.48 |
| Robin_Watts_ | jhabjan: Can you give us a page where it goes wrong? | 08:58.13 |
| kens | Its trickling in for me :-( | 08:58.30 |
| jhabjan | kens: yep, it's a text | 08:58.39 |
| Robin_Watts_ | so page 2 is an example of a page where it looks bad. | 08:58.59 |
| jhabjan | Robin_Watts_: page no 2 | 08:59.02 |
| kens | is still waiting for hte file to arrive.... | 08:59.17 |
| jhabjan | Robin_Watts_: yep | 08:59.22 |
| Robin_Watts_ | jhabjan: Urm... why are the XObjects in that file listed as /Type /XObjcect ? | 09:01.25 |
| kens | hopes that wasn't me..... | 09:01.49 |
| jhabjan | Robin_Watts_: I have no idea... it's not PDF I built... | 09:02.05 |
| kens | Hmm, I'm seeing interpolation on the text, am I missing something ? | 09:03.02 |
| | This is the page with the 'Learnpub' log on it ? | 09:03.39 |
| | Sorry Leanpub | 09:03.52 |
| Robin_Watts_ | kens: yes. | 09:04.03 |
| kens | Well, with TextAlphaBots I see anti-aliasing in the text. | 09:04.17 |
| Robin_Watts_ | To the display device, I get no antialiasing that I can see. | 09:04.25 |
| kens | Oh, I used the tiff device | 09:04.35 |
| jhabjan | this is gs 9.14 output: https://www.dropbox.com/s/ixsgk987do9jlha/Screenshot%202014-04-23%2011.02.56.png | 09:04.59 |
| | and this gs 9.10 output: https://www.dropbox.com/s/d3rtyupciokoj1g/Screenshot%202014-04-23%2011.03.23.png | 09:05.26 |
| Robin_Watts_ | With tif I see AA. | 09:05.31 |
| kens | Yep, its not anti-aliased on the display device, but it is for the TIFF device | 09:05.41 |
| | I have no idea whay that would be | 09:05.58 |
| Robin_Watts_ | With png16m I see AA too. | 09:06.11 |
| | jhabjan: What device are you using ? | 09:06.18 |
| jhabjan | Robin_Watts_: display device | 09:06.28 |
| Robin_Watts_ | jhabjan: Can you open a bug at bugs.ghostscript.com please? Attach the file, state that it's a regression from 9.10 with the display device? | 09:07.47 |
| kens | 9.10 was OK ? | 09:08.02 |
| Robin_Watts_ | kens: Yes. | 09:08.57 |
| kens | Hmm, oh well at least we can bisect it | 09:09.06 |
| jhabjan | kens: yep | 09:09.57 |
| | Robin_Watts_: ok, I will fill a bug report | 09:10.19 |
| kens | Robin_Watts_ : I'm off out in 20 minutes, I'll bisect it this afternoon. As it happens I've just wrapped up all my named destinatiopn stuff | 09:10.29 |
| pedro__ | Robin_Watts_: rest of those commits look good to me. I'm out for a meeting for the rest of the morning btw | 09:11.56 |
| Robin_Watts_ | pedro__: Ta. | 09:12.31 |
| jhabjan | Robin_Watts_: done, http://bugs.ghostscript.com/show_bug.cgi?id=695180 | 09:17.38 |
| Robin_Watts_ | Thanks. | 09:17.58 |
| jhabjan | Robin_Watts_: no worries | 09:19.16 |
| Robin_Watts_ | It's to do with having an image on the page with transparency in it. | 09:22.48 |
| jhabjan | other than that... when Ghostscript is installing, for the GPL version, installer writes some registry keys about Ghostscript version into "SOFTWARE\GPL Ghostscript\". Does commercial version writes into "SOFTWARE\AFPL Ghostscript\" ? | 09:24.46 |
| kens | Nope. There is no installer for the commercial version, we expect people to build it into their own product and therefore have their own instgaller. | 09:25.36 |
| | AFPL = Aladdin Free Public Licence I believe, and we have't distributed GS that way for some time | 09:26.08 |
| chrisl | There is an installer for the commercial version, it should write to SOFTWARE\Artifex Ghostscript\ | 09:26.18 |
| kens | Huh, didn't know that | 09:26.29 |
| | back later | 09:27.50 |
| jhabjan | so, are this two places the only ones in the last years? | 09:27.57 |
| chrisl | jhabjan: in the registry? Yes, should be | 09:28.20 |
| jhabjan | chirsl: yep | 09:28.32 |
| | chrisl: gr8, thanks | 09:28.42 |
| chrisl | TBH, I'd be very surprised if you came across a case of "Artifex Ghostscript", though | 09:29.14 |
| jhabjan | chrisl: I understand... just want to cover all possible scenarios :) | 09:30.12 |
| | chrisl: I think I found AFPL Ghostscript key mentioned here: http://ghostscript.com/doc/8.54/Install.htm | 09:30.36 |
| chrisl | jhabjan: that document is for a release over 7 years old.... I'll check the current ones | 09:31.41 |
| Robin_Watts_ | kens: Simplified file, attached to the bug. It's a transparency issue I think. | 09:32.39 |
| chrisl | jhabjan: the only mention of AFPL in the current docs is in removing GS, which I think is fair, given people may be uninstalling old versions in order to update | 09:33.14 |
| | I need to go for a haircut, back in a while | 09:34.00 |
| jhabjan | chrisl: thanks for clearing out this for me | 09:34.23 |
| Robin_Watts_ | paulgardiner, jogux: pedro reviewed the last 2 commits on robin/master (sot) for me earlier, and commented that I'd forgotten to set -debug on the debug android-sot build on the last one. | 10:03.19 |
| | So I've just updated that. | 10:03.25 |
| jogux_mac | can take a quick look | 10:03.35 |
| jogux_mac | wishes we had a gitweb | 10:03.43 |
| Robin_Watts_ | If one of you could nod at the revised version I will commit it after I run. | 10:03.46 |
| | jogux_mac: yeah :( | 10:03.55 |
| jogux_mac | reckons we could do one using the wiki authentication. but I really need to finish ATS before I start messing with new stuff :-) | 10:04.28 |
| jogux_mac | thinks I need to try this visualgdb thing. although I thought I'd managed to escape windows. :-S | 10:05.34 |
| Robin_Watts_ | The wiki authentication collides with the dashboard authentication at the moment. pain in the ass. | 10:05.43 |
| jogux_mac | oh, how come? | 10:05.53 |
| Robin_Watts_ | dunno, but both things ask me to log in now. | 10:06.08 |
| Robin_Watts_ | runs. bbiab. | 10:06.32 |
| jogux_mac | well the wiki perpetually asks me to login. :( | 10:07.08 |
| | and I don't use the dashboard afaik | 10:07.14 |
| | robin_watts : btw, iirc, if you set GENROOT_OVERRIDE (to to genroot-<configurationname>), you could then avoid having all the indows\x86\vs2010\debug\nonet\romfss2\heap\all_das\debug_puts_enable\excel_edi blahblahblah I reckon. if VS can set env variables per config. | 10:09.44 |
| | and python doesn't need ()s around if expressions :-) | 10:10.09 |
| | you don't need -define=DEBUG_PUTS_ENABLE if you have -debug (that's in the da/test-shell one) | 10:12.57 |
| | (it also does no harm.) | 10:13.57 |
| jhabjan | is there any chance you could take into consideration for making something like Virtual IO module that would allow users to write custom file read, seek, write functions and expose it via Ghostscript API? | 11:05.29 |
| | is there any chance you could take into consideration for making something like Virtual IO module that would allow users to write custom file read, seek, write... functions and expose it via Ghostscript API? | 11:05.29 |
| | (sorry for double lines, it was accidentally) | 11:06.11 |
| chrisl | jhabjan: that's something I've been wondering about. It *should* be possible without breaking loads of stuff, but this is Ghostscript - "should" does not always map to "is" | 11:09.18 |
| | jhabjan: and there is a possible hiccup which is that Ghostscript is not consistent in its use of "gs streams" - in various places (particularly devices) it uses standard FILE * objects...... | 11:11.52 |
| | jhabjan: it would be best to open enhancement requests on bugzilla for things like this..... | 11:16.04 |
| jhabjan | chrisl: ok, got it, thanks... I was just wondering if it's anything that could be easily done in Ghostscript...as it could be helpful feature... | 11:18.24 |
| chrisl | jhabjan: I agree - but I think it wouldn't be as useful in the current architecture, as if we used gs streams throughout the code | 11:19.50 |
| | jhabjan: as it is, it's probably only usable for *input* files | 11:20.10 |
| jhabjan | chirsl: not only for input files... I was thinking more like... everything could be done in-memory... setting input file, output file | 11:22.34 |
| chrisl | jhabjan: no, you can't do output files, that's my point - output files aren't written through gs streams | 11:23.13 |
| jhabjan | chrisl: it would be easy to compile everything into a single dll ( with embedded resource files ) which Virtual IO could serve to GS | 11:23.15 |
| chrisl | jhabjan: the device API uses standard FILE * | 11:24.03 |
| jhabjan | chrisl: ah.. got it.. anyway.. output files could be handled via pipe... | 11:24.28 |
| chrisl | jhabjan: well, you can simulate what you're looking for by hooking the stdin and stdout calls | 11:25.33 |
| | But that's not really ideal - and doesn't really work "properly" for output files :-( | 11:26.27 |
| | jhabjan: but as said, open an enhancement: it's certainly not going to be trivial to implement, so will have to be "officially" considered, prioritised etc..... | 11:29.53 |
| jhabjan | chrisl: ok, thanks | 11:30.15 |
| chrisl | jhabjan: FWIW, why the device APIs (and everywhere else) weren't changed to use streams when streams were introduced is a complete mystery to me..... | 11:31.39 |
| jhabjan | chrisl: no worries ;-) | 11:33.02 |
| Robin_Watts_ | jogux_mac: solutionbuild.py does set GENROOT and GENROOT_OVERRIDE | 11:35.33 |
| jogux_mac | oh well, guess I was wrong about that then. It may be worth adding a GENROOT_OVERRIDE that actually lets us remove all those nasty paths :) | 11:36.04 |
| Robin_Watts_ | That saves it being /genroot/D--cvs--artifex--sot.git/.... | 11:36.17 |
| jogux_mac | ah | 11:36.27 |
| | oh well. | 11:36.29 |
| | shame. but fine for now, anyway. | 11:36.43 |
| Robin_Watts_ | tor8, paulgardiner: I am about to do the MuPDF Android builds. Nothing waiting to be committed by anyone, right? | 12:21.18 |
| tor8 | Robin_Watts_: there's a handful of bugfix commits on tor/master but nothing major | 12:21.42 |
| Robin_Watts_ | tor8: Let's look at them before I get started then. | 12:23.57 |
| | tor8: Wouldn't it be better just to never add a newline ? | 12:25.32 |
| tor8 | Robin_Watts_: yeah, that's a good question | 12:25.59 |
| | I would expect if I select a whole line I'd get the newline, but not getting one would be better than always getting one | 12:26.25 |
| | and it would be simpler code | 12:26.33 |
| Robin_Watts_ | If I select 2 lines, then I would expect to get a newline between them, maybe. | 12:26.47 |
| | But if I select a complete line, I wouldn't expect a newline. | 12:26.58 |
| | But if it's 'flowed' text (i.e. a single sentence split over 2 lines) I would ideally like to not get the newline. | 12:27.25 |
| tor8 | Robin_Watts_: yeah, but then you get into trouble with table cells, and reflowing vs hard breaks etc that we just can't reliably tell | 12:28.07 |
| Robin_Watts_ | Now, we can't easily tell if text is flowed or not, so I could live with always getting newlines. | 12:28.11 |
| | tor8: Right, so always giving newlines within text seems fair. | 12:28.25 |
| | But adding a newline at the end in all cases seems hinky. | 12:28.36 |
| tor8 | yeah. if I remove five lines from the patch I can make it only give newlines between actual lines | 12:29.02 |
| | and never at the end | 12:29.07 |
| | as written it adds between lines, and at the end if the final character in the final line is selected | 12:29.30 |
| Robin_Watts_ | Perhaps we should only add a newline at the end if there have been newlines during the text (and if the final char in the final line is selected) | 12:29.45 |
| | and we should use an enum, not an int for need_newline. (I'm sure I'm guilty of breaking that rule, so feel free to hit me back with it in future) | 12:31.31 |
| tor8 | I think a way that has as few surprises as possible is likely better | 12:31.32 |
| Robin_Watts_ | least surprise is never a bad rule. | 12:31.44 |
| tor8 | so only newlines between text, never at the end, and then the need_newline becomes a boolean value rather than an enum | 12:32.05 |
| Robin_Watts_ | nice. | 12:32.22 |
| | Other reviews all look fine to me. | 12:32.27 |
| tor8 | Robin_Watts_: if you fetch again, there's an updated version | 12:32.28 |
| Robin_Watts_ | tor8: Can you rebase to golden ? | 12:32.54 |
| tor8 | sure, one sec | 12:33.08 |
| Robin_Watts_ | the commit message has gotten corrupted on that commit. | 12:33.25 |
| tor8 | right, will fix | 12:33.43 |
| | done | 12:35.11 |
| Robin_Watts_ | if (!need_newline && saw_text) need_newline = 1; could be simplified ? | 12:36.07 |
| | need_newline |= saw_text ? | 12:36.18 |
| tor8 | if (saw_text) need_newline = 1; you're right | 12:36.46 |
| | this code has been through too many (but not enough) iterations | 12:37.04 |
| Robin_Watts_ | tor8: ain't that always the way :) | 12:37.17 |
| tor8 | that case used to emit the newline directly, but not so much anymore :) | 12:37.23 |
| | any other simplifications you can spot? | 12:37.45 |
| | we might be able to eliminate saw_text altogether, but I think that might be more difficult to understand code | 12:38.33 |
| Robin_Watts_ | I've looked at that, and I think to do that we'd need to split the span loop into two. | 12:39.07 |
| | first a loop until we have a non empty hitbox, then output the newline, then the rest of the loop. | 12:39.36 |
| | If this was speed critical we might consider it, but I think it's fine as is. | 12:39.51 |
| tor8 | yeah. this way is probably simpler. | 12:39.54 |
| | Robin_Watts_: I've pushed a fix for the if-expression | 12:41.04 |
| | if you're happy I'll push that to golden | 12:41.24 |
| Robin_Watts_ | push away! | 12:41.39 |
| tor8 | thanks. | 12:41.53 |
| kens | I've reassigned bug #695180 to Ray because it was his large clist processing for non-clist devices that introduced it. | 13:00.07 |
| | Hopefully Ray can sort out the problem quickly | 13:00.40 |
| henrys | chrisl: I thought that gtk+ stuff had bit rotted - I haven't used it in ages. | 13:18.53 |
| chrisl | henrys: there's a few deprecated calls in it, but it still works, and is nicer than x11 | 13:19.27 |
| henrys | chrisl: funny on ubuntu it doesn't install automatically for me. Do I need the "dev" libraries? | 13:19.51 |
| chrisl | henrys: you mean in the package manager? I've no idea | 13:20.27 |
| henrys | chrisl: no I mean odd the gtk stuff didn't get "autogened" in to the gs build - I would think gtk would have to be available on ubuntu to support everything else. | 13:22.34 |
| chrisl | henrys: but you'll need the "dev" gtk+ packages | 13:23.14 |
| henrys | chrisl: James comment got me thinking about the default device on unix. X11 has always been a poor first impression. | 13:32.04 |
| chrisl | henrys: it's hard to see what else we could opt for without adding a load of new dependencies | 13:32.49 |
| | henrys: plus there might be legal implications if our default build on Unix linked to gtk+ | 13:35.30 |
| henrys | chrisl yeah but geez x11 sucks | 13:39.20 |
| chrisl | henrys: well, strictly speaking there *shouldn't* be a legal problem with making the normal gs build link to gtk, if it's available | 13:40.08 |
| | henrys: and I could have the configure script fallback to not including gtk and issue a warning if the libs aren't found | 13:41.02 |
| henrys | chrisl: gtk is reasonably widespread on unix systems - I'm not sure if it usually is around on qt based systems. | 13:41.13 |
| chrisl | henrys: So? | 13:41.28 |
| | henrys: any toolkit we chose would merit a similar observation | 13:42.29 |
| henrys | chrisl: indeed, let me do some testing of the gtk display maybe you can look at it too if you get a chance. We don't want to change if there are lots of problems | 13:45.12 |
| chrisl | henrys: if we want to make it the default, we should get rid of the deprecated calls, and update it properly to the latest gtk+/glib APIs | 13:46.30 |
| henrys | chrisl: I'll put it on the agenda. | 13:48.32 |
| Robin_Watts_ | In theory, doing an android release should be 4 builds. | 14:25.13 |
| | I'm up to 8 already :( | 14:25.19 |
| tor8 | why so many? | 14:25.27 |
| | too many flavours of arm? | 14:25.40 |
| Robin_Watts_ | armeabi/armeabi-v7a/x86/mips | 14:25.51 |
| jogux_mac | I don't quite understand why mips seems to keep resurfacing lately. I could swear it died ~6 years ago. | 14:35.45 |
| pedro__ | a lot of in-car systems used to use it, but it does seem a bit odd. Possibly just a bunch of companies 'protecting their investment' in all things mips | 14:39.42 |
| paulgardiner_lap | tor8: 5 iOS-related commits on paul/master. Also a new test build on flighttestapp | 14:44.52 |
| chrisl | jogux_mac: I suspect they're doing doing fairly cheap license agreements for MIPS to try to weather the ARM storm..... | 14:45.35 |
| | Crumbs, apparently "75 percent of Blu-ray Disc players are running on MIPS Technologies processors" | 14:46.47 |
| | And I did not know that the Sony PS runs MIPS cores | 14:47.26 |
| | s/PS/PSP | 14:47.32 |
| paulgardiner_lap | tor8: not sure how memory friendly the bitmap-based printing is. On iOS, to make use of a bitmap, you have to pass ownership of it, so I wasn't able to reuse one bitmap. I have to create multiple ones, hand them off to the OS and hope it frees them in a timely fashion. I release the CGDataProviderRef each time, but that doesn't guarantee instant freeing of the data | 15:16.19 |
| | Robin_Watts_: thirdparty-openssl.zip is on casper | 15:17.06 |
| Robin_Watts_ | Thanks. | 15:17.16 |
| tor8 | paulgardiner_lap: right. I'm not sure if there's much we can do about that. in the best case, the iOS code will free it early enough, but if not, we're no worse off than printing a full-page bitmap all at once | 15:18.17 |
| paulgardiner_lap | tor8: yep. It seems to work anyway. I've kept to 300dpi because 600 failed on my iPod touch (it looked like it got killed because of the time it took) | 15:20.05 |
| tor8 | paulgardiner_lap: fair enough. | 15:20.57 |
| mvrhel_laptop | good morning | 15:29.33 |
| Robin_Watts_ | mvrhel: morning | 15:30.15 |
| kens | Hi Michael | 15:31.26 |
| tor8 | Robin_Watts_: two more commits on tor/master for review for whenever you've got a moment | 15:43.48 |
| Robin_Watts_ | tor8: I would be tempted to put a comment by the fz_warn :) | 15:44.58 |
| | otherwise in 6 months I'll add the codepoint back in... :) | 15:45.12 |
| | otherwise both look good. | 15:46.08 |
| | OK, I've got new versions of MuPDF up for beta testing. | 16:03.28 |
| | If people visit: https://play.google.com/apps/testing/com.artifex.mupdfdemo they can become testers and get the new versions. | 16:03.42 |
| | Any testing people can do would be appreciated. | 16:03.52 |
| jogux_mac | wishes Robin better luck than Paul has had persuading people to try the iOS one :-) | 16:04.35 |
| paulgardiner_lap | Robin_Watts_ I get: "Sorry, there is nothing to see for you here." | 16:11.35 |
| | tor8: did you get a chance to check over those iOS commits? | 16:11.56 |
| kens | Robin_Watts_ : Hmm it wants me to sign in with a Google account when I try to visit that | 16:12.25 |
| | Then I get the 'nothing to see' message, presumably because I'm not using an Android device | 16:13.07 |
| Robin_Watts_ | Gah. It seems that in order to have people as beta testers, either I need to add them manually, or they need to join the "MuPDF Beta testers" community on plus.google.com | 16:17.14 |
| | Alternatively, you can get to the apks from: http://ghostscript.com/~robin/MuPDF-44.apk or http://ghostscript.com/~robin/MuPDF-45.apk | 16:18.02 |
| | those are the ARM ones. 44 should run on anything, 45 on anything modern. | 16:18.19 |
| | I have x86 and mips ones for anyone with such devices. | 16:18.46 |
| | bah. It's going to take a day or so for the apks to be visible via the play store. | 16:19.35 |
| paulgardiner_lap | public_html ones will be fine | 16:20.42 |
| Robin_Watts_ | getting set up for the beta testing group might be good for future releases though, as it means people will automatically get upgraded to the beta versions. | 16:21.59 |
| jogux_mac | paul / robin : there's two fairly trivial commits on my master if you want to take a look. no rush. | 16:22.41 |
| Robin_Watts_ | bah. It dies for me. | 16:25.45 |
| | jogux_mac: Both look fine to me. | 16:29.41 |
| kens | Goodnight all | 16:32.42 |
| jogux_mac | robin: thanks | 16:32.50 |
| | pushed now. | 16:34.39 |
| paulgardiner_lap | Robin_Watts_: MuPDF-45.apk dies on start up on my S2 :-( | 16:39.05 |
| Robin_Watts_ | paulgardiner_: Yeah. Me too. It's having trouble with some resources, and I can't see why. Investigating now. | 16:39.41 |
| | I *love* doing releases :) | 16:39.54 |
| henrys | paulgardiner_lap: had a problem with the iOS install on an iPad - emailing you cryptic screenshot in a minute | 16:48.51 |
| pedro__ | Robin_Watts - does the version string in strings.xml get replaced as part of the release build process? | 16:49.59 |
| Robin_Watts_ | pedro__: Only by me :) | 16:50.26 |
| pedro__ | grins - that's the resource its failing on for me | 16:50.41 |
| Robin_Watts_ | http://twiki.ghostscript.com/do/view/MuPDF/AndroidReleases | 16:50.42 |
| | yeah, for me too. | 16:50.47 |
| | Everyone should assume it's my fault for now, and I'll try some rebuilds/tests etc before bothering everyone again. | 16:51.32 |
| | There is a new MuPDF-44.apk there (armeabi) that seems to work for me. | 16:55.44 |
| | Rebuilding armeabi-v7a now. | 16:55.56 |
| pedro__ | Robin_Watts: fine on S4 too | 16:58.31 |
| Robin_Watts_ | My new -45 build tests out here too. Must have been finger trouble in the build. | 16:59.11 |
| | I'll upload that in a mo. | 16:59.15 |
| | oh, I don't believe it. | 17:01.22 |
| | You can't delete an apk off google. | 17:01.28 |
| | so I can't delete -45 and replace it. | 17:01.43 |
| | which means I need to give it higher number. | 17:02.02 |
| | which buggers up all the other numbers. So I need to rebuild all of them. | 17:02.18 |
| | Actually, no, I think I can get away with just this one, but that's just through luck. | 17:02.57 |
| | http://ghostscript.com/~robin/MuPDF-49.apk should hopefully work. | 17:12.52 |
| | Ok, so builds are uploaded to google. I'll wait 24 hours for any feedback then publish. | 17:20.09 |
| pedro__ | Robin_Watts: I've just downloaded 49 but adb install doesn't like it - thinks its a duff apk | 17:22.09 |
| pedro__ | retries | 17:22.36 |
| Robin_Watts_ | 4959474 bytes ? | 17:22.49 |
| pedro__ | just checked - seems to have been truncated by about 400 bytes | 17:23.09 |
| | possibly just ffox although it claimed it succeeded | 17:23.28 |
| Robin_Watts_ | pedro__: odd. | 17:23.47 |
| pedro__ | cool - its fine now | 17:24.23 |
| Robin_Watts_ | phew. | 17:24.44 |
| pedro__ | shifts to home - will give the Good iOS build a whirl later | 17:26.23 |
| | Robin_Watts: based on the old iOS SDK you found in the SDK | 17:26.51 |
| Robin_Watts_ | pedro__: Ah, best of luck. | 17:27.03 |
| pedro__ | what can possibly go wrong? ... | 17:27.22 |
| paulgardiner_lap | henrys: oh okay. I'll take a look. | 17:30.47 |
| | henrys: right. I can add that device to the provisioning profile. but probably I'll need to regenerate the test app | 17:36.15 |
| henrys | paulgardiner_lap: if you feel it needs testing then go ahead. | 17:37.17 |
| | paulgardiner_lap: but if a few others have looked at it, let's just skip it this time. | 17:37.47 |
| paulgardiner_lap | henrys: is that the device you tested on before? | 17:37.53 |
| henrys | paulgardiner_lap: no I think I used my phone last time | 17:38.25 |
| paulgardiner_lap | Ah okay. I'll see how the testing goes. I'll add that device in anyway ready for next time | 17:38.56 |
| | henrys: any other devices you might like to test on, you can register with testflight. I think it will then tell me when I upload something that has a device not included | 17:41.59 |
| henrys | paulgardiner_lap: will do | 17:42.49 |
| tor8 | paulgardiner_mac: ShareAlertMesage typo in comment | 18:55.51 |
| | paulgardiner_mac: other than that, LGTM | 18:57.39 |
| Robin_Watts_ | Oh, I love android tools. | 21:01.15 |
| | So you're doing a debug build, right? You'll want me to strip out the debugging information from your lib then. | 21:01.39 |
| | Woo Hoo! | 21:06.50 |
| | Following some hacking of the build scripts I have it so I can debug SOT on a device from MSVC | 21:07.23 |
| jogux | awesome :-) | 21:22.11 |
| | debugging both the java and the C? | 21:22.31 |
| sebras | tor8: 4cb52a8 lgtm. 0b5c38d tttr. | 22:08.08 |
| tor8 | sebras: tttr? | 22:09.04 |
| | tired? tough? tricky? | 22:09.38 |
| | sebras: git show -w 0b5c38d ought to be easier to read :) | 22:10.55 |
| mvrhel_laptop | Robin_Watts_: ok page extraction in gsview is working nicely now. you can reorder, delete etc and select which ones you want to save | 22:23.16 |
| | this is using the call to pdfclean | 22:24.11 |
| | now I just need to add a few conversion types and that will be all the hard work done | 22:35.39 |
| Robin_Watts_ | tor8: guessing... too tired to read ? | 23:07.26 |
| | jogux: I think so. | 23:07.47 |
| | Though I was only looking at the C at the time. | 23:07.57 |
| | jogu: ah, no. VisualGDB will only let me debug the native code. | 23:17.04 |
| | but supposedly you can use eclipse to debug the java at the same time (both debuggers attach) | 23:17.21 |
| | mvrhel_laptop: Sounds good. | 23:20.28 |
| | Are you still seeing the bouncy baseline bug ? | 23:20.56 |
| markson | is there any chance of the scroll-to-next-screen's worth of a page pathces being accepted? | 23:52.19 |
| | Forward 1 day (to 2014/04/24)>>> | |
IRC Logs
