| <<<Back 1 day (to 2012/10/09) | 2012/10/10 |
LeastOne | When I use gs with -sDEVICE=pdfwrite in interactive mode the output file is not being flushed (its only partially written) ... Does anyone know if there is an option to force a flush of the output file? | 01:58.50 |
mvrhel_laptop | ray_work you there? | 04:51.20 |
| done for the night | 04:59.50 |
| good night all | 04:59.53 |
kens | LeastOne, in order to get the PDF file 'flushed' you need to terminate the device. This is normally done by closing Ghostscript but it can be done in several other ways. The most common is to write your own application using the GS API, rather than our own binary. The other way is to launch GS with one device active, 'save' the context, use a setpagedevice call to switch to the pdfwrite device. When finished with the pdfwrite de | 07:13.23 |
| vice 'restore' back to the original context. This will close the device. Obviously you need to do this for every input file, so this will involve writing a small PostScript program. | 07:13.23 |
kens2 | Robin_Watts : (for the logs) did you read the reply on bug #693367 ? Given that this problem only occurs when they use *their* application, I think the simplest approach is to have them put together a skeleton app and send it to us for us to debug. What do you think ? | 08:50.34 |
| Especially given that teh snippets of code tehy've sent don't actually go the GS API calls. | 08:51.43 |
Robin_Watts | kens2: I haven't read the new attachment yet. | 09:14.07 |
kens2 | Robin_Watts : it mostly says 'the gs application works, ours doesn't. Here's some fairly useless C++ stuff that won't tell you anything' | 09:14.59 |
| The zip archive contains some files, but since they already say those work with GS itself, there's not much we can do with them | 09:16.01 |
| Hence I think it would be best to get them to throw a tiny application together to demonstrate the problem. | 09:16.29 |
Robin_Watts | kens2: sounds reasonable to me. | 09:16.42 |
kens2 | I'm about to go out, if you don't feel like updating the bug I'll do it when I get back, or you can do it before then, whichever you prefer. | 09:17.36 |
Robin_Watts | ok. | 09:18.56 |
| kens: That's not C++. That's (shudder) *java*, I think. | 09:31.49 |
sebras | Robin_Watts: tor8: there are two small patcher over at sebras/master | 11:52.12 |
Robin_Watts | ok. | 12:01.22 |
| Both look fine to me. | 12:01.28 |
sebras | Robin_Watts: great. | 12:02.36 |
| the encryption dict one might be unnecessary, I'm hoping for tor8 to give some input on that one. | 12:03.16 |
Robin_Watts | sebras: I'll push them now. | 12:03.33 |
| Did you look at my "fix android build" one? | 12:03.48 |
| Oh, and you said you had problems with my fz_functions one... | 12:04.03 |
sebras | Robin_Watts: yes, build problems. | 12:05.23 |
Robin_Watts | on what platform? | 12:05.33 |
sebras | but the android-build patch looks good. | 12:05.38 |
| linux. | 12:05.42 |
Robin_Watts | Ta. | 12:05.45 |
| tor8, sebras, henrys: http://blogs.adobe.com/typblography/2012/09/source-code-pro.html | 12:06.23 |
tor8 | Robin_Watts: already using it in my terminal on osx. it's a pretty nice font, for being monospaced. | 12:07.57 |
sebras | tor8: opinions on making mutool show show the encryption dict? | 12:13.29 |
tor8 | sebras: sounds good to me | 12:23.19 |
Robin_Watts | just as well, cos I pushed it already :) | 12:25.33 |
sebras | :) | 12:25.45 |
LeastOne | kens: Thank you for your suggestion, I think I'll be able to make it work. | 13:42.46 |
kens | LeastOne, No problem | 13:44.03 |
Robin_Watts | Stupid reboot requiring windows. brb. | 13:57.32 |
| kens: So are you all up and running on your new uber-box then ? | 14:33.47 |
kens | Yes, all sorted now. Need to pull *lots* of stuff across, but the network is up and running properly now | 14:34.15 |
Robin_Watts | cool. Including VMware images etc? | 14:34.27 |
kens | Not yet | 14:34.35 |
| Need to copy stuff | 14:34.40 |
| I'll get there soon though | 14:34.57 |
Robin_Watts | Did my reply to bug 693367 cover everything ? | 14:35.05 |
kens | Admirably I think | 14:36.01 |
| I don't see how we can debug a problem we can't even see | 14:36.13 |
Robin_Watts | indeed. I'd hope that the dumping of bytes should make it clear where the problem is, and that it's not in our domain. | 14:36.43 |
kens | Mind you, I've no idea how to debug Java on a Mac | 14:36.43 |
| Yeah well they claim their code han't changed... | 14:37.08 |
Robin_Watts | IME Java's debugging extends to printf (or System.out.writeLn or whatever it is) | 14:37.19 |
| I bet it's going to be something like them changing the Locale of the system, and suddenly they aren't passing UTF8 any more. | 14:37.55 |
kens | I wouldn't be surprised, I can't believe its anything to do wiht us | 14:38.23 |
| Not on a Mac :-) | 14:38.33 |
| I have zero confidence in the contracors ability | 14:39.20 |
paulgardiner | Robin_Watts, tor8: ping | 15:30.17 |
Robin_Watts | pong | 15:32.45 |
paulgardiner | I was going to add an fz_ function to ask if javascript support is present. Does that sound ok? | 15:33.50 |
Robin_Watts | so a build time thing that'd be 0 in the non-v8 builds and 1 in the v8 ones? | 15:34.39 |
paulgardiner | yep. I already have two implementations of the pdf_js api which is where the difference is naturally reported | 15:35.43 |
Robin_Watts | sounds reasonable to me. | 15:35.59 |
paulgardiner | ok ta | 15:36.03 |
Robin_Watts | The alternative is that we have a pdf_get_gs function to get that interface. | 15:36.21 |
| And that returns NULL in the non js case. | 15:36.31 |
paulgardiner | That's a thought... but I don't think that interface is suitable for exposing. | 15:38.43 |
mvrhel_laptop | good morning | 15:44.46 |
Robin_Watts | Morning. | 15:46.09 |
| Day #2: Any urges to patricide yet? | 15:47.49 |
mvrhel_laptop | hehe | 15:49.04 |
Robin_Watts | Anyone here have permissions to add a v1.1 tag to the bugzilla choices for mupdf? | 16:32.17 |
| Aha. A Marcosw, just the man. | 16:36.06 |
marcosw | uh oh. | 16:36.11 |
Robin_Watts | Could you add a 1.1 option to the versions list in bugzilla for mupdf please? | 16:36.28 |
marcosw | sure. | 16:36.34 |
| done. | 16:37.22 |
Robin_Watts | Thanks. | 16:39.55 |
ray_laptop | Just found a typo that's probably been there a LONG time. base/gxpcmap.c:382 sets the BandHeight = size.x (instead of y) :-( | 21:58.43 |
| doesn't really hurt because there is a hack elsewhere that force pattern_clist nbands to 1 | 22:04.03 |
Robin_Watts | ray_laptop: Stuff like that is lovely :) | 23:21.40 |
ray_laptop | Robin_Watts: are you commenting on the hack that makes it work anyway ? | 23:22.11 |
Robin_Watts | Hey henrys`. How's the show? | 23:22.16 |
| ray_laptop: No, I mean it's always nice to find stuff like that. | 23:22.28 |
| The "How the hell has it ever worked?" moments. | 23:22.49 |
ray_laptop | Robin_Watts: would be more satisfying if finding it actually mattered (some obscure problem that got fixed). | 23:23.40 |
| Forward 1 day (to 2012/10/11)>>> | |