| <<<Back 1 day (to 2014/05/05) | 2014/05/06 |
jo0nas | chrisl_away: bug files about system shared libtrio: 695211 | 01:37.24 |
kens | ping chrisl (I know you're there really, you can't hide ;-) | 08:09.55 |
chrisl | kens: pong (not sure why I'm still "away")..... | 08:10.22 |
| .away | 08:10.24 |
kens | :-) | 08:10.27 |
| chris what's the status of the mooscript code ? If I check it out will it contain the change sI made for images ? | 08:10.52 |
| And has anything else changed ? | 08:11.06 |
chrisl | Erm, let me check.... | 08:11.54 |
kens | THere's no rush, I just remembered I need to go and try to investigate GSView | 08:14.10 |
chrisl | kens: so, if you use the repo in ghostscript.com:/home/chrisl/repos-priv/ghostpdl-moo.git and the branch mooscript, the last commit is your one with page orientation and images | 08:15.54 |
kens | OK that's pretty much what I trhought, so I should be OK with what I have checked out here, thanks | 08:16.20 |
| Hmm that's a long list of makefile issues | 08:17.28 |
chrisl | And all spurious..... | 08:17.39 |
kens | :-( | 08:17.45 |
chrisl | It's a load of warnings about things nowhere near the changes I made | 08:18.10 |
kens | I thought that was probably the case | 08:18.23 |
chrisl | kens: I *believe* you should be able to commit to the above repo now, too | 08:22.51 |
kens | Oh even better, thanks | 08:22.58 |
| I'll have a go after I either find out what's wrong with GSView or give up in disgust | 08:23.31 |
chrisl | Interesting you count that as an either/or situation - it seems to me that you might find out what's wrong *and* give up in disgust..... | 08:24.44 |
kens | Well possibly. At the moment I'm struggling to see how to make a debug build, and debug it | 08:25.06 |
| OK looks like I may have a debug executable now | 08:28.44 |
chrisl | reboot | 08:29.09 |
VicVii | Hi | 09:11.43 |
ghostbot | privet, vicvii | 09:11.43 |
VicVii | privet | 09:13.02 |
| lol. Q about mupdf for android. Is it possible to retrieve the page's number. This pdf has a section in roman numerals and then a section in arabic numbers. | 09:15.40 |
| I need the roman numeral for the roman section, and the arabic number for the arabic section. | 09:16.00 |
Robin_Watts | VicVii: There is currently no simple method to get that information. | 09:16.52 |
VicVii | Is there a complicated way? If I can get PageLabelNewIndex I can work with it. | 09:25.54 |
| Is there a way for me to get access to the PDF Root Document object on Android with mupdf? | 10:38.03 |
Robin_Watts | VicVii: hi | 10:42.14 |
| MuPDF is a portable C library with an API that includes ways to get the PDF Root Document object. | 10:42.36 |
| We then wrap this in various example tools, including viewers for various different platforms. | 10:42.53 |
| The android viewer is one. | 10:42.58 |
| The frontend is written in java, and it has to have an interface to call down to the C. | 10:43.23 |
| That interface is not fully general - it only supports the bits we needed for the android front end. | 10:43.41 |
| You can see it in platform/android/jni/mupdf.c | 10:44.13 |
| If you extend that, you can probably do what you want., | 10:44.22 |
VicVii | thx | 10:58.48 |
Robin_Watts | VicVii: If you extend the API nicely to do it, then let us know. | 11:04.07 |
VicVii | Will do | 11:04.24 |
tor8 | Robin_Watts: new fix for the bug we discussed yesterday on tor/master | 11:09.07 |
Robin_Watts | tor8: lgtm. | 11:24.09 |
tor8 | Robin_Watts: also, the SLOWCMYK transform from poppler seems to have trouble coping with the file from bug 694801 | 11:26.12 |
Robin_Watts | over/underflows or something? | 11:26.42 |
tor8 | it seems to map both 0,0,0,1 and 0,0,0,0 to black... | 11:26.59 |
| at least if I look at the raw image data and then what comes out of the pixmap transform | 11:27.20 |
| haven't isolated which of the SLOWCMYK transform functions are affected though | 11:27.49 |
| oh, it's fast_cmyk_to_rgb | 11:28.47 |
| so yeah, I'd expect some over/underflow in the integer math | 11:30.03 |
| Robin_Watts: so I've assigned it to you, hope you don't mind | 11:32.20 |
Robin_Watts | ok. | 11:32.28 |
tor8 | it's your unrolled optimized code, I'm not willing to stick my head into that if I can avoid it :) | 11:32.49 |
Robin_Watts | hehe, sure. If I get this layout thing figured out, I'll have a look later. | 11:33.51 |
arjun_ | hello | 11:45.54 |
ghostbot | bonjour | 11:45.54 |
arjun_ | this is arjun heere | 11:46.29 |
Robin_Watts | hello arjun. | 12:05.01 |
kens | can someone remond me of the git submodule magic incantation please ? | 12:11.01 |
| Hmm git submodule update leaves me with an empty folder under thirdparty called mujs, I@m guessing there should be something in there..... | 12:14.56 |
| ah, not initialised | 12:15.40 |
| OK looks better now | 12:16.56 |
| And MuPDF builds, excellent | 12:17.29 |
paulgardiner_lap | tor8: I'm just submitting the iOS app to the App store. I'm being asked for a large app icon 1024x1024. Got anything I can use? Do we have the icon in vector graphics form? | 13:06.37 |
Robin_Watts | paulgardiner_lap: I have a vector version I think, but Tor has the original. | 13:08.06 |
paulgardiner_lap | I expect that'll do nicey. Can you bung it in an email? | 13:09.30 |
Robin_Watts | You want a 1024x1024 PNG ? | 13:10.45 |
| sent | 13:12.33 |
paulgardiner_lap | ta | 13:13.04 |
henrys | chrisl: we never really got clear permission from URW to go AGPL but I'd imagine they'd be fine with it. | 13:18.03 |
chrisl | henrys: as far as the fonts are concerned, there is no difference between the GPL and AGPL, because of the exemption paragraph | 13:20.20 |
henrys | chrisl: yeah still I have mixed feelings if I should check in with them about it. | 13:22.21 |
tor8 | paulgardiner_lap: logo source files are on a branch in tor/logo | 13:25.07 |
chrisl | henrys: I think you should check with them - bare in mind that we can still "distribute" the fonts as GPL separately if URW prefer that. | 13:26.02 |
henrys | chrisl: agreed | 13:31.53 |
Robin_Watts | "bare in mind" = "imagine me naked" ? | 13:34.42 |
| :) | 13:34.48 |
chrisl | Sorry, "bear in mind"..... | 13:37.31 |
henrys | chrisl starts running | 13:38.04 |
paulgardiner_lap | Ew! During attempted submission of iOS app: "Have you added or made changes to encryption features since your last submission of this app?" I'd imagine including openssl counts as Yes | 13:38.23 |
Robin_Watts | http://www.tikihumor.com/wp-content/uploads/sites/37/2013/09/bear-runner.jpg | 13:38.51 |
henrys | Robin_Watts: only 2 had the sense to leave the equipment | 13:39.44 |
Robin_Watts | This is the one I was looking for | 13:41.24 |
| http://media-cache-ak0.pinimg.com/236x/d1/bb/45/d1bb453da254a85d845193667ca0667f.jpg | 13:41.25 |
kens2 | Hmm, as I understand it, trying to outrun a bear is not likely to end well | 13:42.05 |
| Though standing still might not be much better either | 13:42.17 |
Robin_Watts | You don't need to outrun the bear, you just need to outrun your slowest companion. | 13:42.35 |
kens2 | works for the 1st picture, not so well for the 2nd | 13:42.47 |
kens2 | has a T-shirt like tha anyway | 13:42.59 |
Robin_Watts | It's a similar situation to the motto for the helicopter machine gunners in vietnam. "Don't run, you'll just die sweaty" | 13:44.03 |
henrys | I don't think the slowest guy works - the bear is going to catch whoever he gets a bead on first. At 30 mph that gap is going shrink so fast the running difference isn't going to matter. play dead for a grizzly | 14:00.24 |
ray_laptop | Robin_Watts or mvrhel_laptop: can one of you sanity check my patch (regression test OK). The bmpcmp shows a progression, but I can't reproduce it. http://git.ghostscript.com/?p=user/ray/ghostpdl.git;a=commitdiff;h=e459fd958d2165a2b13517b6f1ca8c6b171d20de | 14:01.10 |
paulgardiner_lap | I though it was stand tall with your arms in the air and shout a lot, but I may have been told that by someone who was concerned about starving bears and didn't like me a lot. | 14:03.23 |
Robin_Watts | ray_laptop: Line 1617 too ? | 14:03.41 |
kens2 | I htink the right answer is to be carrying a pepper spray | 14:03.48 |
Robin_Watts | paulgardiner_lap: That's what they told us at katmai. "Look big". | 14:04.03 |
| But then maybe that's because tourists are more expendable than rangers. | 14:04.21 |
tor8 | play dead until it starts eating you, then scream :) | 14:04.44 |
Robin_Watts | then playing is unnecessary. | 14:04.56 |
| ray_laptop: Otherwise it looks plausible to me. | 14:07.59 |
ray_laptop | Robin_Watts: actually, I _could_ do the same thing with line 1443. I had looked at line 1532 and determined that it was not needed because the closepath only does the line if open > 0. I'll go ahead and fix 1443 and 1617. | 14:09.57 |
| Robin_Watts: thanks for spotting that. I guess I was just to eager to get it done. | 14:10.19 |
Robin_Watts | np. | 14:10.25 |
ray_laptop | that's why code review works for me (usually) | 14:10.33 |
Robin_Watts | It'd be another 10 years before you found an example though :) | 14:10.42 |
henrys | Robin_Watts: sigh the USE_RUN_ROP didn't help much at all. | 14:11.51 |
Robin_Watts | henrys: What's the average 'width' ? | 14:12.04 |
| average is the wrong word. | 14:12.12 |
| What sort of width values are we seeing? | 14:12.21 |
| If width is small, then the setup costs hurt us. | 14:12.33 |
ray_laptop | Robin_Watts: actually, if I was motivated, I could construct examples that would fail without the change to 1443 and 1617 (now that I understand the root cause) | 14:12.45 |
Robin_Watts | If it's an unusual rop value, then we end up being almost as slow anyway. | 14:13.05 |
henrys | Robin_Watts: yeah I'm sure the pieces are too small there is well over 1 million calls to mem_gray8_rgb24_strip_copy_rop on just a few pages. | 14:13.19 |
Robin_Watts | but the win comes in that we can do nicely optimised rop routines for common ones. | 14:13.42 |
| at 300dpi ? | 14:13.49 |
henrys | 600 | 14:13.57 |
Robin_Watts | Well, that's 40 Million pixels a page. | 14:14.29 |
henrys | it might be significant if the xor was inlined. | 14:14.30 |
Robin_Watts | We can never get it inlined. | 14:14.42 |
| but we can ensure that we just do 1 function call per run. | 14:15.01 |
| What rop number? | 14:15.08 |
henrys | we could have a separate xor procedure is what I mean | 14:15.18 |
Robin_Watts | We have lots of separate procedures. | 14:15.41 |
| See rop_get_run_op in gsroprun.c | 14:15.57 |
ray_laptop | henrys: BTW, I have to leave today by 8:15 (in an hour) so anything for me should be early in the meeting | 14:16.12 |
henrys | it calls xor_rop_run24 | 14:16.20 |
Robin_Watts | Urm... I don't have that function. | 14:17.17 |
| not xor_rop_run24_const_st ? | 14:17.35 |
henrys | Robin_Watts: yes I abbreviated it | 14:17.52 |
Robin_Watts | ah, well, xor_rop_run24 would be a different beast :) | 14:18.07 |
henrys | ray_laptop: okay | 14:18.26 |
| ray_laptop: I don't think I had anything in particular for you, if you have something bring it up first | 14:19.08 |
Robin_Watts | henrys: We *could* do a more optimised version of that that worked in 64bit chunks, if width was sufficiently large often enough to justify it. | 14:19.18 |
| henrys: If you enable RECORD_ROP_USAGE in gsroprun.c, rebuild and then rerun the file, you'll get a list of the rops used printed. | 14:20.30 |
tor8 | henrys: bug 695112 ... how much do we want to please Raed? fixing that one would necessitate writing our own jpeg or finding another library, more amenable to error handling and correction than libjpeg | 14:21.02 |
| I'm inclined to just WONTFIX it | 14:21.13 |
Robin_Watts | tor8: technically, we can change the jpeg headers to have a max of 65535. | 14:21.46 |
tor8 | Robin_Watts: yeah, but that won't render the right data | 14:22.00 |
Robin_Watts | They only leave it set to 65500 to allow for silly people with 16 bit ints, IIRC. | 14:22.15 |
tor8 | Robin_Watts: the actual image dimensions are in the PDF (2000x3000 or something small like that) | 14:22.18 |
henrys | tor8: agreed but report it to the jpeg folks. | 14:22.53 |
tor8 | but by the time we have read the JPEG header and are in a position to fix it, libjpeg has already put itself into an unrecoverable error state | 14:22.53 |
Robin_Watts | tor8: Is it just the height that is wrong ? | 14:22.54 |
| There is a little used JPEG marker block that means "I lied about the height of the jpeg, it's actually Y". | 14:23.27 |
henrys | 5 minute until the meeting | 14:23.50 |
Robin_Watts | The idea being that you give the height in the header as 0 (or 65535) and then the encoder can put out the marker later. | 14:23.52 |
kens2 | thinks this sounds like another stupid idea form the people that brought us JPEG2000..... | 14:24.50 |
tor8 | Robin_Watts: it is only the height that is wrong in that sample file | 14:25.37 |
Robin_Watts | Then I bet it's got a header of 65535 and that marker at the end. | 14:25.57 |
henrys | Robin_Watts: RECORD_ROP_USAGE is going to give me lengths? | 14:26.55 |
Robin_Watts | It tells us the total number of pixels, IIRC. | 14:27.05 |
henrys | well the old code tells me that since it call the ROP function per pixel. | 14:27.50 |
| from the profile call count that is | 14:28.05 |
Robin_Watts | henrys: Well, 'record_run' is called with 'len' = width. | 14:29.01 |
tor8 | Robin_Watts: if I hack the header size (instead of throwing the error) it decodes fine, but NONE of the image viewers I have tried work on the extracted jpeg image | 14:29.34 |
| Robin_Watts: do you know what the marker is called? | 14:29.42 |
Robin_Watts | You can tweak that to maybe store the 'top 10' widths or something? | 14:29.45 |
| tor8: Looking... | 14:29.54 |
| It's the last 6 bytes or so in the file, IIRC :) | 14:30.12 |
henrys | kens2: I've added the gsprint stuff to the workflowy | 14:31.05 |
kens2 | thanks henry | 14:31.12 |
henrys | kens2: I'm surprised though I wouldn't think print driver folks would put up with that dependency. | 14:32.06 |
kens2 | henrys, its not the print driver, is the spool queue | 14:32.23 |
Robin_Watts | FFDC DNL Define Number Of LInes | 14:33.13 |
henrys | kens2: but you said print ticket wouldn't work without net | 14:33.26 |
kens2 | henrys, not exactly.... | 14:33.34 |
| I can't get the *default* print ticket for he queue (or at least I can't find any way to) wihtout .NET | 14:33.54 |
| But that's all back at teh submission stage, before we go anywhere near a printer driver | 14:34.08 |
Robin_Watts | tor8: It's mentioned in the text/diagram in B.2.1 | 14:34.17 |
henrys | kens2, mvrhel_laptop what about mobile? (windows) | 14:35.25 |
tor8 | Robin_Watts: B.2.1 of which document? | 14:35.27 |
kens2 | henrys I have no idea about mobile and printing | 14:35.40 |
Robin_Watts | itu-11505_t81e.pdf | 14:35.41 |
henrys | ray_laptop is leaving early - anybody have stuff for him? | 14:35.56 |
kens2 | Not me | 14:36.14 |
henrys | kens2: ugh I hate saddling folks with .NET | 14:37.36 |
kens2 | I'd prefer not to go anywhere near it. Our current solution will 'work' for as long as MS supports the 'unsupported' API call. I can probably extend the feature set considerably using some more blakc magic too. But I'm not convinced that long term this is the way to go. That's really waht I'd like to talk about at the staff meeting | 14:38.36 |
henrys | kens2: okay | 14:39.26 |
tor8 | Robin_Watts: I don't have that one ... and it's a for-pay download from itu.int | 14:39.28 |
Robin_Watts | mails. | 14:39.34 |
mvrhel_laptop | sorry I am late | 14:39.42 |
tor8 | Robin_Watts: oh, found it on w3c | 14:39.47 |
kens2 | Robin_Watts beat me to it | 14:39.51 |
paulgardiner_lap | What's a nice easy editor to learn for someone editing html source, preferably with source highlighting? I don't want to suggest vim or emacs because of the learning curve. | 14:40.25 |
henrys | we now have gsview.com | 14:40.36 |
kens2 | Visual Studio ? | 14:40.38 |
tor8 | paulgardiner_lap: sublime text 2? | 14:40.45 |
mvrhel_laptop | henrys, kens2. mobile windows printing. I have no idea what is going on there. i could ask my one contact at MS though | 14:40.58 |
| henrys: oh great about gsview.com | 14:41.13 |
henrys | mvrhel_laptop: a little future guidance from an inside source would be good. | 14:41.41 |
kens2 | I'll poke through the docs and see if they say antyhing about mobile printing. Supposedly this is all supposed to be converging so it *ought* to be the same as 'App Store' apps, but I suspect it isn't | 14:41.51 |
chrisl | paulgardiner_lap: on Linux, I use nedit - little clunky, but usable | 14:41.55 |
paulgardiner_lap | I hadn't considered VS because it does so much more, but yeah that might work. I'll take a look at sublime too | 14:41.58 |
ray_laptop | that gxclpath.c bug has been in there a LONG time. There was a fix to prevent some spurious lines in rev 3.68 (5/9/96) but it's been in there since 3.60 (11/20/95). Boggles the mind that we just tripped over it | 14:42.09 |
Robin_Watts | brackets? | 14:42.12 |
kens2 | paulgardiner_lap : its has good syntax colouring for HTML | 14:42.20 |
mvrhel_laptop | they have made windows printing very easy for windows 8 apps. but not sure about the mobile stuff | 14:42.22 |
henrys | ray_laptop: yeah I couldn't believe that. | 14:42.40 |
Robin_Watts | paulgardiner_lap: brackets is an adobe open source editor thing designed specifically for webby stuff, IIRC. | 14:42.45 |
paulgardiner_lap | kens2: Yeah, other than the sledgehammer/nut aspect, it does fit nicely | 14:43.02 |
Robin_Watts | and emacs on windows can be driven entirely from menus etc. | 14:43.25 |
kens2 | paulgardiner_lap : it depends what you already have and know. Seems like everyone has a good suggestion | 14:43.25 |
paulgardiner_lap | In what way is nedit clunky? | 14:43.46 |
chrisl | It's based on Motif | 14:44.01 |
Robin_Watts | Ultraedit used to be reckoned to be good. There is a copy of that on /mnt/picsel somewhere. | 14:44.11 |
paulgardiner_lap | This is for someone used to notepad | 14:44.26 |
kens2 | running on Linux or WIndows or Mac ? | 14:44.52 |
paulgardiner_lap | Windows | 14:44.57 |
kens2 | well VS or robin's suggestion of brackets sound good | 14:45.13 |
Robin_Watts | http://brackets.io/ | 14:45.15 |
chrisl | Robin_Watts: do you get a Windows download there? | 14:45.35 |
tor8 | Robin_Watts: doesn't brackets run in the browser? | 14:45.38 |
kens2 | the front page makes it look nice | 14:45.38 |
Robin_Watts | chrisl: I do. | 14:45.43 |
tor8 | doesn't trust javascript apps. | 14:45.51 |
henrys | paulgardiner_lap: aren't there free easy to use cloud editors for html? | 14:45.57 |
chrisl | Robin_Watts: Ah, okay, cause I couldn't find one | 14:46.13 |
henrys | paulgardiner_lap: cloud9? | 14:46.42 |
| I didn't have much for this meeting - anyone have any updates for the agenda? | 14:47.44 |
mvrhel_laptop | paulgardiner_lap: I like winedt, which I use for latex. It does highlighting for just about everything even postscript | 14:47.54 |
paulgardiner_lap | henrys: I don't think I'm finding the correct link for cloud 9 | 14:48.05 |
chrisl | henrys: have we had an eta from URW on the fonts? | 14:48.14 |
henrys | c9.io | 14:48.19 |
paulgardiner_lap | mvrhel_laptop: Thanks. Another I'll try out | 14:48.30 |
henrys | https://c9.io/ | 14:48.45 |
paulgardiner_lap | You don't have to pay for it, do you? Heaven forbid! :-) | 14:49.18 |
kens2 | says its free | 14:49.31 |
henrys | https://c9.io/site/pricing/ | 14:49.38 |
kens2 | in italics under the red sign me up button | 14:49.42 |
paulgardiner_lap | Phew! | 14:49.47 |
kens2 | cursive script, not italics <blush> | 14:50.17 |
henrys | paulgardiner_lap: there are lots of them but I'm afraid I don't have a first hand recommendation | 14:51.28 |
marcosw | is there an update on the release date for GhostDocs? Miles wants the comparison of old vs. new vs. word. | 14:51.36 |
paulgardiner_lap | I think I should be able to find at least on good one from all the suggestions. | 14:52.02 |
henrys | GhostDocs kicks off at 8 but we can start it now if everyone is finished. | 14:52.21 |
kens2 | I'm done | 14:52.31 |
chrisl | In that case I can go for the coffee I didn't get before the meeting.... :-) | 14:53.12 |
mvrhel_laptop | I'm done. I have a question later for Robin_Watts. gsview is working nicely now. I fixed a bunch of clunky zooming, navigation issues this past week | 14:53.24 |
henrys | actually I'd like to take the 5 minutes for coffee making | 14:53.29 |
mvrhel_laptop | sounds good | 14:53.39 |
Robin_Watts | : fetches more caffiene too. | 14:53.58 |
mvrhel_laptop | Robin_Watts: so is there a way to get the version number from a dll build of mupdf? gs has this. I was going to use it in populating the about gsview information | 14:54.32 |
| henrys: and what version of gsview are we going to consider this? | 14:54.56 |
kens2 | At least 6.0 | 14:55.21 |
| mvrhel_laptop : ^^ | 14:55.26 |
Robin_Watts | mvrhel_laptop: Urm... a mo. | 14:55.55 |
mvrhel_laptop | Robin_Watts: I guess the question was asked oddly | 14:56.19 |
| is there any API call to get the version number | 14:56.29 |
Robin_Watts | FZ_VERSION is a #define. | 14:57.00 |
mvrhel_laptop | ok great. that will work | 14:57.27 |
tor8 | mvrhel_laptop: Robin_Watts: fz_new_context will check the FZ_VERSION and refuse to cooperate if the library and client headers mismatch | 14:57.55 |
mvrhel_laptop | very good | 14:58.11 |
henrys | mvrhel_laptop: 6.0 right? | 15:00.30 |
Robin_Watts | Any particular reason we shouldn't call it 9.14 ? | 15:01.07 |
henrys | this was the last from russell right? http://pages.cs.wisc.edu/~ghost/gsview/get50.htm | 15:01.23 |
kens2 | 5.0 is the current GSView release, yes | 15:01.36 |
chrisl | Given that's it's mostly mupdf rather than gs, tying it to the gs versions would seem misleading | 15:01.56 |
Robin_Watts | Well, I'd have suggested 1.4, but 1.4 < 5 :) | 15:02.25 |
henrys | I'd prefer continue on with gsview numbers but I don't feel strongly about it. | 15:02.37 |
Robin_Watts | Our plan is to release it every 6 months in lockstep with gs? | 15:02.47 |
chrisl | And mupdf | 15:03.01 |
Robin_Watts | It's simpler for people (and by that I include myself) to have to track a single number. | 15:03.23 |
tor8 | Robin_Watts: drop the decimal to get 14 ;) | 15:03.35 |
Robin_Watts | and the 'branding' on gsview is 'gs'. | 15:03.36 |
chrisl | Well, again, that's misleading | 15:03.58 |
Robin_Watts | muview! :) | 15:04.06 |
tor8 | annual releases and use the date of release, seems to work for a lot of things | 15:04.11 |
henrys | mvrhel_laptop: you decide ;-) | 15:04.18 |
Robin_Watts | tor8: So people will release "gsview 2014" in November 2013? :) | 15:04.48 |
tor8 | Robin_Watts: yup! :) | 15:05.04 |
| or 2012... if they're feeling sneaky | 15:05.18 |
kens2 | personally I'd prefer to call it 6.0, but as long as the version number doesn't go *backward* its fine | 15:05.39 |
jo0nas | how about calling it "gsview XP"? | 15:05.50 |
kens2 | or NT (New Technology) | 15:06.03 |
chrisl | gsview TNG | 15:06.06 |
Robin_Watts | chrisl: make it so. | 15:06.21 |
tor8 | I'd also vote for 6.0 (or gsview .NET) | 15:06.36 |
kens2 | Hmm GS 9 ? | 15:06.41 |
chrisl | :-) Or "gsview ReBoot" | 15:06.53 |
Robin_Watts | So, the first 5 minutes of the GhostDocs meeting has gone swimmingly :) | 15:07.32 |
tor8 | chrisl: J.J. Abrams' GsView the Next Reboot | 15:07.37 |
chrisl | Can't we just wait until mupdf gets to 6.x and then match it to mupdf? 10-12 years should see it ready! | 15:07.55 |
henrys | look like pedro__ is roadblocked with good | 15:08.00 |
| we have a new web person whom I haven't spoken with yet. | 15:08.54 |
pedro__ | yeah, it seems like diminishing returns with the authentication stuff - I've spent about 2-3 days trying various approaches and trawling the forums for similar problems/solutions so it may be more effective to let them come back to us and I could get back to some other bugs/whatever in the meantime ? | 15:09.59 |
henrys | we have all the login info from Miah now | 15:10.03 |
Robin_Watts | Jogu has ATS up and crawling. | 15:10.19 |
mvrhel_laptop | back | 15:10.27 |
Robin_Watts | He's in the process of checking for regressions in the stuff we've done so far. | 15:10.44 |
henrys | pedro__: seems like that would be best if you have the cycles to make some progress on the bugs | 15:10.48 |
chrisl | henrys: perhaps the Artifex and picsel site information should go somewhere central - like the downloads.ghostscript.com stuff? | 15:11.02 |
pedro__ | henrys: I've got the Good 1.6 SDK integrated with the android build but no ability to test the changes properly until this auth issue is resolved | 15:11.08 |
henrys | Robin_Watts: you brought up Memento - has anyone run valgrind on any of this code? Is that something we should do? | 15:11.38 |
Robin_Watts | runs away screaming. | 15:11.59 |
henrys | pedro__: right, just wait for their guidance. | 15:12.05 |
Robin_Watts | ATS does (or did, or can do) various memory squeezing tests, and Picsel use fortify. | 15:12.31 |
| fortify is the thing that inspired memento. | 15:12.44 |
pedro__ | we did do valgrind on some of the DA tests regularly | 15:12.55 |
Robin_Watts | It's just missing a few of the interactive features. | 15:13.01 |
| He's found a couple of problem commits, (ones we've (I've) reapplied from the release branch onto trunk, AIUI) so those need looking at. | 15:13.21 |
| I'll try and look into those after I've got the customer a better answer than "look, just don't use that font, right?" | 15:13.44 |
pedro__ | and memsqueeze tests (inducing progressive memory failures via test scripts) | 15:13.54 |
henrys | I'll give you another reason to run and scream. We now have all the support requests and complaints sent to support@picsel.com for 3 years. | 15:14.28 |
paulgardiner_lap | henrys: For the iOS version of SOG, the blank-screen problem is fixed, but now I'm stalled by the same problem. I haven't looked at the API updates yet, but I don't know how wise it is to start on that until it's possible to verify the version using the old sdk is working. | 15:14.44 |
| stalled by the same problem as Pete that is | 15:15.02 |
Robin_Watts | superglues his finger to his nose | 15:15.08 |
inkbottle | Hi, I've got an academic pdf file which is not searchable. Would there be ways to improve that, look inside and see what's going on? | 15:15.14 |
kens2 | inkbottle : almost certainly not, once the PDF has been made that way, the only way to 'improve' it will be OCR | 15:15.51 |
inkbottle | okay | 15:16.05 |
Robin_Watts | inkbottle: The PDF is probably missing the 'ToUnicode' data. You'd need to get it regenerated for it to be searchable. | 15:16.10 |
chrisl | Robin_Watts: a lot of academic papers in PDF are just scanned, no text at all | 15:16.49 |
Robin_Watts | yeah. | 15:17.04 |
inkbottle | I can "select" text with the mouse, but the content is crap. | 15:17.08 |
| it's not scan | 15:17.12 |
henrys | paulgardiner_lap: yeah so same a pedro__ wait for guidance and work on bugs until it comes, there is not much we can do. I could talk to other folks at Good and see if we can get higher priority. | 15:17.14 |
inkbottle | badly converted, i think it is | 15:17.22 |
chrisl | inkbottle: Ah, then Robin_Watts is correct | 15:17.23 |
kens2 | tehn its a dsubset font with no ToUnicode CMap and the character codes are not selected as ASCII | 15:17.34 |
inkbottle | how do i do to look inside? | 15:17.54 |
kens2 | and that's impossible to turn back into searchable text, generally | 15:18.02 |
inkbottle | okay | 15:18.09 |
kens2 | inkbottle : download mupdf and ask Robin_Watts :-) | 15:18.13 |
henrys | mvrhel_laptop: are you able to build SOT now? | 15:18.29 |
inkbottle | I can have mupdf no problem | 15:18.31 |
marcosw | kens2: no it's not, just convert to a tiff using ghostscript and then load into acrobat pro, which will ocr the file and viola! | 15:18.37 |
Robin_Watts | inkbottle: mutool clean -difggg in.pdf out.pdf | 15:18.44 |
mvrhel_laptop | henrys: yes. I can build and run/debug with visual studio | 15:18.51 |
kens2 | marcosw that's OCR< I did say that would work.... | 15:18.53 |
inkbottle | Robin_Watts: I try that :-) | 15:19.04 |
kens2 | inkbottle : using MuPDF you cna deompress the file, then using the PDF Reference Manual you can figure out what it all means | 15:19.16 |
marcosw | kens2: I don't think acrobat will ocr a pdf file that's not an image. | 15:19.24 |
chrisl | marcosw: and that's not really recovering the information, that's starting from scratch and creating new information..... | 15:19.28 |
mvrhel_laptop | I spent a day stepping through the code but I need to read some of the documentation to understand how things flow | 15:19.32 |
henrys | mvrhel_laptop: and there is a documented way other folks can get to exactly where you are? | 15:19.40 |
kens2 | marcosw it won't no. My point is that once the file is like that, OCR is your only hope | 15:19.45 |
| (probably) | 15:19.49 |
mvrhel_laptop | henrys: yes. the twiki steps worked perfect | 15:20.00 |
| once Robin_Watts twisted my arm to install VS2010 | 15:20.31 |
| Robin_Watts: did have to fix one thing, but that is now checked in Robin_Watts ? | 15:20.54 |
henrys | mvrhel_laptop: oh 2013 won't do? | 15:20.55 |
Robin_Watts | mvrhel_laptop: It is. | 15:21.04 |
| I backported the changes to allow VS2013 to build, for at least some targets. | 15:21.12 |
| But for sanities sake, I would strongly recommend us all using VS2010 | 15:21.25 |
henrys | Robin_Watts: fine by me. | 15:21.36 |
Robin_Watts | 2013 is needed for the Windows 8.1 build, of course. | 15:21.55 |
marcosw | chrisl: fair enough, but I suspect inkbottle doesn't care how about the details, he just wants searchable text. | 15:22.01 |
henrys | Robin_Watts: can 2 VS's be installed at once or does that cause mayhem | 15:22.08 |
| ? | 15:22.10 |
Robin_Watts | henrys: I have many installed. | 15:22.15 |
| No problems. | 15:22.17 |
henrys | never tried it. | 15:22.18 |
| thanks | 15:22.23 |
Robin_Watts | 2005, 2008, 2010 at least. | 15:22.27 |
chrisl | I have a fair selection, too...... | 15:22.32 |
pedro__ | basically sot has host/target tool config files which we set up for each combination | 15:22.41 |
| hence the reluctance to do a full matrix... | 15:22.52 |
Robin_Watts | SOT has to be setup like that because we have *many* cross compile targets. | 15:23.10 |
| some customers supply custom toolchains etc. | 15:23.29 |
chrisl | Still sounds like a nightmare....... | 15:24.31 |
paulgardiner_lap | chrisl: how did you get such a deep understanding of SO in such a short time? :-) | 15:25.10 |
Robin_Watts | chrisl: It's codename was Cthulu | 15:25.29 |
chrisl | paulgardiner_lap: exposure to other, similarly "managed" products..... | 15:25.35 |
henrys | but the variety of devices seems to be shrinking. | 15:26.06 |
Robin_Watts | henrys: The problem is we are embedded in customers designs. | 15:26.34 |
| no one uses MIPS any more - except one of the customers we've got, for example. | 15:26.50 |
chrisl | Still, a build for every OEM multiplied by every hardware type they use | 15:27.21 |
pedro__ | its primarily toolchain setup, plus some postlink gubbins for any app builds/packaging | 15:27.59 |
Robin_Watts | chrisl: Yes. That's the wonder of closed source for you. | 15:28.00 |
chrisl | Robin_Watts: it's not inherent in closed source | 15:28.42 |
henrys | yes but how many platforms did you support in the past that have now been completely displaced by android? | 15:28.46 |
Robin_Watts | a few, certainly. | 15:29.22 |
inkbottle | marcosw: I don't mind having inside details; by the way the mupdftool trick didn't help. I think the ocr solution might be the only feasible solution. Also I suspect actually fixing it could require lots and lots of time; and i'm suppose to read it :-) | 15:29.24 |
Robin_Watts | chrisl: It is when our patent agreement prohibits us sharing the source under any circumstances. | 15:29.48 |
inkbottle | kens2: how do i decompress using mupdf | 15:30.03 |
| ? | 15:30.04 |
Robin_Watts | inkbottle: The mupdf trick was decompressing. | 15:30.12 |
kens2 | inkbottle : what Robin_Watts said | 15:30.14 |
Robin_Watts | It wasn't going to "fix" searchability. | 15:30.21 |
inkbottle | ok | 15:30.21 |
henrys | anything else for the SOT meeting? | 15:30.21 |
inkbottle | so i can read it now | 15:30.29 |
Robin_Watts | A person experienced in PDF can read it now, yes. | 15:30.43 |
chrisl | Robin_Watts: no, you keep the core closed source, and let OEMs handle their own customisations, and executable signing and stuff like that | 15:31.00 |
kens2 | Now you need the PDF Reference, a lot of time, and possibly a stiff drink or two | 15:31.12 |
henrys | I'm going to get this support email database and after I go through it I'll forward it on to paulgardiner_lap and marcosw | 15:31.15 |
Robin_Watts | chrisl: That's what we do. | 15:31.16 |
| We still need to do builds of the libs etc. | 15:31.25 |
pedro__ | chrisl: we did that too - that's what the TGV build was all about | 15:31.26 |
chrisl | Robin_Watts: so we only need one build of the lib per platform? | 15:31.40 |
pedro__ | unfortunately customers don't always want to do their own porting | 15:31.41 |
Robin_Watts | chrisl: a) customers often have their own toolchain. | 15:32.11 |
| b) we tag each lib for security purposes, hence one customer one lib. | 15:32.31 |
paulgardiner_lap | henrys: I'll keep a cat > /dev/null ready for its arrival | 15:32.34 |
chrisl | pedro__: no, which is why I said "handle" - the customisation etc, maybe done under NRE, but is the resposibility of the customer | 15:32.50 |
inkbottle | Robin_Watts: the out.pdf is interesting. It's seems it will need a long work to decipher it, but i can't help thinking it's awesome | 15:33.29 |
henrys | paulgardiner_lap: oh you're going to love this stuff, you don't need iain banks | 15:33.37 |
pedro__ | chrisl: its a more desirable model, but our experience was that 60-80% of the embedded customers would have walked away | 15:34.02 |
paulgardiner_lap | henrys: I can imagine some of it is quite amusing | 15:34.52 |
Robin_Watts | The gs model has always been "customers do their own porting/customisation/embedding", but even in that, we have discovered that we often bear the brunt ourselves. | 15:35.09 |
| "Please help. Having found my ass with both hands, I am now typing with my nose." etc. | 15:35.39 |
henrys | pedro__: I wonder if ghostscript hasn't suffered because we *let* customers do it. | 15:35.53 |
inkbottle | I've got another problem completely unrelated though; I've got a file containing probably "jpx" streams. They happen not to be displayed by konqueror (from some month ago). Could the decompressing trick help me to locate the jpx streams and see if there is something to be fixed there? | 15:37.04 |
| by the may debian's mupdf wont display it either | 15:37.32 |
pedro__ | there are pros and cons I reckon - you'll probably get your own libs integrated more along the lines you'd like/approve, but by the same token the customer exposure can broaden the technology in ways you may not have considered | 15:37.36 |
Robin_Watts | inkbottle: It is possible that the jpx streams trip openjpeg up. | 15:38.08 |
henrys | pedro__: yup pros and cons sure | 15:38.17 |
pedro__ | it mostly came down to resource/availability for picsel - we sometimes had 35-40 customer projects running in parallel | 15:38.20 |
Robin_Watts | We have various openjpeg fixes in (though they may have been fed back by now) | 15:38.27 |
henrys | pedro__: for us it really comes down to the encumbrance of the patents. Without that we'd probably do a source release with an NDA | 15:39.53 |
inkbottle | Robin_Watts: I think it is openjpeg which is complaining. Poppler wont decode it either. But it did some month ago. Imagemagicks "convert" do display it well | 15:40.18 |
Robin_Watts | inkbottle: Well, using a mupdf build from source that doesn't use system libs might solve it. | 15:40.42 |
pedro__ | nods - makes a lot of sense. In Picsel there was a definite get-rich-quick approach and a very tight reign over who got the code (nobody did while I was there) | 15:41.17 |
Robin_Watts | pedro__: Except Tricast Media etc. | 15:42.23 |
pedro__ | well, yeah - but they're family ;) | 15:42.36 |
inkbottle | Robin_Watts: is it not too difficult a task? (the building of mupdf?) | 15:42.44 |
Robin_Watts | inkbottle: It can't be that hard, I do it :) | 15:42.56 |
| git clone... && git submodule update --init && make | 15:43.12 |
inkbottle | Robin_Watts: right must be quite easy then | 15:44.02 |
| ok | 15:44.15 |
| Another thing unrelated (I'm glad there are people on that channel, who would have thought so?). I use a lot the Konqueror facilities to annotate a document. Annotations which do not alter the document. Contrarily to poppler's annotations which changes the document which behaviour I don't want | 15:47.23 |
| Is there another viewer that could do that? | 15:47.48 |
| annotate without modifying | 15:47.58 |
kens2 | You cna't annotate without modifying the document in some sense, unless you store the modifications outside the document, which is conceivable, but unlikely | 15:48.34 |
inkbottle | yes modifications are stored outside | 15:48.51 |
kens2 | Thwn there's no transportable way to do that. | 15:49.00 |
inkbottle | with konqueror | 15:49.01 |
| very convenient | 15:49.05 |
| not transportable at all, right | 15:49.20 |
kens2 | So if you send the file to someone else, annotatrions are gone. Less convenient.... | 15:49.23 |
inkbottle | less indeed | 15:49.32 |
kens2 | THis isn't going to be something you can do outside the OS/shell I would have thought, its going to need tight integration. | 15:50.15 |
inkbottle | but, there is convert action that you can choose, in order to send annotated file to a friend | 15:50.16 |
kens2 | Which presumably modifies the file and puts the annotations in it | 15:50.34 |
inkbottle | tight, yes | 15:50.34 |
| yes | 15:50.43 |
| my annotations are reading annotations, i don't realy want to keep them | 15:51.11 |
kens2 | So if your OS offers it then you a=have what you want, if it doens't you are stuck as far as I cna see | 15:51.11 |
mvrhel_laptop | bbiab | 15:51.56 |
inkbottle | There is xournal which save modification in a separate file too, but I find it less convenient for reading. | 15:55.28 |
| I'll e back some other time, bye | 15:55.38 |
kens2 | night all | 16:24.34 |
mvrhel_laptop | had to redraw the gsview icon in illustrator. turned out pretty good if I say so myself | 17:07.05 |
henrys | mvrhel_laptop: oh cool fun to do stuff like that once in a while | 17:13.56 |
mvrhel_laptop | henrys: yes it was fun | 17:17.26 |
jogux_mac | henrys: apparently I actually missed the meeting, sorry. I was here and watching my colloquy window, but for some reason it decided not to scroll to show messages as they arrived. :( | 17:31.36 |
| henrys : on valgrind; there is stuff in ATS to do it; I'll try and get it running when I get past the more important bits. | 17:32.10 |
henrys | jogux_mac: okay well if they are minor unknit conditions no big deal but illegal reads and writes we do like fixed. | 17:33.59 |
jogux_mac | henrys : also, I'm out on a business trip (east coast US) next week and the week after; but I'll still be around / contactable most of the time / progressing on ATS | 17:34.09 |
deleet | Robin_Watts: in case you're interested, this is all you need for MuPDF to be built on Android via Gradle (the library part, the NDK support is not great): http://pastie.org/private/ntoftn75yk3t91vxl4iqqa | 17:34.24 |
jogux_mac | henrys : yup, indeed. we found valgrind was very valuable for fixing things that would crash for the customer but not for us :-) | 17:34.30 |
ray_laptop | Robin_Watts: I don't know why I can't reproduce it on 32-bit Windows, but there is a progression on tests_private/ps/ps3cet/11-14.PS.psdcmyk.300.1 and what is probably a progression on tests_private/xl/pcl6cet3.0/C411.bin.ppmraw.600.1 The latter wasn't until I added the curveto case that you spotted, and looks like it might be that. | 17:38.21 |
| Robin_Watts: henrys: Do you think it is worth checking the C411.bin file with the debugger | 17:39.00 |
henrys | ray_laptop: is this on your bmpcmp on the dashboard? | 17:41.08 |
| ray_laptop: so that was a result of a banding change? | 17:43.36 |
| C411.bin? | 17:43.45 |
| ray_laptop: so the gxlcpath.c change you have in your repo caused the C411.bin progression? | 17:54.21 |
Robin_Watts | ray_laptop: What's with the other changes in the bmpcmp ? | 18:03.10 |
| Change 1 for example? | 18:03.22 |
ray_laptop | henrys: sorry. phone call. | 18:03.36 |
| henrys: yes, the bmpcmp on the dashboard | 18:03.48 |
Robin_Watts | change 10 looks bad too. | 18:03.51 |
ray_laptop | Robin_Watts: let me look. | 18:04.15 |
Robin_Watts | is 19 a progression? | 18:04.24 |
| I agree that 24 (C411) looks much better. | 18:04.57 |
henrys | it doesn't matter how it looks - just see the non banding result. | 18:05.26 |
| it should match yes? | 18:05.39 |
Robin_Watts | indeed. | 18:05.44 |
| I think it's clear that it's more correct. | 18:05.50 |
| (I'm mid SOT at the moment) | 18:06.05 |
ray_laptop | Robin_Watts: I have no idea what's up with the other changes that show up. Maybe they're diffs since I updated my local repo | 18:06.12 |
henrys | ray_laptop: I verified C411 progression correct | 18:06.31 |
| ray_laptop: I do find the commit message cryptic - I wonder if there is a better phrasing - I know if it is difficult to communicate. | 18:07.29 |
ray_laptop | henrys: I struggled with how to say it. I'll try again with things like Points A, B, C, ... and a more formal description and see if that's better. | 18:10.12 |
| Robin_Watts: I hadn't noticed 10 (Bug693711.pdf pg 34), and will look at it | 18:11.06 |
Robin_Watts | When writing paths into the clist, we take the opportunity to break them at a band level. This means that whole line segments that are not present within a band can be missed out. This code was going wrong. | 18:12.05 |
ray_laptop | Robin_Watts: rather than "going wrong" "messing up when a subpath was closed." | 18:13.12 |
Robin_Watts | I was still composing the next line to describe in which way it was going wrong. | 18:13.51 |
ray_laptop | Robin_Watts: henrys: so you don't think I need to describe how it was going wrong (missing a "catchup moveto") | 18:14.01 |
ray_laptop | waits, apologetically | 18:14.23 |
Robin_Watts | ray_laptop: I'd be tempted to pretty much use your description. | 18:14.32 |
| Just I felt that (for the uneducated reader) a warm up paragraph would help :) | 18:14.53 |
ray_laptop | Robin_Watts: Oh, I see. Thanks | 18:15.06 |
Robin_Watts | (I know that in 2 weeks time, I'll have completely forgotten the context of this) | 18:15.15 |
ray_laptop | henrys: is that good enough for you, or did you want the explanation of the conditions to be better ? | 18:15.58 |
henrys | ray_laptop: I'm good with the warmup | 18:16.12 |
ray_laptop | henrys: OK. Thanks, then. I'll revise it and ask for another quick review after I look into the Bug693711.pdf p34 issue | 18:17.08 |
Robin_Watts | If a subpath ends at an unclipped point X, and the next subpath begins at point X, then the existing code would forget to emit a moveto (or lineto). | 18:17.56 |
| This would cause a subsequent closepath to close the path incorrectly. | 18:18.48 |
| (Delete the '(or lineto)') | 18:18.59 |
ray_laptop | Robin_Watts: actually, it is only a moveto that is relevant, since that establishes the point used for the closepath | 18:19.18 |
Robin_Watts | yeah, hence my deletion :) | 18:19.31 |
ray_laptop | Robin_Watts: crossed in the ether | 18:19.46 |
| Robin_Watts: I'll wordsmith your description a bit, but it sounds basically more concise than mine (thus better) | 18:21.10 |
| Robin_Watts: thanks | 18:21.20 |
Robin_Watts | no worries. | 18:21.25 |
ray_laptop | hmm... The difference (regression) with Bug693711.pdf CAN'T be due to the banding change. It is 72.0 so no banding is used | 18:34.00 |
| I'll open a bug for it after double checking... | 18:34.48 |
Robin_Watts | ray_laptop: pattern clist ? | 18:43.23 |
| oh, that one. wierd. | 18:44.14 |
ray_laptop | OK. That's a pdfwrite problem. There are LOTS of "Error processing content stream" and the ERROR -15 closing pdfwrite | 18:45.44 |
| and the resulting PDF is damaged. | 18:45.55 |
| one for kens | 18:45.59 |
mvrhel_laptop | henrys: so what do we need to do to actually get gsview.com up and running? | 18:46.00 |
Robin_Watts | mvrhel_laptop: Does gsview work on win 7? | 18:46.42 |
ray_laptop | and the resulting pdf, after repair has the missing accent glyph | 18:46.44 |
mvrhel_laptop | Robin_Watts: it should | 18:46.58 |
| I need to do some testing with respect to this | 18:47.10 |
ray_laptop | I'll open a bug for that one. | 18:47.13 |
Robin_Watts | ray_laptop: Does the same problem happen without your change? :) | 18:47.14 |
ray_laptop | I have Windows 7 if you want me to try it | 18:47.24 |
| Robin_Watts: yes | 18:47.28 |
Robin_Watts | likewise. | 18:47.28 |
mvrhel_laptop | I have it here too | 18:47.28 |
| but I will need some testing soon | 18:47.36 |
ray_laptop | I have to run an errand. BBIAB | 18:47.52 |
henrys | phone | 18:48.38 |
| mvrhel_laptop: well we are waiting for the check to clear for gsview then we get a password. The new web guy will set it up exactly as we want and he'll work directly for us, Miles is sending his info to me. | 18:51.17 |
mvrhel_laptop | oh ok great | 18:51.32 |
henrys | mvrhel_laptop: I favor the style of the mupdf site but I don't feel strongly about it. | 18:52.09 |
mvrhel_laptop | yes. being consistent like that would be good | 18:52.22 |
henrys | mvrhel_laptop: I think a content managed site like artifex would be overkill. | 18:54.02 |
mvrhel_laptop | yes | 18:54.09 |
henrys | jogux_mac, pedro__, Robin_Watts and paul (for the logs) we're getting ready to set up an SOT user forum - do we have any strong preferences for news forum systems? | 18:57.36 |
| s/news/QA type/ | 18:58.02 |
Robin_Watts | One that's hard to spam. | 18:58.17 |
| So Captchas etc. | 18:58.51 |
| I think paulgardiner may have some experience of this through having setup a forum for glidos. | 18:59.10 |
| People should need to make an account, and pass a "human" test before being allowed to post anything. | 18:59.30 |
| If there was community upvoting/downvoting that'd be good too, cos it might limit the amount of work we had to do. | 19:00.04 |
| I wonder if StackOverflow do forum hosting for people... | 19:00.46 |
henrys | Robin_Watts: I thought about that but thought that was more for developers. | 19:01.24 |
Robin_Watts | http://www.discourse.org/ ? | 19:03.03 |
henrys | Robin_Watts: although http://stackoverflow.com/questions/tagged/ms-word hmmph | 19:03.29 |
mvrhel_laptop | lunch | 19:06.32 |
| bbiab | 19:06.36 |
henrys | pedro: are you about? | 21:30.33 |
pedro | yup, still here | 22:56.26 |
henrys | pedro: just what I sent in the email was all I had | 22:56.54 |
pedro | that's ideal - I think I had seen that email i one of the server setup docs but didn't actually remember until you emailed - the 'dev support' on the portal just fires people towards the community support | 22:58.12 |
| although the Good support folks do sit on the forums | 22:58.35 |
| I do have one remaining thing I'll try as a backgrounder and that's changing our package name to see if there's some conflict with the Good NOC servers | 22:59.37 |
| but hopefully we'll get something helpgul from support | 22:59.54 |
| helpful | 23:01.20 |
henrys | pedro: what is this secure office library api about? Is this a marketing piece or do we have to do a special build to get all the documented features in place? | 23:13.10 |
Robin_Watts | Is the secure office lib apu just the secure fs stuff? | 23:17.26 |
henrys | Robin_Watts: you were copied in on the API document from joseph, did you see that? | 23:18.25 |
Robin_Watts | I think I saw the email. I haven't read it. | 23:18.41 |
henrys | Robin_Watts: it says securers is optional but typically included in Secure Office | 23:21.39 |
| s/securers/securefs | 23:21.52 |
Robin_Watts | From the fag ends I've picked up... | 23:21.59 |
| If SecureFS is built in then it calls out to some Alien functions. | 23:22.35 |
| Those functions are implemented by the integrator (typically the customer, or picsel doing NRE). | 23:23.04 |
| So the idea is that if an integrator does the work, smart office can access any 'secure' store it wants to. | 23:23.54 |
| and presumably SO is smart enough to avoid doing cut/paste and save between secure and insecure documents. | 23:24.23 |
henrys | is the "Good" configuration using parts of this? I assume they would want to. | 23:26.51 |
pedro | yes, the securefs stuff basically gives a ring-fenced filesystem. Some of the SOL builds use it, but I think SOL is just intended to let customers differentiate their apps (rather than use the whole SmartOffice UI) | 23:27.57 |
henrys | we are getting a sense that this an important market segment: secure office setting so I'm trying to figure out how we "productize" this. | 23:27.59 |
pedro | Good uses securefs to guarantee that docs are not leaked to other apps/clipboards etc | 23:28.35 |
| I believe it would have been a premium feature for Picsel (ie extra cost over the standard SO) | 23:29.19 |
henrys | pedro:SOL is a separate library that is released with a set of appropriate tests | 23:30.39 |
pedro | nods | 23:32.15 |
henrys | pedro: okay that's what I'm after Miles and Scott need to market this stuff so we can say well the secure library adds security to the library blah blah⦠cool | 23:33.18 |
| Forward 1 day (to 2014/05/07)>>> | |