| <<<Back 1 day (to 2015/05/12) | 20150513 |
gliu | I want to use mupdf to do digital signature | 07:00.58 |
| but it seems could not support embeded font. so signed failed. | 07:01.40 |
sebras | gliu: I think the main mupdf developer will be online in an hour or two. | 07:02.13 |
kens | gliu there arent' any MuPDF developers here at the moment, and I don't understand your problem. Probably it would be best if you can come back in a couple of hours | 07:02.24 |
gliu | kens, sebras, thanks, I will come back after several hours | 07:03.25 |
kens | Huh, reboot for this month's patches.... | 07:17.49 |
Robin_Watts | gliu: I'm one of the MuPDF developers, and I don't understand your problem either. | 09:20.38 |
gliu | Robin: I want to do digital signature on IOS. when i sign with a certificate, it throw an exception "pdf device supports only base 14 fonts currently" | 09:43.19 |
Robin_Watts | gliu: oh. | 09:52.58 |
| That's one for paulgardiner I think. | 09:53.05 |
gliu | OK. I think it did not support embeded font now, do you know it there a plan to support it? | 09:56.22 |
paulgardiner | I can't remember the status of digital signatures under iOS. gliu: did you need to alter code to get that far? | 09:58.21 |
gliu | yes. I hope so. | 10:02.48 |
tor8 | paulgardiner: sounds like the annotation creation/editing tried to update a form with a non-base14-font? | 10:02.57 |
paulgardiner | Well yes. I'm wondering whether I pull the font from the form or hardwire it, and whether that choice is taken as far out as the platform code | 10:04.21 |
| Yep. In pdf_set_signature_appearance, there's a call to get_font_info. So we are allowing the form to determine the font to use for the appearance stream | 10:09.15 |
| ... which is the correct thing to do. | 10:09.24 |
| ... but wont work with the current state of the pdf device if that font isn't base 14 | 10:09.55 |
| I guess the either the pdf device needs enhancing (very difficult), or pdf_set_singnature_appearance needs some heuristics to choose a base 14 font that will do. | 10:11.21 |
tor8 | paulgardiner: could look at the font descriptor flags to pick between serif/sans-serif/monospaced | 10:11.57 |
| like we do when picking a substitute font | 10:12.02 |
paulgardiner | yeah, that would make sense | 10:12.19 |
tor8 | paulgardiner: pdf_load_system_font | 10:13.15 |
| you rou/h reuse the logic in there | 10:13.48 |
| could* | 10:15.20 |
gliu | i had try a way. In get_font_info I hardware a font and the signature could be done. But I need support chinese, so some text could not be shown correct. | 10:16.08 |
henrys | chrisl: I didn't catch all of the conversation yesterday did you feel a bug was closed inappropriately? | 13:53.05 |
chrisl | henrys: yes, Ray looked again, and has reopened it - I think he was confused by the report | 13:53.35 |
henrys | chrisl: okay great just didn't want it to slip through the cracks | 13:54.04 |
kens | Ray has reduced the file, its omse kind of transparency blending problem | 13:54.50 |
henrys | kens: ugh michael probably doesn't need another one of those ;-( | 13:55.41 |
kens | Well I thnk Ray is looking at it at the moment, I could be wrong though | 13:55.58 |
| Oddly it seemed to be cuased by a commit of Ray's but I don;t thnk we're very confident about that. | 13:56.23 |
chrisl | I was dubious, given the nature of Ray's change, but it's what the bisect showed, so.... | 13:56.51 |
henrys | the thing I wanted to bring up about language switching yesterday is the idea we are going to have a language "language switching" with mooscript right? PDF, XPS, PS etc. - and I wonder if that shouldn't be the overall architecture instead of this business with separate devices and memory. Feel free to ignore this for now just wanted to put it out there. | 13:58.23 |
kens | We don't have mooscript working. Or at least, not for high level devices | 13:59.00 |
chrisl | I don't understand that - if we ever get it working, mooscript would just be another language | 13:59.17 |
henrys | chrisl: I thought mooscript was going to use the same memory and device as PostScript in GS? no? | 14:00.23 |
kens | I don't thnk it would need to | 14:00.54 |
| What do you mean by 'device' though ? | 14:01.12 |
chrisl | No, that was only ever mentioned because that's how the current language switch (fails to) work | 14:01.12 |
| Besides, it's months since any of us have even looked at mooscript..... | 14:03.02 |
| Actually, we may have discussed sharing the device to allow complex device configuration in Postscript, but I think the way I suggested for language switch would be better anyway | 14:05.34 |
henrys | chrisl: can your ideas be included here: http://twiki.ghostscript.com/do/view/Ghostscript/RefactorForLanguages? | 14:13.59 |
chrisl | I suppose - I've been avoiding the wiki on principle, so far...... | 14:15.28 |
henrys | chrisl: send an email to tech and I'll see it gets in. | 14:15.55 |
| running everything under one instance where languages are really just "parsing modes" would save us a crapload of setup and teardown. | 14:18.05 |
chrisl | But fails utterly in the coping with low memory case | 14:18.48 |
henrys | I won't include that in the wiki because I assume it will be shot down. | 14:18.49 |
| chrisl: yes you lose that. | 14:19.14 |
chrisl | And that's pretty critical given the market this is aimed at | 14:19.31 |
henrys | tor8: do you officially start the sabatical around the meeting time? | 14:21.33 |
tor8 | henrys: june 1st | 14:22.22 |
henrys | tor8: okay send a mail to joann with dates and copy staff so everyone knows. | 14:23.37 |
rayjj | morning, all | 14:24.01 |
henrys | hi rayjj | 14:24.13 |
rayjj | I saw the comments about the bug. I assume that I will continue to 'own' that one, and only consult with mvrhel if I need to | 14:24.44 |
henrys | rayjj: if that's okay, yes. | 14:25.50 |
rayjj | interestingly, the 'cut down' file displays correctly with mupdf, while the original doesn't | 14:26.08 |
| I will open a separate bug for mupdf, ok? | 14:26.35 |
henrys | rayjj: sure | 14:26.50 |
rayjj | OK, bug 695985 is the mupdf bug with the 'a.pdf' file (note that the 'abc.ai' file from bug 695975 looks the same -- has the same problem) | 14:32.32 |
HD | Hello | 14:34.02 |
ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 14:34.02 |
rayjj | hi, HD | 14:34.20 |
HD | I have one quiry regarding mupdf | 14:34.37 |
rayjj | HD: you're in the right place | 14:34.51 |
HD | i wanted to edit pdf using mupdf | 14:34.52 |
| I can edit all the things excepting signature | 14:35.17 |
| Is there any bug in mupdf? | 14:35.34 |
| related to signature? | 14:35.50 |
| FYI: I am working on android lib | 14:36.43 |
Robin_Watts | The signature stuff is still a work in progress. What exactly are you seeing. | 14:36.48 |
HD | I am not getting callback when user tab on signature field | 14:37.22 |
| When in pdf,signature is added from any other app and then i am getting callback | 14:38.13 |
| Is there any update when signature stuff will be completed? :) | 14:41.42 |
Robin_Watts | HD: Not at this time. | 14:45.32 |
| This sounds like a question for paulgardiner | 14:45.42 |
HD | Ohk. | 14:46.03 |
paulgardiner | I'm not understanding exactly what "callback", "tab" and "any other app" means in this context | 14:47.54 |
henrys | bbiab | 14:47.58 |
paulgardiner | The Android version of MuPDF should display a signing dialog if the user clicks on an unsigned signature within a form. Is it that that isn't working for you? | 14:48.56 |
HD | Yes, it is not working for me. | 14:49.29 |
Robin_Watts | HD: Have you done an SSL_BUILD ? | 14:50.09 |
HD | I have also tried MuPDF sample from google play also, but it also not working for signature field | 14:50.29 |
paulgardiner | Robin_Watts: Oh yes, that's a good point | 14:50.31 |
| I guess it must be specific to that file, unless we have a regression. | 14:51.13 |
Robin_Watts | HD: Can you visit bugs.ghostscript.com, make a bug report and attach the file you are using please? | 14:51.20 |
HD | Robin : Sure | 14:52.08 |
| Hi Robin, | 15:04.07 |
| It is taking to much time to submit | 15:04.18 |
Robin_Watts | How large is the file? | 15:04.39 |
HD | HD: Have you done an SSL_BUILD ? > Do i need to build any other lib to get it work | 15:05.36 |
| only 64kb | 15:05.57 |
Robin_Watts | I don't see why it would take a long time to submit then. | 15:06.15 |
kens | I see the report email, it saysit has an attachment | 15:06.32 |
| 64.17Kb | 15:06.40 |
Robin_Watts | When you do the ndk-build bit, you need to specify SSL_BUILD=1 | 15:06.51 |
kens | HD Robin_Watts THe file is attached to the report, it seems OK and I can download it | 15:07.29 |
HD | yes.. i may be of mine.. | 15:07.32 |
| so please, can you let me know is it working on your side? | 15:08.18 |
kens | One for Paul to look at I thnk, I don;t see any signature field on the form, but I know very little about PDF Forms | 15:08.45 |
HD | it has on field whose caption is Signature it self. | 15:09.25 |
kens | I don't see a field with the caption 'Signature' | 15:09.42 |
| I'm using Adobe Acrobat 9 | 15:09.53 |
HD | opps sorry, In hurry i have attached wrong file. | 15:10.00 |
| please refer updated pdf | 15:15.43 |
kens | 89Kb now ? | 15:16.05 |
HD | yes | 15:16.09 |
kens | OK well one for Paul, I do now see a signature field in Acrobat. | 15:16.48 |
| But I don't have an Android device to try MuPDF on | 15:17.00 |
HD | Now its time to leave. Will meet you tomorrow. | 15:18.19 |
paulgardiner | Maybe one for Henry to comment on. I think it might take a while to trace and I'm not on MuPDF just now. | 15:18.26 |
HD | Good night all!! | 15:18.30 |
| It will be helpful, if this issue will be resolve. | 15:18.57 |
paulgardiner | I guess I can take a quick look at the PDF to see if it has what I expect in the field | 15:19.54 |
kens | Seems fair to me | 15:20.05 |
| Warning, its quite a complicated file | 15:20.14 |
HD | Means, you are getting this warning in android ? | 15:22.08 |
paulgardiner | What warning? | 15:23.41 |
| It seems to work for me. Using the android version of MuPDF, if I tap on the signature, a dialog appears offering to "Select a certificate and sign" | 15:46.09 |
| HD^ | 15:46.16 |
mvrhel_laptop | oh jung is not using the projects that for libmupdf and libthirdparty that come in win32 but the old ones that were in the windows directory | 15:46.17 |
kens | paulgardiner : well that seems conclusive to me.... | 15:46.41 |
mvrhel_laptop | that is not going to work so well for him | 15:46.43 |
Robin_Watts | mvrhel_laptop: I got a test solution from him working at 1am yesterday. | 15:51.12 |
| With memento running it doesn't leak. | 15:51.20 |
| At least, the C level does not leak. | 15:51.27 |
| The C++ level looks plausible, but memento doesn't test that. | 15:51.41 |
mvrhel_laptop | Robin_Watts: right. C# should not leak. now the c++ stuff could be an issue | 15:51.42 |
Robin_Watts | I'd like to try to extend memento to capture new/delete too. | 15:51.57 |
mvrhel_laptop | If he is doing any new and no delete | 15:52.01 |
| yes | 15:52.03 |
Robin_Watts | so I'm going to give that a whirl, but I'm being distracted by trying to get windows to stop being an ass on my new laptop. | 15:52.30 |
mvrhel_laptop | I am writing him an email explaining how to manage his code..... | 15:52.47 |
| He has a Frankenstein here | 15:53.06 |
henrys | chrisl: is the ufst version in svn private the latest delivery we have from Monotype? | 15:53.19 |
Robin_Watts | I've installed an M2 SSD in the laptop alongside the original system drive. Will any of MSs/lenovos tools let me move windows onto the SSD? Hell no... | 15:53.26 |
chrisl | henrys: erm, possibly... the one in ufst-6.3 is the latest 6.x we got | 15:54.23 |
henrys | the svn is version 5 | 15:55.51 |
chrisl | henrys: sorry, the one in ufst-6.x | 15:55.57 |
| There is a separate directory in svn for 6.x | 15:56.19 |
| henrys: actually, I have a 7.x release somewhere, but I never did much with it.... | 15:57.28 |
henrys | chrisl: ah I'm looking at URL: svn+ssh://svn.ghostscript.com/var/lib/svn-private/ghostpcl/trunk/ufst | 15:57.28 |
| chrisl: could you throw that up on casper. | 15:57.42 |
| ? | 15:57.45 |
chrisl | Which? 7? | 15:57.53 |
henrys | yes | 15:57.57 |
chrisl | I'll have to find it..... but yes | 15:58.16 |
| henrys: it's going to take a while.... it's a 359Mb tar.gz archive | 16:04.31 |
| henrys: in my home directory on casper - ufst7.1.1.tar.gz | 16:09.08 |
henrys | thanks chrisl | 16:26.43 |
mvrhel_laptop | hmm the display device crashes for me with an x64 build | 16:33.53 |
| this is with VS2013 and windows 8.1 | 16:34.04 |
| win32 is fine | 16:34.10 |
| I was taking a look at that 256 color display bug | 16:34.25 |
| to see if I can get it to fail on windoze | 16:34.31 |
chrisl | mvrhel_laptop: seems to be okay on VS2005/Vista x64 | 16:48.00 |
mvrhel_laptop | ok. I may poke around at it a bit if I get a chance | 16:48.26 |
chrisl | It might be good to try the VS2013 exe on Win7 and a VS2005 exe on Win8.1 | 16:49.13 |
rayjj | hmm... I installed VS 2015 RC Community and it won't build (either PCL or gs) I get two errors when it tries to compile gp_mswin.c: Error c2371 'gp_fseekable': redefinition; different basic typesline 956 -and- | 16:50.59 |
| Error C2084 function 'int snprintf(char *const ,const size_t,const char *const ,...)' already has a body line 976 | 16:51.00 |
| (I was going to have a look at the performance tools) | 16:51.41 |
| I'll go back to real work for the meantime (in case chrisl decides to take a look at it -- if he doesn't then I'll dig into the typedef morass) | 16:53.05 |
chrisl | rayjj: if you can open a bug for it, I'll get to it next time I feel like firing up Windows.... I assume I can use VS2015 on Windows 7? | 16:53.58 |
henrys | okay customer bugs in single digits - you have to fix one before adding another ;-) | 17:06.11 |
mvrhel_laptop | henrys: so there is no way to do a 256 color display set up for me on windows with the display device for bug 695074. who should I pass this thing to? | 17:07.57 |
henrys | hin-tak ;-) | 17:09.07 |
mvrhel_laptop | ha | 17:09.11 |
henrys | pass it to me I want to fool with it. | 17:09.51 |
mvrhel_laptop | ok | 17:09.56 |
henrys | mvrhel_laptop: I assume his regression information didn't help much ... | 17:10.24 |
mvrhel_laptop | no | 17:10.30 |
| Its possible there is something of use in the first comment | 17:11.01 |
| But again, I would need to see what is going on | 17:11.13 |
| Let me see if I can get rid of 695845 now | 17:12.40 |
rayjj | mvrhel_laptop: henrys: there is a 256 color bmp256 device, but the x11 device may do some funky stuff differently (as hin-tak discovered re the 'dynamic color' nonsense) | 17:57.33 |
mvrhel_laptop | rayjj: there is a png256 device too | 17:57.49 |
| that worked fine | 17:57.52 |
rayjj | mvrhel_laptop: right, and also a pcx format one. They all use the same color methods, afaict "pc_8bit_map_rgb_color" | 17:59.01 |
mvrhel_laptop | yes | 17:59.06 |
henrys | rayjj, mvrhel_laptop does anyone understand hin-tak's comment #2 "here's how you set it up but you won't be able to see it"? What does that mean? | 18:01.06 |
mvrhel_laptop | I have no idea | 18:01.16 |
| I actually tried to do that | 18:01.25 |
| and gave up to see who I could hand this off to | 18:01.34 |
| rayjj did you ever hear back from customer 531 about the halftone threshold issue | 18:17.14 |
| bug 695929 is still sitting on my list as a P1 | 18:17.50 |
| bbiab | 18:19.22 |
Robin_Watts | mvrhel_laptop: OK, according to memento, the C++ layer is leaking 2 blocks, totaling just over 8k. Let me try and see why. | 18:24.59 |
rayjj | 8 whole Kb. ;-) | 18:27.56 |
Robin_Watts | rayjj: Yeah. | 18:28.27 |
rayjj | mvrhel_laptop: no, I didn't. I had given Jasper the info on gen_ordered and didn't hear back. Maybe Marcos or I should ping him | 18:30.01 |
henrys | mvrhel_laptop: looks like hin-tak is back on it. | 18:30.13 |
Robin_Watts | Looks like it's gTextResults? | 18:30.18 |
rayjj | mvrhel_laptop: frankly the gen_ordered halftoning is far superior to the gs defaults even with 'sethiresscreen' | 18:30.58 |
Robin_Watts | mvrhel_laptop: You don't delete mu_ctx? | 18:31.23 |
| yeah, so 2 problems. | 18:32.12 |
| Not deleting mu_ctx costs 8K. | 18:32.23 |
| and the gTextResults thing shows up for me, cos it's a static. Is that safe? | 18:32.39 |
rayjj | mvrhel_laptop: I just 'took' that bug, and added the comment with the text that I sent to the customer | 18:34.09 |
| mvrhel_laptop: I also added the halftone screen I sent him that was generated for 90 lpi at 1680 dpi (that gives the strange results) | 18:37.02 |
| henrys: should I leave the follow up to Marcos, or should I ping Jasper ? | 18:37.40 |
mvrhel_laptop | Robin_Watts: sorry I was out for a bit | 19:04.23 |
| hold on | 19:04.44 |
| arg. yes the missing delete on mu_ctx | 19:07.11 |
| ReleaseTextSearch should take care of gTextResults but perhaps I made a mistake there. | 19:10.23 |
| Robin_Watts: I need to go get my daughter at school. | 19:10.34 |
| I will review this more closely when I get back. How hard is it for me to add memento to my solution? | 19:11.07 |
| oh actually its already in there | 19:11.27 |
| but how did you add the new/delete stuff? | 19:11.37 |
Robin_Watts | mvrhel_laptop: The vector itself is not freed, just emptied. | 19:42.52 |
| So it shows up as 8 bytes (in a 64bit build) | 19:43.00 |
| The only worry I have is that should it maybe be part of the context rather than being a static? | 19:43.33 |
| mvrhel_laptop: I have a tweak to Memento here that adds new/delete. | 19:44.25 |
| but not new[] delete[] yet. | 19:44.38 |
mvrhel_laptop | Robin_Watts: yes putting those in the context makes a lot of sense | 21:01.36 |
| let me work on getting that all fixed up | 21:01.41 |
henrys | mvrhel_laptop: yes I can reproduce the problem, but Hin-Tak seems to be doing it so let's him go for a while and see what happens. | 21:41.01 |
mvrhel_laptop | henrys; ok sounds good | 21:41.17 |
| Robin_Watts: I will take a look at the memory issues now if you want | 21:44.55 |
| Since it is my mess | 21:45.17 |
| windows code analysis in vs2013 actually found a couple useful issues | 21:56.30 |
| hmm looks like it is recommended that I change a few things in how I have my calls to the native code structured. will address this a bit later | 22:05.10 |
| Forward 1 day (to 2015/05/14)>>> | |