| <<<Back 1 day (to 2014/06/29) | 2014/06/30 |
pedro_mac | hi folks | 06:58.44 |
kens | Morning | 06:58.48 |
Robin_Watts | morning | 07:44.40 |
pedro_mac | hi Robin | 07:51.33 |
jogux | paulgardiner: the good iOS build is allegedly failing in ATS with platform/tgv/apps/iphone-alien/classes/iphone_alienAppDelegate.m:332:16: error: instance method '-configureUIWithLogo:bundle:color:' not found (return type defaults to 'id') [-Werror,-Wobjc-method-access] | 09:02.21 |
| [gdLibrary configureUIWithLogo:@"good-login-pso.png" - any ideas if it is actually broken, sdk is old, ... ? | 09:02.22 |
paulgardiner | Did I ever commit my sdk-upgrade changes? | 09:03.47 |
| Was just looking this morning and seemed like I hadn't. I remember requesting review and I thought I got feedback. | 09:04.37 |
| If I did then we would need ATS to be using the latest SDK | 09:05.16 |
jogux | I /think/ ATS is using the latest sdk. | 09:05.34 |
| so maybe you haven't. | 09:05.42 |
| hm, no, there's an update commit on master. | 09:06.53 |
| how recently did the sadk change? :) | 09:07.00 |
paulgardiner | I did take on 1.8 quite rencently | 09:10.51 |
| That's probably it. | 09:10.58 |
jogux | is sure I installed this after the heartbleed thing. but maybe I managed to get the wrong version :-S | 09:12.44 |
| I can't actually figure out what version it is from the install | 09:13.12 |
paulgardiner | I think I'm using a later version than we were asked to for the heartbleed thing | 09:13.23 |
| I think my last commit says exactly which sdk to use | 09:13.52 |
| 1.8.xxxx where xxxx is something like 4564 | 09:14.11 |
| but only something like that. :-) | 09:14.25 |
jogux | yeah, that seems to be the latest on their website. just trying installed that in ATS. | 09:16.34 |
paulgardiner | Gah! You know that topmenu animation fix I was working on friday. I thought I'd just check the branch. Turns out the branch has the code I removed in my patch but the topmenu moves correctly. :-( | 09:17.37 |
| I suppose I need to track down how it's been made to work correctly on the branch. Do you think? | 09:18.49 |
jogux | :) I'm not sure it's worth the time, the answer may just be "stuff is different". | 09:21.10 |
| the branch diverged from master > 18 months ago | 09:21.24 |
| Your fix seemed to work well in testing | 09:21.41 |
paulgardiner | I will think of it not one moment longer. :-) | 09:23.11 |
xzibit68 | Hi, all | 09:30.16 |
| What a cookie added in android JNI? | 09:31.55 |
jogux | paulgardiner: upgrading the sdk fixed it. I'd obivously not installed what I thought I had :-S | 09:33.28 |
paulgardiner | excellent | 09:34.35 |
mattchz | morning | 09:35.39 |
paulgardiner | xzibit68: it's a mechanism for fast aborting a render operation being done on a background thread | 09:35.43 |
tor8 | morning matt | 09:36.15 |
paulgardiner | Hi matt | 09:37.45 |
| hi tir | 09:37.47 |
| Or Tor even | 09:37.59 |
xzibit68 | paulgardiner: Thank you! | 09:38.04 |
mattchz | :-D | 09:38.10 |
kens | mattchz : I have some information for you regarding annotations, especially QuadPoints | 09:38.33 |
mattchz | oh, cool, thanks! | 09:38.45 |
kens | The problem is that MuPDF is following the spec. | 09:38.57 |
| Acrobat doesn't..... | 09:39.03 |
mattchz | nice | 09:39.13 |
kens | THe spec says the QuadPoints are given in anti-clockwise order round the vertices of a quadrilateral, Acrobat reads tehm in a different order. | 09:39.34 |
| Technically you cn start anywhere, but GS assumes you will start with the lower left corner | 09:39.54 |
| The next vertex is the lower right corner. Then we jump to the upper left, tehn the upper right | 09:40.14 |
| If you alter the QuadPoints like that, GS renders the highlight 'properly' | 09:40.35 |
| I have 2 other observations though | 09:40.44 |
| 1) THe annotation has no colour specified, even though the appearance stream uses yellow (ie /C [1 1 0]) | 09:41.12 |
jogux | paulgardiner: there's a fairly trivial fix for simulator builds on my master. | 09:41.18 |
kens | You should add the colour to the annotation as well. | 09:41.23 |
| GS defaults to black which hides the text, Acrobat defaults to white so you don;t see the highlight..... | 09:41.47 |
paulgardiner | jogux: okay will look in a couple of minutes | 09:41.53 |
jogux | ta, no urgent. | 09:42.00 |
kens | Finally the Appearance stream doesn't use rounded caps, but that's an aesthetic, not a requirement. | 09:42.14 |
| mattchz : I'm assuming changing the order the QuadPoints are written should be easy, and it should be fairly trivial to add the /C for the colour specification. | 09:43.24 |
Robin_Watts | xzibit68: More specifically the cookie is a way of cheaply passing information into/out of a render as it's going on. | 09:43.32 |
| Passing an 'abort' flag in is one thing it can do. | 09:43.46 |
| Getting progress information for a render is another. | 09:43.57 |
mattchz | kens: cool - thatâs really helpful thanks | 09:48.27 |
kens | NP | 09:48.33 |
mattchz | I wonder if thatâs something paulgardiner wants to look at, or if I should (Iâm supposed to be really looking at Android/iOS specific bugs here) | 09:48.47 |
kens | No idea, it seems core to me, but on the other hand, its probably easy to fix | 09:49.10 |
mattchz | my PDF knowledge is virtually zero, although Iâm happy to stab around randomly if necessary :) | 09:49.19 |
kens | I would imagine its easy to find and sort, though I admit to no actual knowledge..... | 09:49.42 |
Robin_Watts | mattchz: I'm sure paulgardiner has already changed this a few times. | 09:49.57 |
kens | Wherever the QuadPoints are written just change the order of teh operands, and add a /C with [1 0 0] as its operand | 09:50.26 |
| Err sorry [1 1 0] | 09:50.35 |
xzibit68 | Robin_Watts: is cool, thank you! | 09:50.55 |
mattchz | perhaps I should leave this to paulgardiner then; I donât want to make things worse. | 09:51.04 |
| although Iâm starting to run out of mobile specific bugs anyway, I think. | 09:51.25 |
jogux | mattchz: worth looking all this on the bug, if it sounds easy enough we can check with henrys if he wants us to look at it after all the mobile bugs are done | 09:51.56 |
Robin_Watts | jogux: yeah. | 09:52.12 |
mattchz | nod | 09:52.13 |
paulgardiner | Which problem is this? | 09:52.23 |
jogux | worth *logging* all this on the bug | 09:52.26 |
mattchz | Iâll post a link to the IRC logs on the bug | 09:52.42 |
kens | bug 695130 ? | 09:52.43 |
paulgardiner | jogux: doesn't your commit inhibit the output of the arch line in non-sim builds? | 09:53.02 |
| mattchz: right. The spec is wrong. | 09:54.26 |
kens | would prefer to say that Acrobat is wrong, the spec makes more sense.... | 09:54.49 |
paulgardiner | true | 09:55.03 |
| I have half a patch somewhere | 09:56.02 |
mattchz | impressive, as adobe wrote both. | 09:56.05 |
kens | Not uncommon in PDF | 09:56.23 |
jogux | paulgardiner: only if I've done something massively stupid | 09:57.52 |
paulgardiner | reviewer stupidity sounds more likely. I'll take another look | 09:58.31 |
jogux | which is possible, I think there's tabs in there. | 09:58.38 |
| hm. no. gitk was trying to confuse me. no tabs afaics. | 09:59.49 |
paulgardiner | Right. I get it now. Looks fine | 10:03.17 |
jogux | double checks it and spots another existing bug. No n on the end. May as well fix that (though it works) | 10:05.45 |
paulgardiner | I've rebased my menu fix after your build fix so I could push them to golden together? | 10:06.31 |
jogux | would probably rather fix my missing n :) | 10:07.31 |
| there's a new version of my commit up now that fixes that. I'm happy with whatever approach you take :) | 10:08.37 |
paulgardiner | Okay. I'll push them | 10:12.02 |
| tor8: I had a look at generating a 1.5 iOS release for testflight but your user certificate is out of date | 10:18.37 |
tor8 | paulgardiner: okay. what do I need to do? | 10:19.18 |
jogux | paulgardiner: good time to transfer it to the artfiex account then? :-) | 10:19.56 |
| paul: thanks! | 10:20.21 |
paulgardiner | jogux: yeah maybe. I guess we aren't in that much hurry to get the 1.5 app out. | 10:21.05 |
jogux | the transfer should be very quick, it will take less time than renewing the certificates I'd bet :-) | 10:21.28 |
paulgardiner | tor8: I think you just log in click on the certificates section and click on the plus in the top right. | 10:21.32 |
| jogux, tor8: maybe I'll have a go at the transfer then | 10:21.54 |
jogux | matt could do it? | 10:22.02 |
tor8 | paulgardiner: I'd say go for the transfer | 10:22.03 |
paulgardiner | jogux: even better | 10:22.10 |
jogux | paulgardiner: can you give him login details for itunesconnect for the current app? | 10:22.32 |
tor8 | paulgardiner: I believe you got the crucial fixes in for the 1.4-ios tag, so a 1.5 isn't as crucial as it was for the other platforms | 10:22.43 |
paulgardiner | tor8: true. | 10:23.23 |
| jogux: I'll email them now. | 10:23.36 |
mattchz | thanks | 10:32.39 |
| if this is a test flight release, why do we need to transfer the app or change the certificate? | 10:38.30 |
jogux | transferring the app won't hold up the test flight release actually, so yeah :) | 10:42.37 |
| should the 1.5 release include the fixes matt has been doing? (ie. come from master?) | 10:44.57 |
| paulgardiner / tor8 : ^^ | 10:45.10 |
tor8 | jogux: might as well, I think | 10:47.02 |
mattchz | Ah,we have a 1.4 and 1.4-ios tag anyway, so I can tag this 1.5-ios | 10:49.04 |
Robin_Watts | jogux, mattchz, paulgardiner, tor8: If matt is getting close to completing all the mobile specific bugs, I'd like to see him take on the JNI interface to MuPDF and modify the android app to use that. But that's a call for henrys/miles really. We can bring it up at the meeting tomorrow. | 10:50.14 |
tor8 | Robin_Watts: I agree. | 10:50.36 |
mattchz | there are still bugs, but some of them Iâm having trouble reproducing/getting more info about. | 10:52.06 |
| and some are probably in core territory (e.g. the annotations) | 10:52.17 |
tor8 | I reckon I'm supposed to take over the annotation core bugs from paul | 10:53.12 |
Robin_Watts | If we can't reproduce bugs and they won't give us more info, then I think we don't worry about them too much, cos they will probably either be user error, or because they are using an old version of the app. | 10:53.43 |
mattchz | http://bugs.ghostscript.com/show_bug.cgi?id=695271 <â sounds like a generic bug â can I remove the android tag? | 10:55.10 |
tor8 | mattchz: yeah. | 10:56.28 |
mattchz | Ok, Iâll go through the bugs now, and we can decide which ones are generic. | 10:56.44 |
| http://bugs.ghostscript.com/show_bug.cgi?id=695324 <â core too, I think. | 11:00.20 |
mstefanko | hey guys is there a saas service where you can send your PDF to, and they will send you all pages back in image format + page titles? | 11:01.53 |
tor8 | mattchz: yeah. that one too. | 11:03.07 |
mattchz | http://bugs.ghostscript.com/show_bug.cgi?id=695233 too. | 11:03.17 |
tor8 | mattchz: that one is harder to test, we don't have an annotation creation tool that is not on a mobile app :( | 11:03.50 |
| which is one of my higher priority todo items... | 11:04.15 |
mattchz | are the quadpoints etc created in the mobile app? | 11:04.31 |
tor8 | it's in the core, but only exercised by the mobile app | 11:04.44 |
| I've talked with Robin about making a command line tool to add annotations though | 11:04.56 |
| just so we can have something easy to test | 11:05.02 |
mattchz | nods | 11:05.07 |
| Iâm tempted to say we should remove the tag there, if we think it is core related. | 11:05.24 |
tor8 | mattchz: yes. remove the android tag. | 11:05.38 |
mattchz | I think I did the same for the quadpoints bug kens was talking about too. | 11:06.05 |
paulgardiner | mattchz: the apps create the quadpoints, but there is one other place that needs attention which is where we synthesize an appearance stream | 11:08.42 |
mattchz | right, cool. | 11:09.15 |
| tbh, the removing of tags is really just to help me see the wood for the trees. | 11:09.30 |
| There are also various enhancement requests. | 11:09.58 |
| I donât know if those need triaging. | 11:10.12 |
| http://bugs.ghostscript.com/buglist.cgi?bug_status=UNCONFIRMED&bug_status=CONFIRMED&bug_status=IN_PROGRESS&columnlist=product%2Ccomponent%2Cassigned_to%2Cbug_status%2Cresolution%2Cshort_desc%2Cchangeddate%2Cpriority&list_id=15805&op_sys=iOS%204&op_sys=iOS%205&op_sys=iOS%206&op_sys=iOS%207&op_sys=Android&product=MuPDF&query_format=advanced | 11:10.51 |
| Buglist looks a lot smaller now. 19 bugs, of which 11 are feature requests. | 11:11.18 |
tor8 | mattchz: yeah, a bit more manageable now | 11:11.48 |
paulgardiner | pdf_set_markup_appearance is the call. So the apps need updating to generate the quadpoints in Acrobat order and pdf_set_markup_appearance needs updating to expect them in Acrobat order. | 11:12.24 |
tor8 | mattchz: I've got some minor code style complaints found while I was looking over your changes. Robin should have spotted them, but he and I don't always agree | 11:12.58 |
mattchz | no problem. | 11:13.13 |
| let me know, and Iâll try and fix them. | 11:13.24 |
tor8 | mattchz: in the commit that shrinks the store, you've (a) camelCase on a local variable, and (b) declared it in the middle of the function (c99-style) | 11:13.33 |
Robin_Watts | oops. Missing b is particularly bad of me. | 11:14.07 |
mattchz | (b) was bad of me too, because I tried to avoid doing it :) | 11:14.21 |
tor8 | I don't personally care about (a), but then I don't care where you put your braces either (K&R and Allman style are both good to me) | 11:14.42 |
| but other people care about consistency so I try to enforce the rules we have :) | 11:14.56 |
Robin_Watts | just not gnu :) | 11:14.58 |
mattchz | Is C99 in general banned? We should perhaps make the Android/iOS compilers choke on it. | 11:15.03 |
Robin_Watts | mattchz: We try to avoid C99 just so we support as many toolchains as possible. | 11:15.32 |
mattchz | is there a style guide/coding standards description somewhere? | 11:15.32 |
tor8 | we try to stick to C90 with the most common extensions that generally work everywhere | 11:15.33 |
| MSVC chokes on c99 | 11:15.42 |
mattchz | in which case, itâd probably make sense for me to try and make the android build fail on C99isms | 11:16.01 |
Robin_Watts | mattchz: The android jni stuff may rely on C99 in headers maybe? | 11:16.33 |
tor8 | mattchz: docs/naming.txt is the closest we have to a style guide | 11:16.39 |
Robin_Watts | but if you can keep it happy, then great. | 11:16.55 |
mattchz | robin_watts: oh, possibly. | 11:16.59 |
| I guess we managed to do that at picsel by abstracting all the standard lib functions, and not including the system headers. | 11:17.29 |
Robin_Watts | Stupid right to left tabbing. | 11:17.31 |
| mattchz: yeah. Something that comes back to bite us regularly. | 11:17.50 |
kens | heads to lunch | 11:18.40 |
tor8 | mattchz: but the short rules are -- underscore_names, tab for indent, don't align stuff past the indent | 11:18.52 |
mattchz | nod, ta. | 11:19.41 |
tor8 | the last of those rules infuriates robin to no end :) | 11:19.45 |
mattchz | can you give me an example of the last one? | 11:20.05 |
tor8 | vertically aligning stuff after '=' sign or in variable declaration lists, etc | 11:20.21 |
mattchz | ah, okay. | 11:20.31 |
tor8 | or trailing line comments, stuff like that | 11:20.41 |
mattchz | So: | 11:20.58 |
Robin_Watts | mattchz: It's because tor8 insists on using proportional fonts in his editor. | 11:21.03 |
tor8 | it doesn't work reliably with variable tab stops, makes a mess when editing and you have to go and realign everything, and doesn't work with proportional fonts | 11:21.11 |
Robin_Watts | So if he can't have nice looking source, why should anyone else? :) | 11:21.13 |
tor8 | and adds diff noise | 11:21.16 |
mattchz | const char *str = âsome long stringâ; | 11:21.16 |
| rather than | 11:21.18 |
| const char *str = | 11:21.21 |
| âsome long stringâ; | 11:21.25 |
tor8 | if you need to break a line, just indent the continuation by one extra indent | 11:21.53 |
mattchz | oh, I think I see what you mean. | 11:22.16 |
tor8 | a[10] = "x"; | 11:22.18 |
mattchz | int a = b; | 11:22.20 |
tor8 | a[5] = (two spaces) "y"; | 11:22.28 |
| things like that | 11:22.32 |
mattchz | float f = c; | 11:22.32 |
| rather than | 11:22.34 |
| int a = 10 | 11:22.41 |
| float b = 10.0f; | 11:22.47 |
| or something. | 11:22.50 |
| Iâm wittering now, sorry. | 11:22.55 |
tor8 | mattchz: yeah. | 11:22.56 |
| that's it :) | 11:22.58 |
| that sort of code drives me nuts when I have to edit it | 11:23.14 |
| and read it, when I'm using Acme and my favorite (proportional) font | 11:23.27 |
| Robin_Watts: well, I don't use syntax coloring either... so if the code looks good with proportional fonts and no color, it looks good everywhere ;) | 11:24.24 |
mattchz | #religiouswars | 11:24.43 |
tor8 | mattchz: haven't had one in far too long :) how else to start back at work after vacation? | 11:25.09 |
mattchz | hehe. vi or emacs? | 11:25.35 |
tor8 | mattchz: neither: acme. (well, I do use vim quite a lot too. the vi command set is embedded in my genes by now...) | 11:26.23 |
Robin_Watts | acdgacdgacdg:wq | 11:27.37 |
tor8 | mattchz: sadly, acme requires a proper three button mouse to work as intended... | 11:27.53 |
| Robin_Watts: <Ctl-C>:qa!<Enter> | 11:28.38 |
Robin_Watts | C-x C-c | 11:29.01 |
tor8 | <Alt-Tab>killall -9 emacs | 11:29.20 |
| fortunately I've stocked up on http://www.amazon.com/HP-Optical-Button-Mouse-accessory/dp/B0002Y5LZ8 | 11:30.29 |
paulgardiner | jogux: appstore-ios-so build is failing at the point of code signing. I'm sure this has worked before | 11:30.55 |
mattchz | Iâd never heard of acme until now. | 11:30.59 |
tor8 | http://research.swtch.com/acme has a nice introduction | 11:31.27 |
jogux | paulgardiner: working fine in ATS ;-) | 11:31.27 |
mattchz | Iâve started using SublimeText recently. Itâs nice. | 11:31.32 |
tor8 | mattchz: yeah, it's nice but you don't get the source so you can't change it if you want to :( | 11:32.07 |
jogux | paulgardiner: how does it fail? I'm guessing either the private key or the provisioning profile are missing, or it's picked up the wrong key. | 11:32.37 |
tor8 | mattchz: the 'minimap' scrubber is a really neat feature in sublime | 11:32.54 |
mattchz | yeah. the python extensibility seems really cool too. | 11:33.11 |
paulgardiner | It's looking for "iPhone Distribution: Artifex Software Inc. (7M9658FA5G)" which is not what our profile is called | 11:33.23 |
mattchz | not that Iâve used it, but I like the idea in principle. It seems like a modern day reimagining of emacs | 11:33.41 |
paulgardiner | We have "Artifex AdHoc Wildcard App Id" | 11:34.02 |
| I was looking for what determines this in iphone_postlink.py but I'm struggling | 11:35.08 |
jogux | I'm not sure that's in postlink.py anywhere, probably just in the xcode project | 11:35.43 |
paulgardiner | Yeah | 11:35.53 |
| I'm confused about what's changed | 11:36.06 |
mattchz | so, are the iOS testflight builds done in the GUI, and manually uploaded? | 11:36.08 |
| (muPDF) | 11:36.13 |
| Or is there some kind of magic script? | 11:36.22 |
Robin_Watts | watches acme video; gets to the scrollbar explaination and runs screaming. | 11:36.33 |
paulgardiner | mattchz: done in the gui | 11:36.35 |
mattchz | âk. | 11:36.44 |
| for other projects, weâve used a shell script which does the build then calls curl to automatically upload it to testflight. | 11:37.25 |
jogux | paulgardiner: the 7M9658 thing is the name of the key, not the profile | 11:37.38 |
mattchz | itâs a nice thing, although I guess it wonât help here because some people use windows to build. | 11:37.45 |
jogux | how many religuous wars are you planning to start today Matt? :-) | 11:38.52 |
mattchz | not intenional :) | 11:39.02 |
| (for a change) | 11:39.06 |
| did you send me the login details btw? | 11:39.11 |
jogux | paul said he did | 11:39.19 |
paulgardiner | jogux: I may deduce a consequence of that statement regarding keys at some point, but it hasn't happened yet. :-) | 11:39.57 |
jogux | paulgardiner: what's the actual error? | 11:39.57 |
mattchz | ah right. paulgardiner: where did you send them> | 11:40.12 |
| ? | 11:40.13 |
paulgardiner | matt.holgate@emobix.co.uk | 11:40.36 |
mattchz | ah, itâs matt@ | 11:40.42 |
paulgardiner | okay resending | 11:41.12 |
Robin_Watts | would suggest that first.last@ would be a good set of mail aliases for emobix to have :) | 11:41.27 |
jogux | paulgardiner: is it possible you're missing the private key? though I can't imagine how :) | 11:41.34 |
mattchz | paul> thanks | 11:41.36 |
paulgardiner | error is "Code Sign error: No matching provisioning profiles found: This product type must be built using a provisioning profile, however no provisioning profile matching the identity âiPhone Distribution: Artifex Software Inc. (7M9658FA5G)â was found. | 11:42.16 |
| CodeSign error: code signing is required for product type 'Application' in SDK 'iOS 7.1'" | 11:42.18 |
jogux | sounds like you're missing the provisioning profile somehow. | 11:42.48 |
mattchz | hmm, I guess I need inviting to the Artifex development team too? | 11:42.59 |
| (iOS) | 11:43.03 |
jogux | try going to preferences, accounts, artifex, then hit refresh | 11:43.10 |
paulgardiner | Yeah, but we have no profile of that name. Is it just the 7M96... that is important? | 11:43.21 |
henrys | paulgardiner, pedro_mac : we have a new contact at Good who is going to fix everything ;-) Heâs from Apple, I didnât ask him why Apple got rid of him. Anyway if you can think of anything other than talking to the engineers let me know. | 11:43.44 |
jogux | paul: it's "I don't find a provisioning profile that can be signed by key <x>" rather than "I can't find a provisioning profile called <x>" | 11:43.45 |
| matt: invitation sent | 11:45.15 |
mattchz | thanks. | 11:45.27 |
pedro_mac | henrys: that sounds excellent :) I have one query which their support folks seem to be slow and/or ignoring regarding the android activity launch type which I still havenât resolved with them, so that may be something we can escalate if they donât get back to me today | 11:46.00 |
| henrys: obviously thatâs something weâd probably be able to sort by talking with an engineer ⦠;) | 11:46.40 |
mattchz | hmm. | 11:47.00 |
| paulgardiner: the itunes connect details you sent donât seem to work for me. | 11:47.14 |
paulgardiner | mattchz: strange! I seem to be able to sign in with them | 11:49.28 |
mattchz | oh, odd. It worked when I typed the password in rather cânâp it. | 11:50.16 |
| so, I presume you want me to transfer this to the artfiex account then? Please could you send me details fo that? | 11:51.23 |
| I wonder if I need to resign the build. | 11:51.43 |
mstefanko | anybody know of any pdf to png+content APIs ? | 11:52.06 |
paulgardiner | mattchz: You seem to be a member of our team already | 11:53.28 |
mattchz | jogux just added me. But I think I will need the itunesconncet thing for Artifex? | 11:53.51 |
| (separate to the developer account). | 11:54.04 |
| I possibly also need the apple id of the team agent | 11:54.58 |
| for the destination team. | 11:55.05 |
paulgardiner | Okay. Sending that too. | 11:55.34 |
mattchz | thanks Paul | 11:57.18 |
paulgardiner | np | 11:57.27 |
mattchz | ok to go ahead then? | 11:58.07 |
jogux | matt> the transfer doesn't involve any resigning etc, it's just literally just moving the app in itunesconnect between accounts. | 11:58.09 |
tor8 | mattchz: paulgardiner: something I ran into with the mupdf app yesterday ... it wouldn't open CBZ files from safari or Google Drive. | 11:58.19 |
jogux | someone should probably warn Miles he'll get emails. | 11:58.32 |
tor8 | on iOS that is | 11:58.33 |
mattchz | I presume itâs okay to accept the new paid apps agreement? | 11:59.59 |
henrys | tor8: regression? | 12:01.15 |
| tor8: I need you to review Chrisâs missing glyph list so I can get URW started. | 12:02.49 |
tor8 | henrys: I don't remember if it ever worked | 12:02.52 |
| henrys: I have been going through the glyph lists today, assembled the full requirements list | 12:03.08 |
paulgardiner | mattchz: yeah fine | 12:03.13 |
tor8 | just about to diff them against what we have | 12:03.15 |
| henrys: I've found two more candidate glyph lists we might want though | 12:03.37 |
| henrys: the one I've always thought we need is microsoft's WGL4 | 12:03.49 |
henrys | tor8: you can assign it mattchz to look at. I think it deserves a bug report | 12:04.04 |
tor8 | but now I've ran across the W1G (https://en.wikipedia.org/wiki/World_Glyph_Set) which Linotype and Monotype use in advertisement | 12:05.18 |
| and then the MES-2 (multilingual european subset) -- some sort of official unicode subset for european languages | 12:05.47 |
henrys | tor8: well on the gs side UFST compatibility is the goal. If we add more glyphs thatâs okay as long as the priced doesnât increase dramatically. | 12:05.48 |
tor8 | wgl4 has 663 glyphs, w1g has 601 glyphs (most likely a subset of wgl4) | 12:06.16 |
| and MES-2 has 1095 glyphs (including the full WGL4 subset) | 12:06.28 |
| so I think the question is whether we want to settle for WGL4 or go all-out for MES-2 (which makes the fonts even better as substitute fonts) | 12:07.00 |
| my guess would be that the majority of extras in MES-2 are more diacritics, but I haven't looked in detail yet. | 12:07.37 |
mattchz | lunch. | 12:09.32 |
| bbiab. | 12:09.34 |
paulgardiner | henrys: do you think our previous emails (at least the last few) have been passed onto the new Good contact? | 12:15.22 |
henrys | paulgardiner: are you asking me if brad and richard are going to give email to their new boss that demonstrates they donât know how to answer them? I would assume no. | 12:17.12 |
paulgardiner | henrys: :-) | 12:18.01 |
jogux | :-) | 12:19.29 |
paulgardiner | henrys: should I have received an email from the new contact? | 12:21.53 |
| henrys: Oh yeah, and did you trigger a password update for my gmail account? Didn't see anything. | 12:25.42 |
henrys | paulgardiner: no Iâm going to talk to him today and hopefully get an engineer to talk to us. I wish I understood what was wrong there. If they are a sinking ship and all the talent is left we are just wasting our time. But they could be going through a growth spurt and canât find folks to work with customers. | 12:27.18 |
| paulgardiner: ugh sorry forgot doing it now | 12:27.58 |
paulgardiner | The underlying technology seems good | 12:28.03 |
| henrys: thanks | 12:28.10 |
pedro_mac | henrys: theyâre the same guys Picsel dealt with - I guess one of the big differences is that weâre now just 1 of 4 or so office apps on their platform, so they seem less keen on pushing for SO to be progressed | 12:29.40 |
henrys | paulgardiner: try it now | 12:30.09 |
pedro_mac | agrees with paulgardiner - the SDK itself seems to be pretty reasonable | 12:30.09 |
jogux | henrys : the good platform is pretty sticky I believe, once a corporate has deployed it to there devices (and it seems plenty have), changing to another one would be a huge exercise | 12:31.21 |
| and I would imagine would mean they'd also lose all the apps they'd bought via good. | 12:32.22 |
Robin_Watts | jogux: I think you only ever 'lease' good apps, right? | 12:32.49 |
jogux | ah, er, I don't actually know. | 12:33.22 |
| that may be true. | 12:33.32 |
tor8 | henrys: so do you want the glyph lists as a diff for each font, or just a complete listing (less risk of mistakes on their side) of what we want? | 12:33.46 |
henrys | tor8: complete listing | 12:34.09 |
tor8 | henrys: okay. will a semicolon separated text file with unicode value, adobe glyph name, unicode description columns be alright? | 12:35.40 |
henrys | tor8:I thought they were happy with chrisâ format. | 12:36.28 |
| tor8: can you just add to that? | 12:36.49 |
tor8 | chris format was just a list of the glyph names right? | 12:37.09 |
| this will be a more detailed version of that (given the unicode value and character name, should tell them more in case of glyph name ambiguities) | 12:37.38 |
| or was chris format something else? | 12:37.47 |
henrys | I have a link hold on | 12:38.00 |
| tor8: http://www.ghostscript.com/~chrisl/SF_and_H_glyph-images_and_metrics.zip | 12:39.28 |
| tor8: ugh thatâs only part of it. | 12:40.09 |
| 10,0http://www.ghostscript.com/~chrisl/ufst-vs-urw-missing-glyphs.zip1,0http://www.ghostscript.com/~chrisl/ufst-vs-urw-missing-glyphs.zip10,0http://www.ghostscript.com/~chrisl/ufst-vs-urw-missing-glyphs.zip | 12:41.04 |
| I have no idea why that link was copied 3x. it ends at âzipâ tor8 | 12:41.38 |
tor8 | henrys: I got them | 12:41.50 |
henrys | bbiab - dog needs walkin. | 12:41.59 |
mattchz | â I authorize Apple to enable all previously purchased apps to be shared as part of Family Sharing on a permanent basis." | 12:53.11 |
| I presume thatâs ok? | 12:53.13 |
| (not sure you have much of a choice) | 12:53.23 |
| although actually, the transfer seems to have happened without having to agree to that. | 12:55.14 |
jogux | paulgardiner: one of these times, I'll remember to ask you if you've run ATS before approving a commmit ;-) | 13:10.02 |
| paulgardiner: http://intranet.picsel.com/~ats/cgi-bin/results-tgvath.pl?resultid=1155&diff=0 | 13:10.08 |
| paulgardiner: also I suspect we may have found out what that setting does... | 13:10.59 |
paulgardiner | Damn! I must remember to make it unnecessary for you to ask me if I've run things through ATS. | 13:12.51 |
| Very strange result though. I changed the behaviour when the menu it animated off screen. | 13:14.22 |
henrys | paulgardiner, Robin_Watts: is this bluetooth printer bug something for embedded general? | 13:15.46 |
paulgardiner | henrys: I'd imagine so | 13:16.05 |
henrys | didnât jogux have a similar problem? Canât remember if it was bluetooth though | 13:16.29 |
paulgardiner | jogux: some of these changes are improvements. Some are clearly not | 13:17.18 |
henrys | jogux: we have a couple of reports of SOT not finding bluetooth printers⦠| 13:17.25 |
jogux | I was always using ethernet printers. But the behaviour was not reliable for me. | 13:17.46 |
| I don't think I have any bluetooth printers. I'm not sure I knew bluetooth printers were supported for that matter. | 13:18.09 |
paulgardiner | henrys: still not received a password reset email. I wonder if they are failing to forward on from the gmail account. | 13:21.57 |
henrys | paulgardiner: itâs supposed to prompt you next time you login | 13:22.26 |
paulgardiner | oh okay | 13:22.34 |
jogux | paulgardiner: I think there's only a hand full of ATH scripts that enable top bar hiding | 13:22.47 |
paulgardiner | Most of these diffs look correct to me, but the landscape ones are a mystery | 13:23.27 |
| henrys: It's still saying "Please contact your domain IT administrator to reset your password or retrieve your username" | 13:24.53 |
jogux | paulgardiner: I'm wondering about the first diff. Is it possible your changed it so that the top of the document (which was previously visible by default) is now hidden by the toolbar? | 13:29.09 |
paulgardiner | jogux: absolutley, but only in the case that it will become visible again when the menu animates off | 13:30.22 |
jogux | oh right. is that a behaviour change from the old app? | 13:30.53 |
paulgardiner | Do you mean "does the current app store versions do this?"? | 13:31.59 |
Robin_Watts | jogux: How goes the styling stuff? | 13:40.16 |
mattchz | dances the âapple DRMâ dance | 13:49.48 |
jogux | robin_watts : I've actually been stuck with some other (non-Artifex) work for the last day or two, the latest was that I needed to figure out exactly why it was breaking shapes. | 13:49.59 |
| paulgardiner: that was exactly what I meant :-) | 13:50.23 |
tor8 | henrys: is there a per-glyph quote or what number of glyphs do the current quotes apply for? | 13:51.32 |
Robin_Watts | jogux: sure, was just wondering. | 14:00.27 |
| I've got 2 tab files that seem 'odd'. I think one of those is right to left and the other is due to styling changing the indentation. | 14:00.52 |
henrys | tor8: no the quote was per font and I donât know if we can simply assume glyphs are equally priced | 14:02.49 |
tor8 | henrys: okay, because I have two potential sets of glyphs. one adds 450 glyphs, the other 850. | 14:04.02 |
| I'd be fine with the smaller set, but having the full 850 extra glyphs to match MES-2 would be nice | 14:04.31 |
| I've also looked at the times/arial/courier fonts microsoft ships with Win7 | 14:04.52 |
henrys | tor8: and I assume chrisâ glyphs are a subset of the 450? | 14:05.29 |
tor8 | they're HUGE ... between 2300 and 2700 glyphs | 14:05.31 |
henrys | tor8: wow | 14:06.43 |
tor8 | yes (but I will double check) | 14:06.52 |
| the diffs chris has are about 190 extra glyphs | 14:07.11 |
henrys | tor8: we certainly wouldnât want something like that in a mobile release? | 14:10.36 |
Robin_Watts | Well, MS (presumably) have a full time font department, that have been doing 'stuff' since Win 3.1 | 14:11.14 |
| I'd therefore assume that they've churned out pretty much every imaginable glyph :) | 14:11.30 |
tor8 | henrys: no, the majority of the extras look to be arabic, hebrew and combining characters | 14:11.33 |
jogux | henrys : I wonder what the stats would be on the size of that font vs a retina ipad launch image... | 14:11.43 |
tor8 | the ttf fonts from microsoft sit at between 600 and 800k | 14:12.20 |
| each | 14:12.22 |
paulgardiner | jogux: no old app has different behaviour. Although the rulers spring back with the menu, the document top is never pulled up to where it would be covered by the menu returning (although you can manually place it there of course) | 14:12.45 |
jogux | paulgardiner : ah, right. that kind of makes sense to me. | 14:13.11 |
Robin_Watts | jogux, paulgardiner, pedro_mac: tabbing review on robin/master then if you have time. | 14:13.39 |
| http://intranet.picsel.com/~ats/cgi-bin/results.pl?grid=54 <- Those are the ATS tests. | 14:13.57 |
henrys | wonders if Monotypeâs compression patent has expired and if it could be applied to tt fonts. just thinking out loud | 14:14.05 |
paulgardiner | jogux: so, do you think it's worth looking at what's going on with the version on the branch so as to try to understand how that behaviour is being achieved? | 14:14.48 |
| jogux: on master both things seem to be controlled by a single variable | 14:15.34 |
tor8 | henrys: I was wrong, there are more glyphs in chris set that aren't in the WGL4 | 14:16.29 |
| his list adds 187 glyphs, (179 unique, there are 8 duplicates with different names like Dslash and Dcroat) | 14:17.21 |
| 39 which aren't in the WGL4 | 14:17.26 |
jogux | paulgardiner : yes, I think it probably is worth looking further then (contrary to my statement this morning :( ). I guess there must be a second variable somewhere that controls only the rulers. | 14:20.06 |
paulgardiner | yeah. I'll take a look | 14:22.14 |
mattchz | coo. Iâve finally got a 1.5 build for ios on testflight which appears to work. | 14:28.23 |
| please let me know if it works for you guys. | 14:28.29 |
| trivial commit for review for changing to the new account: http://git.ghostscript.com/?p=user/matt/mupdf.git;a=commitdiff;h=1ce58585a018a5b4d04ef31b7af1d4a754d704cd;hp=5587d318d8495d15fb9157ea8c0a925e6e590e00 | 14:29.23 |
Robin_Watts | mattchz: Seems reasonable. | 14:30.45 |
mattchz | ta | 14:30.54 |
jogux | mattchz : seems to install/run for me | 14:32.54 |
mattchz | cool, ta :) | 14:33.03 |
| let me know if you want me at some point to prepare an appstore build | 14:33.12 |
| tor8: so presumably you want maxSize renamed to maxsize (not max_size) ? | 14:36.18 |
jogux | mattchz : you should rephrase that as "what needs to happen before I can submit this to the appstore?" :) | 14:36.40 |
tor8 | mattchz: both maxsize and max_size are fine | 14:36.52 |
mattchz | so, is camel case verboten everywhere? | 14:37.19 |
| jogux> shrug | 14:37.23 |
Robin_Watts | max_size is preferable, I think. | 14:37.33 |
mattchz | most stuff in that file seems to use maxsize | 14:37.52 |
tor8 | mattchz: we generally use 'cap' for capacity rather than max_size when dealing with growing arrays | 14:38.13 |
| though in this case I'd probably call it new_size or desired or goal or something to that effect | 14:38.50 |
paulgardiner | Robin_Watts: tabbing patch LGTM | 14:39.05 |
Robin_Watts | paulgardiner: Thanks. | 14:39.12 |
mattchz | tor8: fair enough. new_size it is then. | 14:40.28 |
| tor8: howâs that: http://git.ghostscript.com/?p=user/matt/mupdf.git;a=commitdiff;h=3dff43ff0f5064f64c29a619079859d0e789fd26 ? Was there other issues that needed fixing? | 14:43.16 |
Robin_Watts | mattchz: That looks good to me. | 14:46.47 |
tor8 | mattchz: trailing whitespace should be stripped. | 14:46.57 |
mattchz | I thought I had done that? | 14:47.11 |
tor8 | (and if I had written the code, the ternary expression would be lacking a set of parentheses, but that's just me) | 14:47.12 |
mattchz | I tend to (over)use brackets (makes things more readable imho) | 14:47.51 |
tor8 | int_\nfz_shrink_store and if (ctx == NULL)_\n | 14:47.56 |
| mattchz: then you're just like robin. at long as you don't parenthesize "return" I can live with it | 14:48.44 |
mattchz | :) | 14:48.54 |
| oh, I hadnât noticed that trailing space. ta. | 14:49.19 |
tor8 | I also don't compare against NULL... I'd write if (!ctx) rather than if (ctx != NULL) | 14:49.27 |
| but again, robin does like you do | 14:49.31 |
| sorry, make that if (ctx) | 14:49.46 |
Robin_Watts | tor8: And that's exactly why you're wrong :) | 14:49.56 |
tor8 | if (truthy) | 14:50.07 |
mattchz | fix: http://git.ghostscript.com/?p=user/matt/mupdf.git;a=commitdiff;h=594b619ba795d2ccfa61cf2b1733195457a55044 | 14:50.36 |
| tor8: I donât normally bother comparing against NULL either. I thought Iâd probably be chastised if I did that though :) | 14:51.05 |
tor8 | mattchz: right. no need to do so in the future in mupdf at least :) | 14:51.36 |
mattchz | (I do if (ctx) or if (!ctx) too) | 14:51.37 |
tor8 | commit LGTM | 14:51.40 |
mattchz | ta | 14:51.43 |
| quite frankly though, Iâm very happy to do if (ctx != NULL) as an alternative to the frankly repulsive if (NULL != ctx) some people seem to love ;) | 14:52.18 |
| pushed. | 14:52.59 |
tor8 | mattchz: javascript programmers can't do the latter ... a mistake there would really mess things up. "undefined" in javascript is assignable... | 14:53.07 |
mattchz | nod. I just find it terribly unreadable (the NULL != foo thing) | 14:53.36 |
tor8 | yeah. it's too backwards to parse if you're not used to it. | 14:54.03 |
mattchz | if (EPAGE_NO_ERROR != e) | 14:54.13 |
| eww | 14:54.15 |
tor8 | OTOH I find that if (ctx) has fewer parsing steps than if (ctx != NULL) | 14:54.18 |
mattchz | yeah. I like if (ctx) | 14:54.29 |
Robin_Watts | If I'm completely honest, I'd probably use if (ctx) and if (ctx == NULL) | 14:56.26 |
| but I find that deeply dissatisfying. | 14:56.44 |
| and I can't bring myself to use if (!ctx) | 14:56.58 |
mattchz | tbh, any of those are readable. | 14:57.52 |
pedro_mac | Robin_Watts: I think the tabbing change looks good and makes sense. We still have a number of tab-related things to sort (like properly supporting center and right-aligned tabs) but this seems like a good step along the way. Test results look pretty much an improvement, although a couple of changes are not necessarily âbetterâ - just different (shift-jis-ppt.ath, 3rd failure) and the 53rd image failure - ppt âColour schemeâ slides ha | 15:01.25 |
| their title slightly off-centre now). | 15:01.26 |
Robin_Watts | I've checked, and those are all now correct. | 15:02.05 |
pedro_mac | cool. Iâm happy with that then. Most indentation in the docs seems to line up much better now | 15:03.42 |
| minor typo in the commit message coninved->convinced | 15:05.06 |
Robin_Watts | oops, ta. | 15:05.45 |
pedro_mac | does that all the time :) | 15:06.27 |
Robin_Watts | hmm. It seems I have 2 more reviews on robin/master that haven't been looked at/pushed yet. | 15:30.05 |
| The ThumbnailGenerator_setSubSampleThreshold one | 15:30.36 |
| and the clean/rebuild targets for the VS solution. | 15:30.59 |
jogux | I presume globalManagerData->clientConfig is calloced/memseted (as otherwise subSampleThershold may be used uninitialised?) | 15:32.37 |
Robin_Watts | jogux: It is. | 15:32.48 |
| Other code makes similar assumptions. | 15:33.13 |
jogux | cool, that one LGTM then | 15:33.23 |
Robin_Watts | Ta. | 15:33.41 |
mattchz | btw, android peeps: be careful if you upgrade your SDK to the latest stuff. | 15:34.22 |
| itâs caused me a whole load of breakage. | 15:34.35 |
| my eclipse is now stuffed up. | 15:34.56 |
Robin_Watts | mattchz: What has changed? | 15:35.11 |
pedro_mac | aww - jdk breakages? | 15:35.20 |
jogux | robin_watts : I don't really seen anything obviously wrong in the VS one | 15:35.28 |
Robin_Watts | Ta. | 15:35.36 |
mattchz | Well, thereâs the L, Wearable and TV SDKs | 15:35.38 |
| but the updated platform/sdk tools seem a bit moosed. I canât seem to update my ADT to version 23 due to dependency issues. | 15:36.06 |
| And apparently, Graham was saying that they broke proguard. | 15:36.16 |
jogux | eclipse has a none-stuffed up state? | 15:36.21 |
mattchz | heh | 15:36.27 |
Robin_Watts | eek, breaking proguard is bad. | 15:36.32 |
mattchz | https://code.google.com/p/android/issues/detail?id=72419 | 15:36.58 |
| âFYI even other important commands are missing zipalign, sqlite3, and so on are missing. This version is really BROKEN so it's not acceptable the level Medium. Fix asap this mess. | 15:37.15 |
| Jun 27 (2 days ago) | 15:37.15 |
| " | 15:37.16 |
Robin_Watts | "Fix asap this mess." I just love the entitled feeling that developers get. | 15:38.25 |
mattchz | Well, tbf, Google have pretty much broken everything. | 15:38.57 |
pedro_mac | mattchz: Iâll stick with v22.6 then :) | 15:39.05 |
Robin_Watts | mattchz: yes, but nonetheless, there are workarounds and google are working to fix it. | 15:39.28 |
| We get lots of people who turn up here and insist that we should instantly drop everything and fix things that don't work for them. | 15:40.19 |
mattchz | I do wonder if it had been tested at all. | 15:40.22 |
| Oh, indeed, and itâs not right that people are like that. | 15:40.45 |
| This is a fairly spectacular screw up on Googleâs part though; they clearly didnât test it at all. | 15:41.16 |
Robin_Watts | mattchz: Yeah. I bet it was people saying "we must have this out for Google IO!" | 15:42.00 |
mattchz | Robin: yeah - almost certainly. | 15:47.47 |
Robin_Watts | 1 more commit on robin/master. ATS tests running now. | 15:51.26 |
| With this in, and with jogux's styling stuff, we should have Artifex2014 rendering properly now. | 15:52.12 |
| I might have a look at right to left tabbing for a bit. | 15:52.50 |
mattchz | FWIW, the workaround suggested by Google doesnât fix many of the problems. | 15:55.00 |
| (restoring from a Time Machine backup, now). | 16:08.56 |
mvrhel_laptop | Robin_Watts: are you there? | 17:06.29 |
Robin_Watts | I am. | 17:06.45 |
mvrhel_laptop | can I ask you a quick question | 17:06.54 |
Robin_Watts | sure. | 17:06.59 |
mvrhel_laptop | so it is about the format of the edr and how I go about adding in and extracting additional information | 17:07.33 |
Robin_Watts | ok. | 17:07.47 |
mvrhel_laptop | I am not sure where I look to go about doing that | 17:07.57 |
Robin_Watts | What sort of extra information do you want to add ? | 17:08.09 |
mvrhel_laptop | essentially, I need to be able to add in information (and extract and use it during render time) that is equivalent to the dPt information. This is extra information for a particular data point in a plat | 17:09.08 |
| plot | 17:09.10 |
| Robin_Watts: for example, color, or gradient | 17:09.19 |
| 3D | 17:09.21 |
| if I can get this through the edr then I think we have everything we need to draw it properly | 17:09.38 |
Robin_Watts | mvrhel_laptop: Do you have an edr dump of a document with a chart in to hand ? | 17:10.03 |
mvrhel_laptop | I can get the information out of the xml source file | 17:10.04 |
| Robin_Watts: no. how do I do that? | 17:10.15 |
| this is the sort of thing that will help me | 17:10.30 |
Robin_Watts | mvrhel_laptop: When you run test-shell on a file, it outputs a file. | 17:10.34 |
mvrhel_laptop | ah ok | 17:10.37 |
| then yes I do | 17:10.42 |
Robin_Watts | the name is given on the command line, but it's generally called out.txt | 17:10.53 |
mvrhel_laptop | yes | 17:10.57 |
Robin_Watts | Can I see it please? :) | 17:11.23 |
mvrhel_laptop | yes. hold on let me get this finished running the I will shar | 17:12.02 |
| share | 17:12.03 |
Robin_Watts | fab. | 17:12.08 |
mvrhel_laptop | cool. hold on I will paste it | 17:12.50 |
| http://pastebin.com/qHJcktbH Robin_Watts | 17:13.37 |
Robin_Watts | So we have some master page data first... then the actual document Edr, then the Layout dump at the end. | 17:14.43 |
mvrhel_laptop | right. I see the chart data | 17:14.51 |
Robin_Watts | The document Edr starts at line 122. | 17:14.52 |
mvrhel_laptop | Style, idx = 29 Content = (dbl = 5.000000) | 17:14.55 |
| Style, idx = 30 Content = (dbl = 10.000000) | 17:14.57 |
| Style, idx = 31 Content = (dbl = 15.000000) | 17:14.58 |
| Style, idx = 32 Content = (dbl = 20.000000) | 17:15.00 |
| Style, idx = 33 Content = (dbl = 25.000000) | 17:15.01 |
| It is a simple chart | 17:15.09 |
| each one of these points has its own special color though | 17:15.20 |
| some have an sRGB value | 17:15.25 |
| some a gradiant | 17:15.28 |
| one a gradient | 17:15.36 |
| and one a modified luminance value of a palette color | 17:15.50 |
| this information is not making it through Robin_Watts | 17:16.01 |
| so what I need to understand is the best practice to add this | 17:16.14 |
Robin_Watts | So the marker information is living in a group (line 149) | 17:16.21 |
| You could add another group of type 'markercolors' and then have a series of entries with the color data in. | 17:17.01 |
mvrhel_laptop | ok. that makes sense to me. | 17:17.11 |
Robin_Watts | (I'm making this up as I go...) | 17:17.21 |
mvrhel_laptop | :) | 17:17.25 |
| we all are | 17:17.27 |
Robin_Watts | The existing marker data looks to be 'dbl' (double) | 17:17.53 |
mvrhel_laptop | Robin_Watts: those are the "values" of the bar chart points | 17:18.08 |
Robin_Watts | so you could use a 'dbl' to hold a 24bit RGB color. | 17:18.10 |
| yeah. | 17:18.12 |
mvrhel_laptop | Robin_Watts: I will look how other color information is stored in the EDR with some simple examples and follow those | 17:18.33 |
Robin_Watts | but are you saying that some markers in a chart might be a gradient, and others might be an RGB color? | 17:18.44 |
mvrhel_laptop | now that I see this output format, I have a better understanding | 17:18.50 |
Robin_Watts | or is it the case that all the colors in a given chart will be specified in the same way ? | 17:19.11 |
mvrhel_laptop | Robin_Watts: yes. there seem to be 3 types. sRGB values. an accent value from the current pallette and a gradiant | 17:19.21 |
| Robin_Watts: generally a palette is used, and the graph parts are indexed on the palette | 17:19.46 |
| however, you can override these colors and set special colors | 17:19.58 |
| SOT handles the palette case OK | 17:20.11 |
Robin_Watts | In that case, we need to distinguish the different types. | 17:20.16 |
mvrhel_laptop | but the special colors are missing | 17:20.16 |
| well, it will say "accent 1" for example if it is grabbing accent 1 from the pallete | 17:20.40 |
Robin_Watts | ooh, actually... | 17:20.40 |
| If you look at the group on line 145, there is a 'privData' pointer. | 17:21.02 |
mvrhel_laptop | Robin_Watts: so what we need is some special information that may or may not be present. If it is present we will use it | 17:21.03 |
Robin_Watts | to something of type Edr_Chart_SaveData. | 17:21.14 |
mvrhel_laptop | I will encode the palette index, the gradient details or the sRGB color | 17:21.22 |
Robin_Watts | I *think* the general rule is that anything that can edited needs to be in the Edr. | 17:21.37 |
| but 'constant' stuff can be put into the private data. | 17:21.50 |
mvrhel_laptop | ok this is stuff that can be edited ideally. at least it can in word and excel | 17:21.57 |
| I would think we would want to do so in SOT eventually | 17:22.11 |
Robin_Watts | yeah, but are you going to write the code to do that? | 17:22.16 |
mvrhel_laptop | no :) | 17:22.20 |
Robin_Watts | OK, you're keener than me then :) | 17:22.21 |
mvrhel_laptop | step one is to render it correctly | 17:22.33 |
Robin_Watts | I might have been tempted to put the data into Edr_Chart_SaveData to start with. | 17:22.43 |
mvrhel_laptop | but I have no interest in adding the capability to edit in in the UI mess | 17:22.53 |
Robin_Watts | (maybe I'm wrong and that's not possible, but if it is, it'll be a hell of a lot easier and will use a lot less memory etc) | 17:23.10 |
mvrhel_laptop | Robin_Watts: ok so I will try to add into Edr_Chart_SaveData | 17:25.23 |
| I will attempt to finish my parsing and see if I can add in some new content and get it back out and use it | 17:25.51 |
| however today is supposed to be vacation so I will probably not get to it until tomorrow. getting evil eye from wife right now.... | 17:26.13 |
jogux | robin_watts : Edr_Chart_SaveData appears to not contain much at the moment, so I suspect this would be a departure from the intended model | 17:26.57 |
mvrhel_laptop | jogux: so is what I just said a bad idea? | 17:27.20 |
| brb | 17:27.31 |
jogux | mvrhel_laptop : just because I'm picking holes in this plan doesn't mean I have a better one :-) | 17:27.57 |
Robin_Watts | jogux: Ah, rats. | 17:28.14 |
| My MSVC just locked up, hence I'm still trying to get to Edr_Chart_SaveData | 17:28.38 |
jogux | robin_watts : I think it's just got a filename in it, and something else. (for the xml format anyway). | 17:28.39 |
| for xml, it's entire contents is: | 17:29.14 |
| Unicode *partName; /**< Chart XML filename, known as part name */ | 17:29.15 |
| int index; /**< Chart series index */ | 17:29.16 |
Robin_Watts | I suspect it's supposed to be "the information required to save the chart back from the XML". | 17:29.53 |
| So perhaps putting the data into the Edr is a better plan. | 17:30.55 |
jogux | for the binary formats, it's a reference to where in the file the original data is - so I think the essence here is that it's purpose is to allow the place the chart was defined to be found. | 17:31.49 |
Robin_Watts | jogux: So, a commit I did earlier should have changed the bitmaps etc. | 17:32.14 |
mvrhel_laptop | jogux: ok. So it will be stored ad Content = (blah blah blah) | 17:32.22 |
Robin_Watts | do I still need to rerun a group task? | 17:32.24 |
mvrhel_laptop | in the edr? | 17:32.29 |
| s/ad/as/ | 17:32.34 |
jogux | robin_watts: you could | 17:32.34 |
Robin_Watts | jogux: What number? | 17:32.57 |
mvrhel_laptop | for each point. if they have some special color case | 17:33.04 |
| to me that would seem easier | 17:33.19 |
jogux | robin_watts: the description will be either "tgv ath" or "tgv ath lomond" deepending which yuou need | 17:33.36 |
Robin_Watts | jogu: 5 and 163 ? | 17:33.56 |
jogux | robin_watts: yes | 17:34.08 |
Robin_Watts | Ta. | 17:34.11 |
| mvrhel_laptop: The way the data is currently stored, we have a group that then contains 1 entry for each point. | 17:34.46 |
mvrhel_laptop | yes | 17:35.00 |
Robin_Watts | If you want to be able to have a 'sparse' set of data, that's trickier. | 17:35.02 |
| i.e. if you have a chart with 10 points, and just 2 of them have to be special ones, we can't quite use the normal trick. | 17:35.26 |
mvrhel_laptop | Robin_Watts: ok. If any are "special" the others could indicate the "accent" color to use from the palette | 17:36.26 |
| so we would have something for each one | 17:36.37 |
Robin_Watts | OK, so we have a markercolors group. | 17:36.51 |
| And we have 10 entries in that. | 17:36.57 |
mvrhel_laptop | right | 17:37.00 |
Robin_Watts | If an entry is an int, it gives the accent color to use. | 17:37.12 |
| If it's a double it's an rgb value. | 17:37.19 |
| If it's a group then it could have gradient data in? | 17:37.29 |
mvrhel_laptop | Robin_Watts: I need to dig into the gradient specification. we will plan on something like that | 17:37.54 |
Robin_Watts | jogux: Is this all sounding like we're misusing edr like this? Is there a better way of doing it? | 17:38.09 |
mvrhel_laptop | that will be the next phase | 17:38.11 |
Robin_Watts | I don't know if there are 'tuples' etc. | 17:38.18 |
mvrhel_laptop | EpageStrokeColor = (color = (4f,81,bd/ff)), | 17:38.33 |
| like this Robin_Watts | 17:38.36 |
Robin_Watts | mvrhel_laptop: I'm not saying you should feel obliged to implement all of it at once, just that we should have a vague plan. | 17:38.39 |
mvrhel_laptop | Robin_Watts: yes | 17:38.45 |
Robin_Watts | mvrhel_laptop: That's a style. | 17:38.55 |
mvrhel_laptop | Robin_Watts: but this is how the sRGB color is encoded | 17:39.09 |
jogux | robin_watts : I'm learning this at roughly the same speed you are; the moving of charts into edr so they could be editted in post-first-admin stuff I believe :) | 17:39.19 |
mvrhel_laptop | in the ooxml too | 17:39.20 |
Robin_Watts | mvrhel_laptop: Right, but a style is something you put on an edr node. | 17:39.47 |
jogux | robin_watts : the edr already uses inline style | 17:40.00 |
| or should I say abuses. | 17:40.04 |
Robin_Watts | oh, inline style, right, yes. OK. | 17:40.13 |
jogux | the objects of type 'Style' | 17:40.14 |
Robin_Watts | Gotcha, sorry. | 17:40.22 |
jogux | so we could add an inline EpageStrokeColor rule to sort colours | 17:40.45 |
| there must be some existing mechanism for storing gradients too (though I don't recall what it is) | 17:42.00 |
Robin_Watts | We have things like "EpageTabstopArray", so clearly it is possible to have styles with a variable number of data entries in. | 17:42.25 |
mvrhel_laptop | jogux: yes. I see some of this stuff in the chart code now. however it is stored for the whole series. not for individual points | 17:42.40 |
Robin_Watts | Aha, EpageGradient | 17:42.43 |
mvrhel_laptop | let me show you where | 17:42.43 |
jogux | mvrhel_laptop : awesome | 17:42.59 |
mvrhel_laptop | jogux: getOptionalColour | 17:45.01 |
jogux | urgh. it's almost 7pm. | 17:45.01 |
mvrhel_laptop | in drawingml-chart.c | 17:45.07 |
| jogux: if you need to go that is fine | 17:45.14 |
| oops wrong one. hold | 17:45.34 |
| on | 17:45.36 |
| getOptionalColourProperties | 17:45.49 |
| jogux and Robin_Watts : so this function attempts to get these properties for the whole series rather than for a single data point. We need to do something like this for each data point node | 17:47.00 |
| this is called in addChartSeries | 17:47.34 |
Robin_Watts | ah, that's in the parsing side, yes. | 17:47.42 |
mvrhel_laptop | and it is handed the series. | 17:47.42 |
| in cases where someone has specified special colors etc for the chart parts, it returns nothing since it does not drill down to the data points | 17:48.21 |
Robin_Watts | right. so all that's code you're going to have to add. | 17:48.39 |
mvrhel_laptop | yes | 17:48.43 |
| I almost have that done | 17:48.57 |
| but I was worried about how to pack into the edr and read back out | 17:49.09 |
| also, when reading back out, I need to worry about if we are actually drawing or if we are packing into a word or excel format? | 17:49.35 |
| Robin_Watts: that was actually a question | 17:49.43 |
| the drawing part I think I can handle. not so sure about going to MS formats.... | 17:50.05 |
| or PDF.... | 17:50.10 |
Robin_Watts | We never transcode formats. | 17:50.22 |
jogux | mvrhel_laptop : pdf is generated from the display list so probably will just work | 17:50.24 |
Robin_Watts | with the exception of PDF, which is generated from the display list, as jogux says. | 17:50.37 |
| so essentially all you need to worry about for now is loading. | 17:50.53 |
mvrhel_laptop | but if I added new content wont it need to know how to do it? | 17:50.53 |
jogux | so the only extra bit to worry about is saving to pptx. | 17:51.08 |
Robin_Watts | at some point we need to add code to save the new data. | 17:51.10 |
mvrhel_laptop | i.e. this data point rect fill gets this special color | 17:51.11 |
| to pptx? | 17:51.22 |
| You know what it is probably getting late for you guys. lets continue this tomorrow during the SOT meeting | 17:52.04 |
jogux | err. whatever the source document format was :-) I meant excel of course. | 17:52.09 |
mvrhel_laptop | jogux ah ok | 17:52.17 |
| this is docx. but I created the chart in xlsx and copies and pasted to word | 17:52.45 |
| I assume the same chart format will work with all 3 | 17:52.56 |
Robin_Watts | mvrhel_laptop: I'm going to be about for a bit on and off, but I know this is a vacation day for you. | 17:52.57 |
| yeah, it's all oxm; | 17:53.10 |
| ooxml | 17:53.13 |
jogux | mvrhel_laptop: yes. actually I guess this code is common between all 3. | 17:53.16 |
mvrhel_laptop | ok ool | 17:53.18 |
| cool | 17:53.26 |
jogux | mvrhel_laptop : the display list is a step further one than edr, where it the chart has (I hope) being converted into simple paths etc - so the existing pdf export code already knows how to handle paths. | 17:53.44 |
mvrhel_laptop | oh. | 17:54.01 |
| I see | 17:54.08 |
| thanks for that clarification! | 17:54.13 |
| hold on a sec | 17:54.28 |
jogux | the display list starts after "---Layout dump: Start----------------------" in the output from the testshell | 17:54.42 |
mvrhel_laptop | oh | 17:56.51 |
| a readable displaylist | 17:57.22 |
| so I have to do 3 things. 1) parse the ooxml and get the data 2) pack into the edr 3) convert from edr to display list | 17:58.45 |
| jogux, Robin_Watts is that correct? | 17:58.58 |
| sorry for the basic questions. | 17:59.43 |
jogux | yes. and 4) get data back out edr when users edits/saves. | 18:00.23 |
| actually that may not be necessary as we'll not be changing it. | 18:00.49 |
| (the save process may be managed to preserve things it doesn't know about) | 18:01.03 |
mvrhel_laptop | jogux: ok that makes sense. I think 4) can come in another phase if we decide we want to add that type of editing to the app(s) | 18:01.45 |
| thank you for clearing all of this up for me | 18:02.02 |
jogux | yes, hopefully. | 18:02.11 |
mvrhel_laptop | I now have a much clearer picture of things | 18:02.12 |
jogux | no problems | 18:02.12 |
jogux | heads home . night all | 18:02.27 |
mvrhel_laptop | have a good night jogux | 18:02.36 |
jogux | thanks | 18:03.00 |
tsbtmn | I am a longtime mupdf user and I've noticed that in version 1.5 the 'j' & 'h' up/down panning keys now also move to the next page once the bottom or top or bottom of the current page has been reached. Previously, they would stop at the top or bottom of a page. Is there a way to configure mupdf to its prior behavior? | 18:12.57 |
Robin_Watts | tsbtmn: short answer, no. | 18:15.58 |
tsbtmn | Robin_Watts: Thanks. For whatever reason, my muscle memory doesn't seem to be adapting well to this particular change. | 18:19.51 |
dddvvv | hello Artifex. can someone answer me some questions about mujs? | 18:54.57 |
| please :) | 18:55.10 |
Robin_Watts | dddvvv: We can try. | 18:59.07 |
| (Don't ask to ask, just ask :) ) | 18:59.14 |
dddvvv | thanks. i have a little startup and I want to make the existing tools easier to use. instead of programming in c++ i decided to go for an8 hours and embedded a js engine in the layouter. | 19:02.56 |
| do you think it is feasible to embed mujs. or should i use tiny-js? | 19:03.22 |
| currently we dont need speed. | 19:03.34 |
Robin_Watts | mujs sounds ideal. | 19:03.55 |
| but honestly, I'm not the guy to talk to. You really want tor8. Tor8? | 19:04.15 |
dddvvv | do you plan to have a release soon? | 19:04.20 |
Robin_Watts | dddvvv: There should be a version on mujs.com. | 19:04.44 |
| We use it in mupdf and it works just fine. I am not aware of any holes in its support. | 19:05.00 |
dddvvv | great! you made my day | 19:05.12 |
| :) | 19:05.18 |
Robin_Watts | It is released under the GNU GPL. | 19:05.25 |
| If you want to use it with non GPL software, then you'll need a commercial license. | 19:05.36 |
dddvvv | mupdf is cool stuff. if it uses it than it is solid. | 19:05.37 |
Robin_Watts | same author for the two. | 19:05.46 |
| If you come back here tomorrow you can talk to tor with more questions. tor is in sweden, so it's evening there. | 19:06.40 |
dddvvv | thank you Robin. i will stick around. about GPL, I actually do care about the software I am using. you can be sure I will contribute effectively. | 19:09.00 |
henrys | paulgardiner: are you still about ? | 19:18.26 |
paulgardiner | Yep | 19:19.13 |
henrys | paulgardiner: I suggest any outstanding question be sent/resent to 1,0Matt Sturges msturges at good dot com | 19:20.11 |
paulgardiner | Oh right. Will do. | 19:20.49 |
henrys | paulgardiner: just say miles and henry referred you | 19:21.02 |
tor8 | Robin_Watts: dddvvv: I'm around, for a short while longer if you have any questions | 19:35.38 |
dddvvv | tor8: thank you. i am hacking on it right now. getting it to compile in visual studio 2012, hoping there are no problems with c99. | 19:41.59 |
Robin_Watts | We routinely build it in 2005 | 19:43.30 |
dddvvv | that put me at ease. | 19:45.10 |
Robin_Watts | It builds on linux/macos/windows/android all without any problems. | 19:45.31 |
dddvvv | it builds on my machine too. straight of the box, great software! :) | 19:48.43 |
tor8 | dddvvv: cool! | 20:11.40 |
| I'm logging off now. I'll be back tomorrow. | 20:12.19 |
Shelly_ | henrys: are you still around? | 20:41.59 |
Robin_Watts | evening Shelly_ | 20:42.39 |
Shelly_ | hi Robin, working late again? | 20:43.03 |
Robin_Watts | just trying to get some tests to go through. | 20:43.14 |
Shelly_ | I am catching up on the logs, look like things are going well for you on the office integration | 20:44.40 |
Robin_Watts | dunno if I'd say well. | 20:45.11 |
marcosw | mvrhel_laptop: did you see the email this morning from Zoltan asking about type 103 images and transparency? | 20:45.31 |
Shelly_ | we'll agree on progressing! | 20:45.41 |
| for the logs, henrys: I have emailed you an update on some bounty bug fixes and the OpenJPEG2 code has been patched for GS memory management, I mailed then 2 weeks ago but still no reply.... | 20:47.19 |
| Have fun Robin, catch you round. | 20:48.01 |
Robin_Watts | nght | 20:48.08 |
| I think I need to make changes in the way the font manager does getWidths so it can properly account for spaces. | 23:20.47 |
| Forward 1 day (to 2014/07/01)>>> | |