| | <<<Back 1 day (to 2014/06/19) | 2014/06/20 |
| mvrhel_laptop | dinner break. much to debug in my v2 profile creation code.... | 01:14.23 |
| | plus need to review chrisl's code | 01:14.40 |
| | hi rayjj | 01:15.06 |
| | is it hot there? | 01:15.09 |
| | I sent a quick reply about the trans stuff. I can't imagine how they are going to get that all working | 01:15.41 |
| | bbiaw | 01:16.29 |
| | ok success on the first v2 icc profile | 05:20.21 |
| | one more v2 profile type to debug tomorrow then I think this stuff will be done | 06:44.14 |
| | night all | 06:44.24 |
| pedro_mac | night mvrhel_laptop | 06:44.40 |
| mvrhel_laptop | pedro_mac, then it is on to SOT for me... so I guess I better get some good rest | 06:45.18 |
| pedro_mac | indeed - should be fun :) | 06:45.50 |
| kens | Hmm, Matt has been busy :-) | 07:54.05 |
| Robin_Watts | kens: Those didn't all happen overnight, but yes :) | 08:01.57 |
| kens | 5 fixes in a week (or so) seems pretty good to me, from a stganding start | 08:02.43 |
| Robin_Watts | absolutely. | 08:03.33 |
| pedro_mac | hi folks | 08:25.00 |
| Robin_Watts | morning | 08:25.11 |
| kens | Hi Pete | 08:25.40 |
| Robin_Watts | http://twiki.ghostscript.com/do/view/MuPDF/EmbeddingMuPDF | 08:27.53 |
| | If anyone can spot any mistakes, or can suggest anything else that should be in there, please let me know. | 08:28.11 |
| mattchz | morning | 08:28.23 |
| Robin_Watts | I need more figures in there for memory use, and to fill in the last couple of sections etc obviously. | 08:28.41 |
| sebras | Robin_Watts: wouln't shared libraries save not just ROM, but RAM as well..? | 08:30.09 |
| Robin_Watts | How? | 08:31.31 |
| pedro_mac | depends if its XIP I guess | 08:31.38 |
| Robin_Watts | by ROM I mean "program memory" and RAM I mean "data memory". | 08:32.00 |
| sebras | pedro_mac: good point. I didn't think of XIP. it was a long time since I though about NOR. :) | 08:32.02 |
| pedro_mac | :) | 08:32.09 |
| Robin_Watts | pops off to blast bakeolite. | 08:32.47 |
| sebras | Robin_Watts: hm... the part about allowing custom allocators. is this true for the libraries we use too? | 08:33.13 |
| | Robin_Watts: I think zlib and libjpeg are ok with it, but I'm unsure about the rest. | 08:33.30 |
| | curl, openjpeg...? | 08:33.36 |
| | "underlying graphic library" -> "underlying graphic_S_ library" as used in the heading. | 08:34.38 |
| chrisl | sebras: we currently have someone looking at OpenJPG, I fixed/hacked libjpeg a while back.... don't know about the others | 08:41.09 |
| sebras | chrisl: right. | 08:46.18 |
| chrisl | So the one we *know* is currently a problem is OpenJPG | 08:46.46 |
| sebras | chrisl: ok. hopefully upstream will accept any allocator patches. | 08:47.11 |
| chrisl | sebras: that's been the delay - they are accepting patches, but they are *really* slow doing so. Otherwise it would have been finished months ago :-( | 08:48.04 |
| sebras | chrisl: yes I know. I reported a but ~9 months ago, and they have just recently started to care about it. | 08:48.51 |
| | chrisl: despite me actually sending a file that reproduces the problem. oh well... | 08:49.11 |
| chrisl | sebras: they do seem to go through periods of rapid activity, and then longer periods of no activity | 08:51.22 |
| sebras | chrisl: sounds like there is no paid developer on that project then. | 08:51.47 |
| | chrisl: which is bit of a pity, because it seems to be the only viable alternative for decoding/encoding openjpeg. | 08:52.33 |
| chrisl | sebras: I suspect there are a few who work on it in a professional capacity, but they only pay attention to it when their requirements aren't being met - then we get the rapid activity, new release, then back to the doldrums again | 08:53.19 |
| sebras | chrisl: oh. in that case it might be possible for artifex to throw a bag of money at them and see what happens. | 08:54.13 |
| | chrisl: of course this is possible for a " | 08:54.35 |
| | non-business-backed" project as well, but unless they have someone in place to receive the donations then I guess it will not have the desired effect. | 08:55.25 |
| chrisl | sebras: I think that's the problem: I have the impression that the OpenJPG devs who work on it "professionally" do so as *part* of a "real" job - it's not the main focus of their work | 08:56.52 |
| | I have to go for a hair cut - bbiaw...... | 09:34.11 |
| mattchz | robin/paul: mind reviewing my change please? (last 4 commits) | 09:44.00 |
| | http://git.ghostscript.com/?p=user/matt/mupdf.git;a=summary | 09:44.01 |
| gians | hi guys .. i'm, Gians ... is there someone that can help me with a error compiling example.c mupdf 1.5 ? | 10:30.29 |
| | i think i missed a lib, therefore i linked all lib as wrote on example file .. | 10:32.58 |
| | i get this error: http://pastebin.com/tgK4VF5u | 10:33.18 |
| | os is Mac 10.9 | 10:33.51 |
| kens | Loos like missing crypto stuff. I'm afraid I can't help with MuPDF builds though | 10:35.29 |
| chrisl | gians: there probably will someone who can help at some point - you may just have to hang around for a while....... | 10:35.29 |
| jogux | it's missing openssl it seems | 10:35.49 |
| kens | jogux : that was my guess, but I don' tknow enough to say | 10:36.10 |
| | chrisl I've got something really weird going on, are you free for a minute ? | 10:36.58 |
| gians | but mupdf require openssl lib ? | 10:37.10 |
| chrisl | kens: yeh, shoot | 10:37.11 |
| kens | This bug #695322 'stranbge characters' | 10:37.35 |
| gians | i compiled from source mupdf 1.5 without problem | 10:37.46 |
| kens | I've reduced the file to the point where it draws two space glyphs, both in the same type 0 font | 10:38.00 |
| jogux | gians: you probably need -lcrypto in your linker command | 10:38.10 |
| kens | The first time, teh string of data is 0x0020, the second time its 0x2020 BUT only with pffwrite.... | 10:38.26 |
| Robin_Watts | gians: MuPDF will use crypto for digital signatures. | 10:38.29 |
| kens | chrisl I'll just mail you my reduced file. | 10:38.37 |
| chrisl | kens: do I need to fonts, too? | 10:38.47 |
| kens | fonts are int the PS file | 10:39.03 |
| gians | ok ok thanks! try to add -lcrypto to linker | 10:39.44 |
| kens | I put a breakpoint in zxshow, and looked at the string of data and with a rendering device, its 0x0020 the first time and 0x0020 the sewcond time. With pdfwrite its 0x0020 the first time and 0x2020 the second. | 10:39.59 |
| Robin_Watts | gians: BIO looks like an openjpeg problem. | 10:40.07 |
| | gians: If you're building with HAVE_OPENSSL defined, then probably, on balance, you ought to have openssl available :) | 10:41.05 |
| chrisl | kens: give me a minute - I'll need to change branches and rebuild...... | 10:41.06 |
| kens | chrisl this is why it fails of course. I have a sneaky suspicion that pdfwrite is borking the string being passed, because instead of doing what any sensible person would, and doing <0020> show, they instead call a function to do... <0020> show. | 10:41.07 |
| Robin_Watts | mattchz: I'll look at your reviews, but not for a bit. I have a doctors appt in 20 mins. | 10:41.28 |
| mattchz | no rush at all. I'm working on non-mupdf stuff atm anyway | 10:41.42 |
| gians | Robin_Watt: thanks.... i'm looking for | 10:42.43 |
| | Robin_Watt: is there a way to compile MuPdf with HAVE_OPENSSL=NO ? | 10:43.29 |
| pedro_mac | paulgardiner> did you see Brad's email? when I was saying yesterday about the app ids having to match I forgot that you'd still be activated using our own GC rather than the GoodShare one | 10:44.55 |
| gians | compiled ok! with -lcrypto to linker command | 10:45.33 |
| | thanks all | 10:45.47 |
| kens | chrisl OK I just proved it. pdfwrite is stomping on the string in the bound procedure | 10:46.39 |
| chrisl | kens: well, that's clearly bad! | 10:47.00 |
| pedro_mac | paulgardiner: if you delete and reinstall your iOS SO it would be best to use one of the original GoodShare access keys too, as they expire tomorrow I think | 10:47.07 |
| kens | Yeah, I guess it never occured to whoever was doing it that someone would reuse the same string. | 10:47.23 |
| pedro_mac | and we can keep the ones Brad just sent for additional devices/test | 10:47.24 |
| kens | Now I need to see where we are doing that, and fix it by copying the string data | 10:47.39 |
| chrisl | kens: I'm shocked that hasn't been a problem before | 10:47.45 |
| kens | Well, normally strings don't get reused when you 'show' them | 10:48.00 |
| chrisl | Not normally, but it's not *that* rare - I think some of the QL tests do it | 10:48.33 |
| kens | At least I know what to look for now | 10:48.36 |
| paulgardiner | pedro_mac: right. and i should use com.picsel.smartoffice2gd rather than com.picsel.smartofficegd? | 10:48.37 |
| kens | chrisl it has to be a composite font too | 10:48.48 |
| chrisl | Yeh, that's probably why we got away with it | 10:49.02 |
| kens | I owuld imagine so yes | 10:49.10 |
| | Its the scan_cmap_text routine that breaks it I htink | 10:49.22 |
| pedro_mac | paulgardiner: yup | 10:49.32 |
| kens | I'm just not quite sure why right at the moment | 10:49.32 |
| chrisl | kens: quick question: in gs_init.ps there is a procedure "/.systemvar" which gets a value out of systemdict - how would you feel about a "/.systemexec" which does the same, but execs the result? | 10:50.22 |
| paulgardiner | I'll give that a go, although seeing as our own test app is working it's no longer an essential test | 10:50.29 |
| | ... for iOS at least | 10:50.39 |
| kens | chrisl, not a problem as far as I can see | 10:50.47 |
| chrisl | kens: cool, it might make for a neat fix for a bug Shelly is looking at | 10:51.08 |
| kens | Is there a reason not to just use .systemvar exec ? | 10:51.11 |
| pedro_mac | sure, but I'd be keen to test that the apps the customers use actually work ok. I also noticed a new Good Pro app on android - looks like a GfE replacement | 10:51.28 |
| chrisl | kens: just space and easy reading, really | 10:51.40 |
| kens | OK, well not a problem from my POV | 10:51.54 |
| chrisl | okay, we might concoct another idea, but we'll see | 10:52.15 |
| chrisl_r500 | And now to start the traditional wrestle with cust 532's simluator...... :-( | 10:53.37 |
| kens | Oh what fun..... | 10:53.46 |
| kens | wonders when marcos is coming back from vacation, I could do with a pause in the customer bug reports | 10:54.17 |
| | chrisl didn't we have someone wittering about stopping text making it into pdfwrite output at some point ? I think we came up with a solution too | 10:57.28 |
| chrisl | You mean text disappearing in the output PDF? | 10:58.30 |
| kens | searches irc logs | 10:58.32 |
| | chrisl, he wants the text to disappear, I htink I found it | 10:58.46 |
| chrisl | Oh, yes, you hacked the tr mode, or something? | 10:59.15 |
| kens | Yes, I think I see it here: | 10:59.36 |
| | http://ghostscript.com/irclogs/2014/05/21.html | 10:59.38 |
| paulgardiner | pedro_mac: ah ha! I guess you must have changed the password policy at some stage to perform one of the tests GD require. I wondered why I got asked to change my password. Good: that's one test I've already conducted and don't need to do again. :-) | 11:00.17 |
| pedro_mac | paulgardiner: yeah, I've been working through the security/data leakage test in the pre-veracode checklist | 11:01.02 |
| | just added a 'needs a number' criteria | 11:01.17 |
| chrisl_r500 | kens: yeh, I remember now - none of us could work out why hacking the tr didn't work...... | 11:01.30 |
| kens | because it was a type 3 :-) | 11:01.39 |
| jogux | robin_watts: I have buyer's regrets over this pptx issue. | 11:01.57 |
| chrisl_r500 | Aha! | 11:02.04 |
| jogux | robin_watts : also I may be blind, but the output I'm getting doesn't look obviously wrong compared mac ppt: https://www.dropbox.com/s/sdp940j9hsdtagt/Screenshot%202014-06-20%2012.05.14.png | 11:05.51 |
| pedro_mac | jogux: interesting | 11:06.36 |
| pedro_mac | sees similar opening Art5.pptx in Office 2010 on windows7 | 11:08.22 |
| | (both show as 18pt in the font dropdowns) | 11:08.54 |
| Robin_Watts | jogux: hmm. | 11:09.30 |
| paulgardiner | pedro_mac: well that didn't work. I provisioned using the same email as for Good Share and using another of the keys they gave us and still Good Share wont allow me to Open In SOG | 11:11.06 |
| Robin_Watts | jogux, pedro_mac: OK, if I open Art5.pptx on Microsoft Office, I get "Row 2" in very large text. | 11:12.29 |
| | This is office 2007. | 11:12.37 |
| jogux | interesting. | 11:12.45 |
| Robin_Watts | Now, officially the stance we take is that if Libre and MS get it right, we should too. | 11:13.02 |
| | so perhaps my cutdown file isn't a perfect example. | 11:13.43 |
| jogux | I'll recheck the none-cutdown one, thanks | 11:14.26 |
| Robin_Watts | Try opening Artifex2014.pptx and looking at page 10 | 11:14.29 |
| | On 2007 the list is all in 18pt, and the rest of the text is in 16pt | 11:15.09 |
| | In SOT the list appears as 22pt (or was it 24?), and consequently everything falls off the bottom of the screen. | 11:15.49 |
| pedro_mac | Robin_Watts: its actually a great example of the inconsistent MS Office behaviour between versions tbh ;) | 11:16.17 |
| Robin_Watts | pedro_mac: Yeah :( | 11:16.25 |
| pedro_mac | Robin_Watts: on page 10, the top-level (TxLevel-1) is 22pt by the looks of it | 11:19.36 |
| Robin_Watts | pedro_mac: The top level TxLevel-3 is 22pt too, I think. | 11:20.02 |
| | and the list contents are TxLevel-3. | 11:20.13 |
| | BUT... they are placeholders, so shouldn't inherit from the top level TxLevel-3. | 11:20.37 |
| | (by "top level" TxLevel-3, I mean "the TxLevel-3 defined in presentation.xml") | 11:20.58 |
| pedro_mac | nods | 11:21.20 |
| jogux | https://www.dropbox.com/s/j1uaif9wfekz61a/Screenshot%202014-06-20%2012.23.00.png is from SO; am I looking at the right page? | 11:23.30 |
| Robin_Watts | Yes, but that's not how SO renders it for me. | 11:24.11 |
| | Is this with my patches? | 11:24.20 |
| jogux | yes | 11:24.25 |
| | in a mac testshell build | 11:24.29 |
| Robin_Watts | Maybe my patches are working?!?! | 11:24.33 |
| jogux | that would certainly be one conclusion :-) | 11:24.45 |
| | let me check in a full SO build. | 11:24.50 |
| Robin_Watts | I was using a testshell build without my patches to see this fail. | 11:25.13 |
| jogux | actually let me check without your patches, that'd be a good test. | 11:25.58 |
| Robin_Watts | am rebuilding without patches now. | 11:26.50 |
| jogux | I see it wrong without your patches. | 11:27.41 |
| Robin_Watts | I win! | 11:27.52 |
| jogux | you do | 11:27.57 |
| Robin_Watts | W.T.F. | 11:28.03 |
| jogux | should I attempt to understand why your solution seems to be working? | 11:28.08 |
| Robin_Watts | I suspect that someone should :) | 11:28.17 |
| jogux | ok | 11:28.23 |
| jogux | runs it through ats too | 11:28.27 |
| Robin_Watts | Ta. | 11:28.33 |
| jogux | the style stuff seems a good place to add some extra debug output, it seems very lacking just now :( | 11:30.59 |
| Robin_Watts | jogux: I've been meaning to ask... can the debug stuff be enabled/disabled at runtime other than by the options in the frontend? | 11:31.48 |
| jogux | EPAGE_DEBUG="module.tag" will do it | 11:32.13 |
| Robin_Watts | Like can I do: Alt-Ctrl-Q in Visual Studio and call a C function ? | 11:32.18 |
| | or 'call xxxx(....)' in gdb etc. | 11:32.36 |
| jogux | ah, yes. DEBUG_SET_STRING macro, if you track that down to the function, you can call that with module.tag | 11:32.38 |
| Robin_Watts | Fab, ta. | 11:32.49 |
| pedro_mac | Epage_setDebugId() or similar ? | 11:33.06 |
| jogux | might be DEBUG_SET_FROM_STRING. it's something like that. | 11:33.07 |
| | Epage_setDebugId is takes the enum values iirc | 11:33.22 |
| | the string forms are generally a bit easier. | 11:33.30 |
| pedro_mac | nods | 11:33.42 |
| Robin_Watts | OK, so my changes horribly break Artifex2014.pptx | 11:34.00 |
| | Lots of returns seem to go missing. | 11:34.07 |
| | So, I visited the doctor cos of pain in my ear. Apparently I have a huge lump of wax. | 11:36.20 |
| | He asked if I'd ever had a perforated eardrum, and I foolishly said "I think so, as a kid" | 11:36.37 |
| | and he said "ah, well, then I need to refer you to the hospital for treatment as I can't syringe them here then." | 11:37.12 |
| | "It'll be a long wait." | 11:37.30 |
| | Bugger. | 11:37.33 |
| pedro_mac | Robin_Watts: :( - no old remedies ? | 11:38.49 |
| jogux | robin_watts : can't you get vacuumed? | 11:44.10 |
| | robin_watts: which page is horribly broken? | 11:45.01 |
| | I suspect I should add this document into ATS | 11:45.38 |
| Robin_Watts | page 2 for a start | 11:46.25 |
| | pedro_mac: No, I'm generally fine with my ears these days. Might try olive oil etc. | 11:46.47 |
| | and page 4... | 11:47.02 |
| | and 8 | 11:47.13 |
| | sorry, 2, 5, and 9. | 11:47.25 |
| | stupid off by 1 numbering :) | 11:47.31 |
| jogux | something is definitely up with 2. | 11:49.14 |
| Robin_Watts | Did anyone review matts stuff? | 11:55.05 |
| | commits all look fine to me. | 11:57.19 |
| paulgardiner | Oh, I'll take a look | 11:59.38 |
| Robin_Watts | paulgardiner: I just didn't want to review if someone else already had. Turns out the only big patch there is one we'd basically OKd before. | 12:03.34 |
| paulgardiner | Last 4 right? I'm surprised at needing changes to buils with Eclipse because I always use Eclipse | 12:04.32 |
| | Robin_Watts: I was looking mainly because I was expecting them to be in the Android app | 12:05.06 |
| Robin_Watts | paulgardiner: el_mendi may appear later today with questions about the android app. He's a free user that has been trying to extend MuPDF as part of a college course. | 12:05.55 |
| paulgardiner | Robin_Watts: oh okay. Thanks for the heads up | 12:06.18 |
| Robin_Watts | he has got it to the stage where the android app can update a given page by adding an image to the page contents. | 12:06.37 |
| | the android app drops the display list from the appropriate pagecache entry, and invalidates the bitmaps. | 12:07.17 |
| | So the page updates correctly. | 12:07.26 |
| | But when he exits the app, it doesn't give him the option to save the updated file. | 12:07.55 |
| paulgardiner | I'd imagine that's a problem in the C library not the java. The java just asks the library if the doc has changes, IIRC | 12:10.14 |
| Robin_Watts | sebras: All our libs allow the use of custom allocators, except openjpeg. | 12:22.14 |
| | and Shelly has just fixed that in openjpeg I think. | 12:22.25 |
| | so it's just a question of us updating to the latest version, I believe. | 12:22.40 |
| chrisl_r500 | Robin_Watts: no the work is on-going - the OPJ devs rejected the first patch he sent, then a couple of months later, applied it, and did a release. But it's only the first step.... | 12:24.41 |
| Robin_Watts | ok, but it's getting there. | 12:24.58 |
| | and if it becomes a showstopper, we can fix it in mupdfs version directly, so I think I'm good to claim it. | 12:25.24 |
| chrisl_r500 | Yeh, getting there - as quickly as the OPJ devs and Shelly's day job allows | 12:26.08 |
| rayjj | I may regret it, but I responded to Zoltan :-/ | 12:46.52 |
| henrys | rayjj: responding to Zoltan while on vacation ⦠really ? | 12:48.14 |
| chrisl_r500 | I'm not sure that spending an hour and ten minutes tracking down a build problem in cust 532's simulator (again!) is the best use of my time :-( | 12:54.16 |
| rayjj | chrisl: it worked "out of the box" for me. Did you start it with: setsim.bat gemini_prn_es ? | 12:55.49 |
| | with 'alphard2' it wouldn't build, but the most recent emails from then had mentioned gemini_prn_es so that's what I used. | 12:56.39 |
| chrisl | rayjj: it was still the AUXDIR being undefined, so it was trying to write files into the root directory of the drive | 12:56.49 |
| rayjj | heading out for the day. bbl8r | 12:57.09 |
| henrys | paulgardiner: my request for a Good UK contact was rejected ;-( | 13:01.16 |
| paulgardiner | Really! On what basis | 13:01.37 |
| | ? | 13:01.39 |
| henrys | they claim nobody in the UK has the right skill set | 13:03.01 |
| Robin_Watts | henrys: All their Ukers are hampered by being able to find their ass with both hands. | 13:03.42 |
| henrys | and now he wants to have another conference call, OMG, only to tell us heâll have to ask the UK guys to answer our questions | 13:04.21 |
| | dunces are difficult | 13:05.02 |
| | Robin_Watts: this truly does border on comical be glad youâre not involved | 13:06.12 |
| paulgardiner | On the iOS side, Picsel had developed a test app that communicates with SOG in the way that other GD apps might. I hadn't been able to get that working before, but now it is working, so I've been able to do the testing I wished with that. | 13:06.47 |
| Robin_Watts | henrys: Oh I do. | 13:06.48 |
| | How can mupdf take more memory to render 100 pages of pdf_reference17.pdf at 1dpi than it does at 72dpi?? | 13:07.48 |
| chrisl_r500 | More objects being cached? | 13:08.23 |
| paulgardiner | Some of the tests they ask for require communication specifically with Good Share, which we can't do (not with the instructions they've given so far at least), but the Picsel test app tests the same things. | 13:08.29 |
| Robin_Watts | chrisl_r500: caching disabled, I think. | 13:09.04 |
| chrisl_r500 | Robin_Watts: an interesting quirk of software, then...... | 13:09.33 |
| paulgardiner | There's possibly only one outstanding problem on iOS and that's not really GD specific. It's a problem with standard Open In from other apps. Get that sorted and run through as many of their tests as we can, and maybe we could submit a new version | 13:11.19 |
| Robin_Watts | henrys: http://twiki.ghostscript.com/do/view/MuPDF/EmbeddingMuPDF | 13:11.40 |
| | Again, if anyone has comments, ideas for stuff to add etc, please tell me. | 13:11.56 |
| Robin_Watts | lunches | 13:11.59 |
| henrys | Robin_Watts: reading | 13:13.09 |
| paulgardiner | jogux_mac: what did you say about soft threading the other day? | 13:28.30 |
| kens2 | decides ray can reply to Zoltan's latest missive..... | 13:28.47 |
| | FWIW pdfwrite does get the correct output from that file, so as Ray says its clearly possible to ge the information without having teh pdf14 compositor in place. | 13:29.31 |
| jogux_mac | paulgardiner: I remove all the soft threading support code on iOS | 13:29.33 |
| | paulgardiner: Although it was disabled because of a bug (a bug I believed I could see how to fix) I can't see any advantage to soft threading on a modern OS. | 13:30.10 |
| jogux | someday I'll find out wtf the I've configured colloquy to use jogux_mac. | 13:30.37 |
| paulgardiner | Okay, I was just checking that there wasn't a fix on iOS SO that I might need to do something extra to get also for SOG. I'm seeing Pal_Thread_doMutexLock assert because of m->tData->current being NULL. | 13:32.31 |
| jogux | ooo. dodgy. I've not seen that. | 13:35.55 |
| | the most likely explanation for that would be an epage function called from outside the context of an epage thread. | 13:36.11 |
| | do you have a backtrace? | 13:36.17 |
| paulgardiner | I see it if I open a document in SOG (the first stage of opening where you see the pages and you have options for share etc) and then I go to another app and do Open In SOG | 13:37.10 |
| | I think the same happens in SO | 13:37.23 |
| | http://pastebin.com/J5cEL7Ln | 13:38.56 |
| jogux | paulgardiner: that does indeed look like a bug to me | 13:40.06 |
| | the preShutdownFn is running outside of the scope of scheduler, so would not be allowed to take locks. | 13:40.22 |
| paulgardiner | /* Loading a new document can only be allowed if we can | 13:40.28 |
| | * successfully shutdown the application */ | 13:40.29 |
| | Seems to be a new bug too. App store version doesn't seem to suffer from this | 13:45.45 |
| jogux | paulgardiner: oh really? boo :( | 13:48.05 |
| | I'd imagine someone has tweaked what the shutdown function does lately then. | 13:48.22 |
| paulgardiner | So, although deep in the schedular, is this to do with not closing an existing doc that has changes without saving? | 13:50.54 |
| jogux | paui: it's not in the scheduler afaics, that's the problem | 13:54.15 |
| paulgardiner | Hmmm maybe not. "Shutdown Handlers" I remember something like that being introduced, some considerable time ago and perhaps unnecessarily, as a way to clean up resources, instead of using thread shutdown and letting threads clean up naturally via their error handling. | 13:54.38 |
| jogux | paul: but, yes, the issue is it's asking ue2fileviewer "do you have unsaved changes", and ue2fileviewer is trying to take locks to answer that question. | 13:54.40 |
| | paul: iirc, they were implemented as a way to allow the picsel side of the application to object to the shutdown request if (eg) there were unsaved changes | 13:55.24 |
| paulgardiner | shutdown handlers were? | 13:56.13 |
| jogux | mm | 13:56.30 |
| paulgardiner | henrys: a point has just come up that you might want to weigh in on (discussing on Skype) | 14:02.10 |
| henrys | paulgardiner: okay just a minute | 14:02.42 |
| paulgardiner | No hurry. | 14:03.08 |
| jogux | paulgardiner: an easy way to fix that is likely to be to get the app to store a flag saying "I currently have unsaved changes", then the shutdown function can simply check the flag. (I think. I've not looked in detail.) | 14:05.55 |
| | robin_watts: " the exact details of the example viewers for these OS is not covered here." - is not -> are not? | 14:06.46 |
| Robin_Watts | jogux: Ta. | 14:07.03 |
| paulgardiner | jogux: sounds like a plan, but I'd be interested to find out what has changed to cause the prob | 14:07.32 |
| | lem | 14:07.37 |
| jogux | paulgardiner: I wonder if it didn't check for unsaved changes in this case previously :-) | 14:07.49 |
| paulgardiner | Possibly. I wondered that too. I can check | 14:08.17 |
| | ... but someone would have tested the fix... hmmm | 14:08.34 |
| Robin_Watts | Oh great. | 14:15.53 |
| | I have a docx file with just a single paragraph of text in 14pt Arial in it. | 14:16.21 |
| | LibreOffice breaks the first line in a different place than MS. | 14:16.41 |
| jogux | robin_watts: From a scan the rest make sense to me, and I learnt some new things about mupdf. I was surprised that the 72 dpi rendering memory usage is lower than the 1dpi. | 14:20.02 |
| Robin_Watts | jogux: Me too! | 14:20.14 |
| jogux | robin_watts: it might be useful to include an ARM binary size too perhaps? | 14:24.16 |
| Robin_Watts | jogux: Yeah. | 14:24.41 |
| jogux | Picsel used to do fun graphs of "with <x axis> of RAM, you can process 95% of files of size <y>". They were a reasonable amount of effort to prepare though. | 14:25.28 |
| Robin_Watts | jogux: Only if you trouble yourself to make them truthful :) | 14:26.01 |
| jogux | :-) | 14:26.07 |
| | Picsel made the mistake of asking engineers to do them then :-) | 14:26.23 |
| pedro_mac | â¦and based on a fairly unrepresentative set of docs, with lots of duplicates] | 14:26.23 |
| henrys | I think the only person happier than sabrina about the artifex meeting in London is Tor :-) | 14:27.51 |
| chrisl | henrys: London in December? I think Miles has been sniffing gas fumes! | 14:29.13 |
| Robin_Watts | chrisl: "Downtown" London. | 14:30.11 |
| | Claridges, I hope. | 14:30.16 |
| henrys | I thought London was a âChristmasâ destination ⦠| 14:30.28 |
| chrisl | Still, it's going to be cold and soggy...... | 14:30.36 |
| pedro_mac | Robin_Watts: Travelodge :) | 14:30.48 |
| Robin_Watts | pedro_mac: urk. Tor/sebras stayed in the travelodge at Marylebone. It wasn't nice. | 14:31.48 |
| el_mendi_ | paulgardiner: Hi, I need some help saving my document in Android | 14:31.51 |
| Robin_Watts | The Landmark next door is much nicer :) | 14:32.01 |
| henrys | chrisl: my âcolorado coldâ definition might be a bit different than yours. | 14:32.05 |
| el_mendi_ | paulgardiner: are you available? | 14:32.14 |
| pedro_mac | nods - Travelodge is that way by design.. | 14:32.20 |
| paulgardiner | el_mendi_: very busy but we can have a go | 14:32.41 |
| chrisl | henrys: undoubtedly, but the softy Californians might struggle! | 14:32.58 |
| henrys | chrisl: true | 14:33.16 |
| Robin_Watts | el_mendi_: Random thought... why don't you hardwire hasChanges in MuPDFCore.java to return true? | 14:33.28 |
| | That way you'll always be prompted to save, and we can at least see if it works ? | 14:33.44 |
| | henrys: "crisp and cold" is very different to "soggy and dank" though. | 14:34.25 |
| el_mendi_ | I am prompted to save because I changed the "dirty" parameter of the pdf_document variable. But when I save it the page in cuestion gets totally empty | 14:35.01 |
| | Like it doesn´t recognise the new contents of the page and thinks it is empty... I don't know | 14:36.31 |
| paulgardiner | el_mendi_: totally empty when viewed in Adobe Reader (or similar)? | 14:37.13 |
| jogux | if any of you guys want to pop up to scotland whilst you're in the UK I can put you up :-) | 14:37.14 |
| Robin_Watts | Did you try manually calling fz_write_document? | 14:37.36 |
| el_mendi_ | paulgardiner: I've not tried to visualize it in another reader... wait | 14:38.06 |
| henrys | jogux: I was going to ask if you guys wanted to come down to London for a day, weâd pay for travel and accomodations. | 14:38.53 |
| jogux | henrys: I would come down to have a drink with you guys anyway, so if you pay for it that's even better :-) | 14:39.25 |
| henrys | jogux: do they let you over the border ? I heard you guys were trying to revolt or something ;-) | 14:39.31 |
| jogux | henrys: TBD in September ;-) | 14:39.43 |
| el_mendi_ | paulgardiner: Yes , the pages where I inserted an image and saved after are empty. | 14:39.45 |
| pedro_mac | henrys: would be great to meet everyone, but as jogux says we'd also be glad to see you here if you want somewhere even colder ;) | 14:40.00 |
| Robin_Watts | Glasgow is the European city of Culture, don't you know. | 14:40.35 |
| pedro_mac | henrys: its one of those votes where nothing would happen for years anyway | 14:40.37 |
| paulgardiner | el_mendi_: I'd suggest doing this experiment with a fairly simple document and then inspecting the resulting PDF in a text editor to try to figure out what's wrong with the page that makes it appear blank | 14:41.00 |
| el_mendi_ | Ok | 14:42.29 |
| kens2 | jogux : that'll give you a couple months to apply for a new passport :-) | 14:45.59 |
| jogux | :-) | 14:46.19 |
| Robin_Watts | kens: jogux is english, so you and he will be part of the prisoner exchange. | 14:46.53 |
| kens2 | thought Glasgo was European city of culture back in 1990 | 14:46.55 |
| | Robin_Watts : then he'll need a passport to get back | 14:47.09 |
| Robin_Watts | kens2: It either is now, or just was. | 14:47.17 |
| kens2 | But at least not a new one | 14:47.18 |
| jogux | Glasgow has the longest shopping street in the UK, or something. | 14:47.19 |
| | we also have the friendliest junkies | 14:47.35 |
| Robin_Watts | and the highest proportion of unicycling bagpiper darth vaders. | 14:48.05 |
| chrisl_r500 | Naybe I should encourage my mother to hang onto my grandmother's old flat in Edinburgh, at least until October....... | 14:48.05 |
| | Or even "Maybe" | 14:48.21 |
| Robin_Watts | chrisl_r500: As an offshore tax haven, you can do better... :) | 14:48.34 |
| kens2 | I wonder if I can get dual citizenship.... | 14:48.53 |
| chrisl_r500 | I was thinking I could have dual passports.... or even move back | 14:49.01 |
| pedro_mac | chrisl_r500: depends where it is of course :) | 14:49.06 |
| jogux | an Irish person living in Scotland may end up entitled to 4+ different passports :) | 14:49.13 |
| kens2 | Apparently my wife is entitled to apply for an Irish passport.... | 14:49.52 |
| pedro_mac | doesn't get why its so radically different to any other commonwealth country | 14:50.01 |
| Robin_Watts | Anyone can *apply* for an irish passport... | 14:50.10 |
| pedro_mac | too much nonsense going on | 14:50.11 |
| chrisl_r500 | pedro_mac: it's in Musselburgh - a dire place, if ever I saw one! | 14:50.23 |
| kens2 | Robin_Watts : OK but she could actually get one | 14:50.30 |
| | Hmm, intriguign bmpcmp, I must test the original files now | 14:50.58 |
| pedro_mac | chrisl_r500: still pretty popular place though, and you have a beach (cold)! and Luca's ice cream parlour. It just gets better | 14:51.03 |
| kens2 | The beach is nice | 14:51.43 |
| | I seem to remember it was in a film too, though I can't recall which one. Myabe Gregory's Girl | 14:52.15 |
| chrisl_r500 | Portobello is nice, and has a nice beach - that's where I grew up. Musselburgh doesn't have much of a beech since they filled up the water front with ash from Cockenzie power station | 14:53.16 |
| kens2 | Oh, that's changed since I was htere then | 14:53.51 |
| | It has been a long time | 14:53.55 |
| chrisl_r500 | And anywhere that calls itself the "Honest Toun" can't be good! | 14:54.00 |
| pedro_mac | is probably thinking of Portobello actually - been a while since I lived/worked in .ed ;) | 14:54.15 |
| | nope, definitely musselburugh, google tells me. | 14:55.17 |
| chrisl_r500 | The north side of Musselburgh is Fisherow, which is okay | 14:55.27 |
| kens2 | wonders how a change to pdfwrite can result in 34-all.ps rendering differently to pgmraw | 14:56.38 |
| | I think I'll ignore that one | 14:56.47 |
| chrisl_r500 | 34-all.ps has occasional indeterminisms | 14:57.09 |
| kens2 | Yeah bmpcmp showed no differences for the rendering case, but does ofr hte pdfwrite case. Now I don;t know if that's correct or not..... | 14:57.57 |
| | The other diff is definitely a progression though, so that's nice | 14:58.30 |
| chrisl_r500 | IIRC, 34-all.ps is one of the tests that pokes around where it shouldn't, so..... | 14:58.34 |
| kens2 | Oh well, in that case.... Given that it won;t run for me here because it tries to execute startjob | 14:58.57 |
| henrys | wow I just noticed you guys only have about 1500 hours of sunshine a year compared to 3000 here. Thatâs a big difference | 14:59.46 |
| pedro_mac | henrys: surely for Glasgow its closer to 150? ;) | 15:00.11 |
| kens2 | But plenty of rainfall | 15:00.30 |
| paulgardiner | jogux: the mutex assert - I think what's happening is that most of what needs testing by preShutdownFn is being achieved wholly at the UE2 level, until DocumentInfoView_isVisible (which calls UE2Panel_getVisible and bang). I could cache the visible state with a flag in DocumentInfoView, but I wonder how difficult it is to arrange for preShutdownFn to run within the schedular. | 15:05.51 |
| jogux | paulgardiner: I think one of the other functions might do it; the risk is that shutdown becomes unbounded-ish if we start letting the scheduler run whatever it wants. | 15:06.59 |
| | I don't know entirely why it matters if something is visible. surely visibiliity is irrelevant compared to unsaved changes or not. | 15:07.18 |
| paulgardiner | Something to do with checking whether the viewstate needs saving. | 15:08.15 |
| henrys | kens2: Iâll take catalinâs support question getting close to end of business for you. | 15:09.41 |
| paulgardiner | Probably if you in info view you know you have no unsaved state | 15:09.42 |
| | s/you/you're/ | 15:10.08 |
| jogux | paul: it does seem slightly curious. | 15:10.28 |
| paulgardiner | I'll see if I can find the commit that added the check for DocumentInfoView being visible | 15:11.10 |
| kens2 | henrys not seen it yet | 15:11.32 |
| | ah, ok, thanks henrys | 15:12.12 |
| henrys | kens2: it looks like a pcl issue anyway if - itâs not exclusive to pdfwrite | 15:14.09 |
| kens2 | henrys if display shows it, then its not pdfwrite :-) | 15:14.25 |
| henrys | s/if// | 15:14.33 |
| el_mendi_ | paulgardiner: I've downloaded this PDF doc: http://fzs.sve-mo.ba/sites/default/files/dokumenti-vijesti/sample.pdf and with this PDF when I insert an image, save the document and reopen it with MuPDF all the content of the page that is before the image remains, but everything else of the page disappears. But when I open the document with any other reader is like nothing happened. | 15:14.53 |
| paulgardiner | How are you inserting the image? | 15:16.24 |
| el_mendi_ | Like Robin_Watts told me, using the function pdf_clean_page_contents | 15:17.19 |
| paulgardiner | jogux: I think 3516e2a66828823f3a5c8eff13b0598a243486a2 may be the culprit | 15:17.26 |
| Robin_Watts | el_mendi_: What do you mean "all the content of the page that is before the image remains"? | 15:17.38 |
| el_mendi_ | Inserting the image in the buffer of the page with a pdf device | 15:17.44 |
| | wait, I'm going to make some screenshots | 15:17.58 |
| Robin_Watts | ok. | 15:18.08 |
| el_mendi_ | This is the original page: http://postimg.org/image/6c75ua36r/ | 15:22.17 |
| | I insert the image: http://postimg.org/image/qnj0gkvj7/ | 15:22.34 |
| | And this is what it saves: http://postimg.org/image/b9kqfe55v/ | 15:23.06 |
| Robin_Watts | el_mendi_: boggle. | 15:25.47 |
| | suppose you remove the code that inserts the image. Just have the clean_page_contents do nothing with the buffer. | 15:26.26 |
| | does that then save correctly? | 15:26.33 |
| el_mendi_ | I don't know | 15:27.40 |
| Robin_Watts | do that test. That will tell us if the problem is in the clean_page_contents (which is new logic and not thoroughly tested) or whether it's in the insertion of the image. | 15:28.32 |
| paulgardiner | jogux: I could back that out, I could try caching the visible state for DocumentInfoView or I could try doing the test within the schedular. What do you reckon? I'm thinking probably try option 2 first, if that turns out harder than it seems take option 1. | 15:30.46 |
| el_mendi_ | The same problem, I have called pdf_clean_page_contents(doc, pc->page, cookie, NULL, NULL); and when I save the document it is the same | 15:31.31 |
| | the content disappear | 15:32.08 |
| | but not all | 15:32.15 |
| Robin_Watts | el_mendi_: OK, so the problem is in the clean_page_contents call. Damn. | 15:33.00 |
| | Can you go to bugs.ghostscript.com and open a bug there please. Attach the file you are using, and the code changes, and I'll try to have a look at some point. | 15:33.46 |
| el_mendi_ | Ok | 15:34.14 |
| jogux | paulgardiner: I'm still not at the stage of understanding why visibility matters! | 15:34.52 |
| el_mendi_ | Robin_Watts: with the file I am using what do you mean? | 15:39.29 |
| Robin_Watts | el_mendi_: The PDF file you are trying to annotate. | 15:39.49 |
| paulgardiner | It's refusing to shutdown if the viewstate would need saving (i.e. which page you are looking at and where on the page). Not sure why you'd want that come to think of it. | 15:39.50 |
| el_mendi_ | The code I suppose is the function clean_page_contents no? | 15:40.01 |
| paulgardiner | jogux: ^ | 15:40.54 |
| Robin_Watts | el_mendi_: Basically, I want you to give me everything in the bug that I need to reproduce the problem locally. | 15:41.14 |
| jogux | paulgardiner: err. interesting. | 15:41.18 |
| el_mendi_ | Ok | 15:41.39 |
| jogux | paulgardiner: I guess that sort of makes sense ish, in terms of being able to reopen to where the user currently is. | 15:41.41 |
| Robin_Watts | I'm too busy to devote much time to this, so if you want me to help you out you need to make it really easy for me. | 15:41.51 |
| jogux | paulgardiner: and the state database is new, so that would explain why it's a new issue I guess | 15:42.07 |
| paulgardiner | jogux: but wouldn't you allow shutdown and just make sure to save the viewstate when you do. It's not like you need to ask the user | 15:43.18 |
| mattchz | bye | 15:43.32 |
| jogux | paulgardiner: shutting down would dispose of the state :-) | 15:43.35 |
| paulgardiner | So where is the save of the viewstate done? | 15:44.06 |
| jogux | somewhere in ue2fileviewer I think | 15:44.20 |
| | presumably somewhere in the stacktrace of your assert | 15:44.29 |
| | paulgardiner: I agree with your point about not needing to ask the user. | 15:47.01 |
| paulgardiner | I'm lost. I would have thought if we were able to save the viewstate of the previous document as part of loading a new document, we wouldn't need to refuse shutdown if the viewstate needs saving - if that makes any sense | 15:47.05 |
| jogux | actually thinking about it I'm also confused about why we're trying to shutdown. | 15:47.46 |
| paulgardiner | So I wouldn't expect to see the saving in or around my stack trace unless the test is unnecessary | 15:47.51 |
| | :-) | 15:48.02 |
| jogux | paulgardiner: yes, true, that seems a reasonable expectation. | 15:48.47 |
| | perhaps we're only shutting down as a way to make sure unsaved changes to the doc are handled? | 15:49.19 |
| | I think there's a limit to how much longer I can talk about this without actually going to look at the code :) | 15:49.38 |
| paulgardiner | :-) | 15:49.49 |
| jogux | I'm afraid to stop looking at Robin's code in case in stops working | 15:50.14 |
| paulgardiner | No worries. Let me bash on at it a while | 15:50.41 |
| el_mendi_ | Robin_Watts: All the code from the android part too? | 15:50.49 |
| Robin_Watts | el_mendi_: Yes please. | 15:51.00 |
| | It may be easiest to just zip everything up and attach it. | 15:51.17 |
| el_mendi_ | yes, that's what I was thinking | 15:51.49 |
| Robin_Watts | You're not using any silly IDE or anything are you? | 15:52.02 |
| jogux | robin_watts : hm, yes, there is definitely something going on with this patch that's breaking some things. Argh. | 15:57.42 |
| paulgardiner | jogux: from the commit message, is seems shutdown is being interrupted so as to save the viewstate and then shutdown is retried. The test for document-info-view being visible may be an optimisation - no need to save view state if doc-info-view is open and hence no need to interrupt the shutdown | 16:01.27 |
| | s/is/it/ | 16:01.35 |
| jogux | that sounds logical. | 16:01.50 |
| paulgardiner | jogux: I wounder whether it is harmful to save the view state even when the doc-info-view is open. If not I could just remove that test. | 16:03.32 |
| jogux | what's the review number? | 16:03.50 |
| paulgardiner | 3516e2a66828823f3a5c8eff13b0598a243486a2 | 16:04.09 |
| | Oh 95812 | 16:04.22 |
| pedro_mac | I was thinking theyâd done a lot of additional reviews after we left ;) | 16:04.42 |
| jogux | paulgardiner: I like how the review comments end with "this is becoming more and more crazy" | 16:11.17 |
| paulgardiner | Yeah. Eeek! Multidocs | 16:15.05 |
| jogux | sssh :) | 16:15.16 |
| | (none of the customers get multidocs afaik ;-) ) | 16:15.42 |
| Robin_Watts | I didn't think multidocs was ready yet ? | 16:16.48 |
| jogux | robin_watts: aiui it might be working but is only tested on winrt within SOL. maybe. | 16:17.12 |
| | paul: it does seem like it would be possible to update saveViewStateOnShutDn when changing views | 16:18.53 |
| | the whole querySaveViewStateOnShutDn function feels pretty evil | 16:20.20 |
| el_mendi_ | Robin_Watts: The ADT for Eclipse | 16:20.51 |
| Robin_Watts | Ah, well, there is precisely 0% chance of me installing eclipse to look at the bug. | 16:21.15 |
| jogux | mattchz may well have eclipse setup if that helps. | 16:21.58 |
| Robin_Watts | but I can try to run it through stuff here. | 16:21.59 |
| el_mendi_ | Ok | 16:22.13 |
| paulgardiner | jogux: not sure I understand. Do you mean keep saveViewStateOnShutDn up to date to reflect the full complexity of what querySaveViewStateOnShutDn calculates, so that that function can be changed to return the flag? | 16:22.21 |
| jogux | paulgardiner: that was what I was thinking, yes. | 16:22.36 |
| pedro_mac | Robin_Watts: presumably can still just make with ant? | 16:22.42 |
| paulgardiner | jogux: what about the option of just removing the call to DocumentInfoView_isVisible? | 16:22.52 |
| | I imagine it's unnecessary. | 16:23.04 |
| Robin_Watts | pedro_mac: Yes, if I can make it with the usual build commands there is no problem. | 16:23.08 |
| jogux | paulgardiner: I don't understand why that would be necessary, so yeah, why not remove :) | 16:23.15 |
| paulgardiner | I'll give that a try | 16:23.29 |
| Robin_Watts | It's just that I've been bitten before by people giving me stuff and saying "oh, but I imported everything into Eclipse, so you'll need to use that..." | 16:23.37 |
| pedro_mac | nods - can be a bit of a nightmare | 16:24.14 |
| pedro_mac | also has it set up here, for different reasons ;) | 16:24.34 |
| el_mendi_ | But I don't know why the contents of the page are updated fine to redraw the page but not to save the document... | 16:25.36 |
| Robin_Watts | el_mendi_: That is a good question. | 16:31.06 |
| | henrys: Your latest customer query about splitting the first 300 pages from a PDF... | 16:34.00 |
| | Surely that's a classic case of "use mupdf" ? | 16:34.11 |
| | mutool clean -difggg in.pdf out.pdf 1-300 | 16:34.40 |
| kens2 | agrees wholheartedly with that one. | 16:35.06 |
| Robin_Watts | oops: mutool clean -ggg in.pdf out.pdf 1-300 | 16:35.15 |
| | No need to decompress. | 16:35.21 |
| | and possibly can get away with -g or -gg. | 16:35.37 |
| kens2 | pdfwrite is the wrong tool for htis job | 16:35.40 |
| henrys | Robin_Watts: ah true good point - nonetheless it shouldnât crash Iâll recommend mutool now | 16:37.22 |
| kens2 | henrys 'crash' to a customer can mean 'exits with error' | 16:37.44 |
| Robin_Watts | henrys: Unless I hear different, I'm going to assume the EmbeddingMuPDF document is fine :) | 16:37.58 |
| henrys | Robin_Watts: I enjoyed it ;-) | 16:38.15 |
| Robin_Watts | jogux, pedro_mac: I have a docx file here with a table defined in it that is given a specific Y coord to appear at (the idea being that other content should flow around it) | 16:43.06 |
| jogux | saw | 16:43.13 |
| Robin_Watts | Looking in epage, it looks as though the word agent respects tblpX, but not tblpY. | 16:43.30 |
| pedro_mac | Robin_Watts: yup, saw the bug | 16:43.41 |
| Robin_Watts | Do you know of any Edr limitations that might explain that ? | 16:43.43 |
| jogux | robin_watts : can you review first review on my master? just a new ATH script as a stick in the ground for this doc - http://intranet.picsel.com/~ats/cgi-bin/results.pl?grid=41 | 16:44.02 |
| Robin_Watts | i.e. is it just the case of "never saw a file that needed it", or is it "ooh, we can't do that..." | 16:44.05 |
| jogux | robin_watts: hm. unsure. if it was an image I think it would work. | 16:44.27 |
| Robin_Watts | fetches now. | 16:44.42 |
| | I'll give it a whirl and see. | 16:44.50 |
| jogux | any issue with a non-image type may be related to layout needing to figure out it's size first I guess. | 16:45.41 |
| Robin_Watts | ATH test script looks good. | 16:46.12 |
| | were there more reviews ready to go there ? | 16:46.22 |
| jogux | robin_watts: the rest are basically ready to go pending a bit more testing, but everyone ignored me when I said so :-) | 16:47.11 |
| Robin_Watts | The gitignore one seems a no brainer. | 16:47.33 |
| | The unused/thread stack code one... | 16:48.04 |
| | in the commit message you probably mean "no reason not to" | 16:48.23 |
| pedro_mac | Robin_Watts: do you know which anchor type its relative to? (paragraph/page/margin?) - I canât say for sure but I think the support isnât complete for anchors in general, and vertical would certainly be less likely to be supported than horizontal given how we flow content if its paragraph-relative. joguxâs comment about tables also makes sense; I think most of the absolute positioning Iâve seen relates to embedded images/containers, | 16:48.47 |
| | structured content like tables | 16:48.48 |
| Robin_Watts | margin, I think. | 16:49.20 |
| jogux | robin_watts: I do. | 16:49.47 |
| Robin_Watts | otherwise that commit looks plausible. | 16:51.34 |
| jogux | I really can't think of any advantage of soft threads on any modern OS. | 16:52.16 |
| Robin_Watts | ios8 compiler error seems fine too. (Really Apple?) | 16:52.45 |
| | All the others look fine, except a minor thing with the last one... | 16:56.29 |
| | The comment says "this file can be deleted", but I bet we'll get customers that delete it not through xcode, and are then left with the project trying to find the deleted file. | 16:57.17 |
| jogux | okay, so rephrase as can be removed from project? | 16:59.01 |
| Robin_Watts | That'd be perfect. | 16:59.10 |
| jogux | I think I need to recreate this sample xcode project from scratch; other than the fact that it builds and works I can find very little else that's actually correct about it :-S | 17:04.03 |
| pedro_mac | Robin_Watts: when I initially looked at that bug I think I was at the point where I reckoned it just handled tablepX because that was the easy thing (mapping to indentation of the rows) | 17:06.35 |
| Robin_Watts | pedro_mac: I've just updated the bug with what we know, and I'm moving on... | 17:07.00 |
| jogux | robin_watts : okay, both those issues fixed, I tested printing and it didn't blow up, okay to push to golden? | 17:07.17 |
| pedro_mac | farynuff | 17:07.18 |
| Robin_Watts | jogux: Sure. | 17:07.29 |
| pedro_mac | ok, Iâm off - have a good weekend all | 17:11.58 |
| Robin_Watts | you too! | 17:12.07 |
| jogux | robin_watts: thanks! pushed | 17:18.27 |
| | Forward 1 day (to 2014/06/21)>>> | |
IRC Logs
