| <<<Back 1 day (to 2013/06/07) | 2013/06/08 |
marcosw | mvrhel_laptop: sorry, I messed up your cluster job. I'm restarting it now. | 04:44.48 |
sebras | malc__: good morning. | 07:41.29 |
malc__ | sebras: tor said that he passed the info to you that remote /F doesn't quite work | 08:17.59 |
sebras | malc__: he mentioned that, yes. but did you discuss it? | 09:38.20 |
| malc__: do you know why..? | 09:38.26 |
malc__ | sebras: no. i do know how to reproduce though, not why | 09:53.12 |
sebras | malc__: do you have a sample file? | 10:35.11 |
| malc__: I'm happy to take a look. | 10:35.24 |
| robin_watts_mac: have a look at my repo. there's a "d'oh"-commit in there. :-P | 10:39.21 |
malc__ | sebras: http://www2.lauterbach.com/pdf/commandlist.pdf | 10:45.37 |
sebras | malc__: oh.. you use trace32? | 10:50.20 |
malc__ | sebras: aye | 10:50.26 |
sebras | malc__: I used it a lot at my old workplace. | 10:50.42 |
| fantastic debugger but ugh! what a user interface! :-P | 10:50.52 |
malc__ | sebras: agreed :) | 10:50.58 |
| but as long as it allows me to debug this bigLITTLE thingy i'm happy | 10:51.15 |
sebras | ah. over JTAG I presume? | 10:51.35 |
malc__ | sure | 10:51.40 |
sebras | are you using ETM as well? | 10:51.44 |
| that is the feature that makes trace32 fantastic imho. | 10:51.56 |
malc__ | actually i've no idea what ETM is | 10:51.56 |
sebras | malc__: it's a part of the ARM core which records all instructions and data/register accesses and packages this info for a HW debugger.\ | 10:52.32 |
malc__ | oh, then yeah, i'm definitely using that | 10:52.48 |
sebras | this makes it possible to put HW breakpoints in the code and then you get a full sequence of what the CPU was doing up to that point. | 10:52.56 |
malc__ | erm.. | 10:53.10 |
| hum.. | 10:53.15 |
| no i'm not apparently | 10:53.18 |
sebras | so basically it means you can go back in time. | 10:53.20 |
malc__ | sebras: sortof like ocamldb reverse debugging, neat. | 10:53.34 |
| would have to give it a shot | 10:53.39 |
sebras | ok :) so anyway, I'll take a look at the pdf. | 10:53.48 |
| is there any particular link that is problematic or is it all of them? | 10:54.00 |
malc__ | sebras: was stomped yesterday.. didn't know how to write to IO memory via trace32.. any idea? | 10:54.18 |
sebras | malc__: I would try data.write or something like that but I haven't touched trace32 for 1.5years. I'm stuck in userspace programming (gstreamer) with my new employer. | 10:55.20 |
malc__ | sebras: accept my deepest condolences | 10:56.00 |
| glib... shrug | 10:56.04 |
sebras | malc__: thanks. I think I need that. :) | 10:56.11 |
malc__ | [ftw. data.write didn't work] | 10:56.15 |
| fwiw rather | 10:56.20 |
sebras | malc__: is it the links to other files on page 1 of commandlist.pdf that causes problems? | 11:01.58 |
| malc__: I added some extra debugprints and ran mupdf through valgrind bug I can't see that there would be anything causing segfaults..? | 11:02.37 |
malc__ | sebras: it's not about segfaults, it's about those links containing empty targets | 11:03.36 |
| as for segfaults that's different issue (still bidi and use after free related i guess), nothing to do with your /F patch | 11:04.09 |
sebras | malc__: you need to elaborate a bit more. I print the utf8-string in ld.ld.gotor.file_spec and it looks like filenames where. | 11:05.53 |
| where -> here. | 11:05.56 |
malc__ | sebras: sec | 11:11.41 |
| sebras: uh.. right you are.. that said, fz_warn is annoying | 11:14.03 |
sebras | malc__: http://pastebin.com/S614xKeM this is what I see... | 11:14.16 |
malc__ | sebras: http://pastebin.com/raw.php?i=neiWz7jZ | 11:15.31 |
sebras | malc__: yes, sorry about those warnings, they shouldn't be there. | 11:21.10 |
| I used them for debugging as you saw me do just now, but I forgot to remove them before checking in. ?-) | 11:21.32 |
malc__ | sebras: naughty! | 11:23.23 |
sebras | yes, and no one caught this in code review. :( | 11:26.03 |
| malc__: is the prints the only reason you worry? | 11:26.21 |
malc__ | sebras: w.r.t. remote actions - yes (probably mislinked something initially) | 11:27.11 |
sebras | malc__: http://git.ghostscript.com/?p=user/sebras/mupdf.git;a=blobdiff;f=pdf/pdf_annot.c;h=1d7ca98d066038dc2e5285f0944ee29e1f0df953;hp=4fe01a0bbfc4d76eb7dbd5f7a7722c5c5ab4fd4a;hb=9390304b7ef3c965b2f780f92356a0de0d20cee3;hpb=40599eb00e746c0aa6e86a46115a109e073a8890 if you apply this patch when they'll disappear. | 11:28.11 |
| I mentioned this to Robin before in the hopes that he or tor8 will apply the patch on origin/master. | 11:28.33 |
malc__ | sebras: tack | 11:29.57 |
sebras | malc__: they're at a meeting in the .us right now so it will probably take some time until they merge it. | 11:30.47 |
malc__ | it's saturday so they are probably boozing instead | 11:32.00 |
sebras | malc__: I spoke to tor8 just now and he pushed my new patch to mupdf master. | 11:40.39 |
malc__ | sebras: tusen tack | 11:41.54 |
chrisl | It's 7:40am here - it's a bit early for boozing - even for me! | 11:42.26 |
malc__ | jetlag usually takes care of that | 11:42.50 |
chrisl | It is slightly tempting just to crawl back to bed..... | 11:44.09 |
malc__ | sebras: thanks once gain, time to flee | 11:59.14 |
vtorri | sebras, when will the mupdf and gs devs be back ? | 12:02.48 |
robin_watts_mac | vtorri: monday. | 12:07.03 |
vtorri | thank you | 12:07.11 |
robin_watts_mac | but we'll be jetlagged then, so tuesday. | 12:07.16 |
vtorri | haha | 12:07.22 |
| i'll be busy anway | 12:07.31 |
robin_watts_mac | The Heisenberg Uncertainty Principle: will I be wearing the same T-shirt as Paul today? | 12:08.05 |
marcosw | http://bugs.ghostscript.com/buglist.cgi?regetlastlist=6270 | 13:30.45 |
e98 | I know I'm not the only one wondering why mupdf can't just also display images from the filesystem like it does from archive files | 13:54.17 |
| has there been any thought given to this before? | 13:54.26 |
robin_watts_mac | e98: It does. | 14:16.40 |
| I added code within the last month or so so it can open PNGs, JPEGs and TIFs. | 14:17.07 |
| What platform are you on? | 14:17.23 |
e98 | robin_watts_mac: right now linux i386 | 14:18.10 |
robin_watts_mac | e98: I believe that should work. | 14:18.27 |
e98 | I was thinking I could just tell it | 14:18.27 |
| $ mupdf vacation/*.jpg | 14:18.37 |
robin_watts_mac | e98: What version of mupdf are you using? Built from git? | 14:18.54 |
e98 | I discussed this with the llpp developer and he said only if mupdf supports loading image files from the fs, will he add support | 14:19.12 |
| yeah git master | 14:19.17 |
robin_watts_mac | e98: Ah, we only open 1 file at a time. | 14:19.26 |
| so mupdf vaction/IMG0001.jpg | 14:19.39 |
e98 | 1.2-266-g40599eb | 14:19.43 |
| I'll try | 14:19.52 |
robin_watts_mac | We might be tempted to add the ability to open a directory of images: | 14:20.17 |
| mupdf vacation/ | 14:20.23 |
| but that's not there yet. | 14:20.30 |
| If you want to code that, we'd love to see the patch. | 14:20.50 |
e98 | llpp has a nice overview moder or two and is pretty fast, so I thought it should also view images directly | 14:20.51 |
tor8 | directory of images, or directory of any file type we can open? | 14:21.08 |
robin_watts_mac | directory of images. | 14:21.14 |
e98 | loading a single jpeg works | 14:21.21 |
robin_watts_mac | right. | 14:21.28 |
tor8 | we get questions about how to combine multiple pdf files into one document view | 14:21.47 |
robin_watts_mac | tor8: paulgardginer has tweaks to make the android file do that. | 14:22.34 |
tor8 | I'm thinking a directory/multi-document aggregation document type could do the job | 14:22.58 |
robin_watts_mac | One pf our android customers wanted to be able to display n 1page pdf files as 1 n-page pdf file. | 14:23.07 |
| tor8: yeah, that would work. | 14:23.14 |
| might be a nice way to do PDF portfolios too. | 14:23.22 |
tor8 | robin_watts_mac: oh yeah! | 14:23.30 |
paul_gardiner | The ReaderView class was originally developed within an app that displayed a directory of images | 14:26.52 |
| I still have it somewhere | 14:27.05 |
sebras | tor8: robin_watts_mac: would you extract the portfolios from the original pdf into memory chunks and parse the embedded pdfs from there? | 14:28.54 |
tor8 | sebras: probably. | 14:29.15 |
sebras | I have a number of portfolio samples should you want them. | 14:29.15 |
robin_watts_mac | I can't remember. | 14:29.36 |
| I know that PDF portfolios with foreign types have the file embedded in them. | 14:29.57 |
| I can't remember if we can access PDF files from within portfolios without extracting the whole things first. | 14:30.25 |
sebras | robin_watts_mac: I don't think you can. | 15:11.12 |
| not unless you have a stream which extracts the embedded pdf from within another object. | 15:11.37 |
| and then you'd have to juggle two distinctly separate xrefs... | 15:11.49 |
robin_watts_mac | juggling xrefs is not a problem. Seeking within a single stream is. | 15:42.29 |
tor8 | marcosw: git log --format=fuller | 16:47.32 |
marcosw | http://bugs.ghostscript.com/report.cgi?resolution=---&x_axis_field=bug_status&y_axis_field=assigned_to&width=600&height=250&action=wrap&format=table | 21:06.07 |
robin_watts_mac | mvrhel_laptop: Given that you're not running, did you have some git stuff you wanted to look at ? | 21:49.28 |
| or the smart motion on android? | 21:49.35 |
mvrhel_laptop | oh either or both would be great. | 21:49.44 |
robin_watts_mac | or do you want to just relax a bit before going out? | 21:49.48 |
mvrhel_laptop | where to do you want to meet | 21:49.51 |
robin_watts_mac | Back in the meeting room? | 21:49.59 |
mvrhel_laptop | sounds good | 21:50.03 |
| be there in 5 minutes | 21:50.11 |
robin_watts_mac | see you in 3 mins then :) | 21:50.13 |
mvrhel_laptop | or 3... | 21:50.17 |
robin_watts_mac | :) | 21:50.25 |
| Forward 1 day (to 2013/06/09)>>> | |