| | <<<Back 1 day (to 2012/10/04) | 2012/10/05 |
| mvrhel_laptop | henrys: you broke the build ;( | 01:11.17 |
| | Robin_Watts: how hard is it to revert the last commit? | 01:11.34 |
| henrys | for windows? | 01:11.45 |
| mvrhel_laptop | no look at the logs | 01:11.51 |
| | or the regression | 01:11.57 |
| | dashboard | 01:12.00 |
| henrys | it passed the local push | 01:12.44 |
| mvrhel_laptop | thats what they all say :) | 01:13.00 |
| | or thats always my excuse | 01:13.06 |
| henrys | oh I see xps compile broke and that probably isn't tested by a "pcl" push | 01:14.32 |
| mvrhel_laptop | ah yes | 01:14.43 |
| henrys | give me a second I'll fix it. | 01:14.45 |
| mvrhel_laptop | that makes sense | 01:14.47 |
| | ok | 01:14.48 |
| | brb | 01:14.51 |
| henrys | oh no everything is broken because I didn't add a file. | 01:17.23 |
| | if I just fix it will the clusterpush use it's last valid checksums or does it save 0's or something foul for a broken build. | 01:18.39 |
| | I'll give it a go, it can't get much worse. | 01:21.21 |
| | it seems to be running | 01:22.58 |
| mvrhel_laptop | henrys; ok cool | 01:35.01 |
| henrys | looks like it passed so I guess it did the right thing and looked at the last valid crc's. | 01:36.17 |
| mvrhel_laptop | yes | 01:36.33 |
| | bbiab to check my local push | 01:38.31 |
| | I appear to be coming down with a cold. Calling it a night | 02:38.24 |
| kens | Robin_Watts : (when you read the logs) COuld you look at bug #693367 please ? Its not entirely clear, but it 'looks' like they are using the 9.07 release code, and 'may' have a problem with filenames which *might* be related to teh Unicode changes.... | 04:22.00 |
| henrys | kens:yikes what are you doing here? | 04:22.29 |
| kens | Hi henrys | 04:22.37 |
| | Woke up at 2:30 am in some pain, my shoulder was hurting. | 04:22.50 |
| | Its fine now, but, ewll, I;m awake..... | 04:22.59 |
| henrys | I wanted to ask you what your current project is, Miles likes a report after the in person meeting and I forgot what you're up to. | 04:23.18 |
| kens | Well I finished up the linearisation, my next big job is adding colour management to pdfwrite (and by implication ps2write) | 04:23.45 |
| henrys | thanks | 04:24.00 |
| kens | I talked to Michael about this at the staff meeting, and most of it is already in place from his POV. | 04:24.09 |
| | There's a couple of things I need, which will have to wait until Micahel has finished his current customer work, but I can move on wit it anyway | 04:24.35 |
| | I'm just trying to clear up a few bugs, get my desk tidy as it were, tehn I can destory pdfwrite again ;-) | 04:24.58 |
| henrys | yes, I think we all have quite a few small todos to catch up on before we go off breaking new stuff ;-) | 04:25.35 |
| kens | Well, finishing the linearisation allowed me to close 5 bugs, which helped :-) | 04:26.08 |
| | There's one free user stange one I want to look at, and one from Marcios (seg fault on regression tests), but I need a 64-bit Linux to check that on eout | 04:26.51 |
| | At the moment my VM capability is messed up. | 04:27.06 |
| | Rather than play with ti, I'm waiting till my new computer arrives (promised for today) and I'll set everything up new on there. | 04:27.35 |
| henrys | what make did you get? | 04:28.12 |
| kens | Custom build | 04:28.17 |
| | A company called PC Specialist | 04:28.25 |
| | I wanted a SSD for the primary drive, and a RAID 1+0 for the main storage | 04:28.46 |
| | and none of the main suppliers offered that as a configuration | 04:29.09 |
| henrys | I hear everyone recommending the SSD's | 04:29.22 |
| kens | Well, they are at least 3x faster than a spinning disk :-) | 04:29.38 |
| | So I'll put the OS on there for faster boot up | 04:29.49 |
| | and things like Acrobat, Visual Studio etc. | 04:30.05 |
| | source code, test files, what have you can all go on the RAID | 04:30.19 |
| henrys | yeah I have a friend of mine who works at Seagate actually about a couple miles from me who works on disks - he's hoping it can get him to retirement and last a few more years. | 04:30.32 |
| kens | Oh I'm sure there are a few years left in the technology | 04:30.50 |
| | flash is still expensive, and has a limited rewrite cycle | 04:31.06 |
| henrys | yes that's true | 04:31.21 |
| kens | Which is why I'll put the apps on the SSD (don't change much) and the volatile data on the spinning disks | 04:31.30 |
| | While I do back up my machine regularly, I like the security of a RAID mirror | 04:31.58 |
| | Looks like I'm going to have to poke Matsubara and ask again for copies of the files they are having trouble with. | 04:33.04 |
| henrys | btw you still have a warning about comparison always false:gdevpdf.c:1414, there are also a bunch of things about %ld not matching int64 but those are never interesting. | 04:39.18 |
| kens | henrys I'd very much like the detail on the comparison only false test | 04:39.40 |
| | I don't get that here, but I'd like to at least be sure its benign | 04:39.53 |
| | The %d thing is only for debug statements, so I'm not too worried about those. | 04:40.15 |
| | debug print in fact. | 04:40.26 |
| henrys | line 1414 in the current code .... HintByte > 255 | 04:40.58 |
| kens | Ah, let me look | 04:41.08 |
| | Ha, well that is correct, ti doesn't make sense.... | 04:42.17 |
| | I even remember intending to change that, it should be 254.... | 04:43.08 |
| | I'll do it now. | 04:43.18 |
| henrys | that would make all the difference. | 04:43.25 |
| kens | Its unlikely ever to trip, its a security to prevent a buffer overrun. | 04:43.49 |
| | But it should be correct. | 04:43.55 |
| | Hmm Matsubara is claiming this problem is on the Mac, that *is* a surprise. I was expecting it to be Windows | 04:45.16 |
| | And now the cluster is upset. Looks like connectivity problems | 05:03.44 |
| | Or Amazon has died again possibly | 05:04.01 |
| henrys | hmm I broke things earlier but then I fixed it I think. | 05:05.32 |
| kens | It snot hte build that's broken, the cluster nodes can't talk to casper.ghostscript.com | 05:05.55 |
| henrys | yike that looks like something different. | 05:06.00 |
| kens | I can ping it though, so I'm a bit puzzled | 05:06.18 |
| | maybe the cluster master process has died | 05:06.28 |
| henrys | yup I'm logged into casper. | 05:06.39 |
| kens | Well, I can try resubmitting hte test | 05:07.01 |
| | Seems happy enough at the moment | 05:08.01 |
| henrys | strange | 05:08.25 |
| kens | Oh, I wonder if its restarting my job, I haven't finished restrating the test and its already going | 05:08.36 |
| henrys` | chrisl:I see what looks like tor moving the ufst to git - are we still using the svn repo for git? | 07:06.19 |
| | I'm sorry for the ufst not git obviously | 07:06.36 |
| chrisl | henrys': as far as I know, we're still using svn for all the "private" stuff - luratech, ufst etc | 07:07.15 |
| | henrys': of course, now that git does "subtree" operations, some of the reason of keeping all that in svn is gone.... | 07:08.47 |
| henrys` | thanks, have a good day guys hope kens gets a nap! | 07:24.30 |
| kens | goodnight henrys | 07:24.40 |
| chrisl | 'night henrys | 07:25.06 |
| sebras | tor8: morning toro. | 08:50.58 |
| tor8 | morning se-bash | 08:54.30 |
| | -tien | 08:54.33 |
| chrisl | tor8: henrys was asking earlier if you were doing something with the svn-private stuff (ufst, luratech etc), moving it to git? | 08:57.32 |
| sebras | tor8: U+5341 U+516B U+5341 U+94B1 | 09:00.33 |
| tor8 | chrisl: I haven't given it much thought. I did convert some stuff from the old svn, it's in my /home/tor/src/private on casper | 09:03.22 |
| | more specifically luratech and ufst | 09:03.39 |
| | but that was a year and a half ago, so I can't really remember what I did | 09:04.01 |
| chrisl | tor8: that's okay, i guess henrys saw that and wondered if we were moving stuff over. I'm not keen at least until the git subtree stuff is better established | 09:04.28 |
| Robin_Watts | We can't do private stuff with git subtree can we? | 09:33.04 |
| | cos it's got to be private :) | 09:33.12 |
| sebras | Robin_Watts: fetch over ssh? | 09:37.36 |
| Robin_Watts | Right, but how annoying is it going to be for people that do 'git submodule update' and find that it fails because they don't have access to the private repos ? | 09:38.55 |
| sebras | Robin_Watts: oh.. so you're saying that it's not possible to do submodule update on just a subset of all the submodules. | 09:41.03 |
| | you're probably right of course. | 09:41.15 |
| Robin_Watts | sebras: Such would be my fear, yes. | 09:41.55 |
| tor8 | yeah, git submodule annoyingly stops on the first error. they may have fixed that in newer versions though. | 09:45.02 |
| sebras | tor8: it would have been nice if one could update the submodules per origin. | 09:46.15 |
| | eh.. remote. | 09:46.20 |
| chrisl | Oh, for some reason I had it my mind that subtree would also let you clone subtree, or do "sparse" clones...... | 09:52.58 |
| Robin_Watts | chrisl: typically you just do: git subtree init (once only) that pulls in all the remotes. | 10:00.39 |
| | Then you do git subtree update each time to update 'em all. | 10:00.53 |
| | (or you can do both in 1 hit: git subtree update --init) | 10:01.16 |
| chrisl | Robin_Watts: sure. But until/unless we can "pull" just a subdirectory from a repo (unlikely to happen), we're better with svn for the private repo | 10:17.53 |
| Robin_Watts | chrisl: Absolutely, that's what I was trying to argue. | 10:22.13 |
| | Using subtree is a pain in that you have to do 2 commands to update to a new git each time (git pull --rebase, then git submodule update). | 10:23.03 |
| | It would be WAY worse if we had to handle each submodule separately. | 10:23.18 |
| chrisl | Hopefully, the submodule stuff will improve over time.... | 10:25.08 |
| paulgardiner | Robin_Watts: What was the pain with subtree? | 10:37.10 |
| | Or should I read that after s/subtree/submodule/ ? | 10:38.49 |
| Robin_Watts | sorry, yes. | 10:43.42 |
| paulgardiner | ah right. got it | 10:44.02 |
| Robin_Watts | Sheesh. | 11:35.59 |
| | So if an android permission is "signature", it means you can only get that permission if you are signed with the same key as the permission granter. | 11:36.37 |
| | So if the downloader is a system thing (and it is), and it grants permissions to the /all_downloads/ path only to "signature" holders, it means that only apps signed with the system key can access them. | 11:37.46 |
| | i.e. no third party app can ever get access. | 11:37.56 |
| | kens: bug 693367 | 11:43.25 |
| | My recent unicode changes only affected windows. | 11:43.47 |
| kens | Robin_Watts : yes, but I have a sneaky suspcicion it *is* a Windows problem, but he's opened the reoprt using a Mac :-) | 11:48.04 |
| | Of course, wihtout example files.... | 11:51.12 |
| Robin_Watts | The problem is that I don't believe the content of the example files matters. | 11:51.59 |
| | I think it's just the NAME of those files that matters. | 11:52.09 |
| kens | agreed | 11:52.14 |
| Robin_Watts | and if he sends us those files via bugzilla, the name will get converted, and the problem lost. | 11:52.41 |
| kens | If he zips them ? | 11:52.50 |
| Robin_Watts | zip uses 8 bit names, I think. | 11:53.01 |
| kens | Well then we are probably stuck | 11:53.10 |
| Robin_Watts | And of course, the most interesting point.... | 11:53.21 |
| | I've not committed any unicode changes yet. | 11:53.29 |
| kens | Ah, curiouser and curioser | 11:53.41 |
| | He claims its the 9.07 pre-release | 11:54.15 |
| | and is OK with 9.06 | 11:54.21 |
| | But from the descirption I don't see how it can be a pdfwrirte problem, or anythign except the filenames | 11:54.38 |
| Robin_Watts | indeed. | 11:54.43 |
| kens | So I guess I'm puzzled. If he sends the files and they work here, then I can say with some reliability its the filenames | 11:55.26 |
| Robin_Watts | Well, he says the files work for him when renamed, so I can't see how it can be anything other than the filenames. | 11:56.56 |
| kens | can you suugest a course of action to investigate it ? | 11:57.50 |
| Robin_Watts | I am doing so now. | 11:59.49 |
| kens | OK thanks Robin_Watts you are probably thinking clearer than me right now | 12:00.44 |
| Robin_Watts | Bugger. Should probably have said "matsubara-san" rather than "scott" :( | 12:17.54 |
| kens | :-) | 12:18.06 |
| | Goodies! | 13:21.03 |
| | New PC has arrived.... | 13:21.08 |
| Robin_Watts | kens: Nice. | 13:46.18 |
| | tor8: What VS did you build the 1.1 binaries with ? | 13:46.32 |
| tor8 | 2010, I think | 13:46.55 |
| Robin_Watts | Ah. | 13:47.09 |
| | So that might explain why they don't work on XP ? | 13:47.20 |
| tor8 | it's hard to check now, because I had to reinstall windows when I upgraded motherboard and cpu | 13:47.29 |
| | Robin_Watts: that would probably be the case, yeah | 13:47.43 |
| Robin_Watts | Damn. Is it worth us rebuilding with VS2005? | 13:48.02 |
| tor8 | Robin_Watts: want to recompile with 2008 and upload again? | 13:48.06 |
| Robin_Watts | 2005 would be better (goes back even further, and loses us nothing). | 13:48.22 |
| | Either your or I should, yes, cos I think that may be behind bug 693360 | 13:48.43 |
| tor8 | performance? how's the optimizer on 2005 compared with newer msvcs? | 13:48.51 |
| Robin_Watts | I've never seen a difference. | 13:49.04 |
| | I use 2005 all the time. | 13:49.12 |
| tor8 | 2005 doesn't have an express edition though, or have I missed it? | 13:51.01 |
| Robin_Watts | possibly not. | 13:51.11 |
| | http://apdubey.blogspot.co.uk/2009/04/microsoft-visual-studio-2005-express.html | 13:54.13 |
| | There you go. | 13:54.16 |
| | Morning mvrhel_laptop. Are you still buried in 330 stuff? | 14:33.26 |
| mvrhel_laptop | hehe. Hi Robin_Watts. Yes reading through an email from them right now on some clarification that I asked for | 14:34.09 |
| Robin_Watts | ok. I've hit a point in investigating that SEGV where I think I could do with a nudge from either you or ray. | 14:34.39 |
| | but it can wait until you're feeling less pressured. | 14:34.58 |
| mvrhel_laptop | ok. if ray is able to help you that would be good. I am going to go eat breakfast now then I will be back to add in support for the use of overriding device link profiles that vary with object type and may map to NCLR or to CMYK even when the destination profile is NCLR. | 14:38.39 |
| | also need to look at still doing overprint simulation even when we replace a spot color with process colors. This will require changes in the overprint compositor to "blend" instead of simply writing a particular color | 14:40.08 |
| Robin_Watts | Bah. Vanishing tor8. | 14:58.24 |
| sebras | Robin_Watts: I feel the same, but over another protocol. :-/ | 15:06.10 |
| Robin_Watts | sebras: I just pushed your 2 patches. | 15:07.01 |
| | and one of tors. | 15:07.11 |
| | 3 new patches up on my repo for review. | 15:07.24 |
| sebras | Robin_Watts: great, thanks! I'll have a look at your patches tonight and bug you about them during the weekend. :) | 15:09.43 |
| Robin_Watts | :) | 15:09.53 |
| kens | Robin_Watts : MuPDF isn't building did yo ufix it in the set of patches ? | 15:11.51 |
| Robin_Watts | kens: In what way isn't it building? | 15:12.40 |
| kens | the cluster regression says it isn't | 15:12.51 |
| | An error occurred that prevented the local cluster run from finishing: | 15:13.06 |
| | macpro reports: compileFail: mupdf | 15:13.06 |
| | compileFail: mupdf | 15:13.22 |
| | makemupdf.out (last 50 lines): | 15:13.22 |
| | ^ | 15:13.22 |
| | apps/x11_main.c:3:10: fatal error: 'X11/Xlib.h' file not found | 15:13.22 |
| | #include <X11/Xlib.h> | 15:13.22 |
| | ^ | 15:13.22 |
| Robin_Watts | Our survey blames henry. | 15:14.03 |
| kens | I couldn't possibly comment | 15:14.21 |
| | But all your regression tests are failing to run | 15:14.29 |
| Robin_Watts | I've taken macpro out of the cluster and rescheduled the missing jobs. | 15:16.13 |
| kens | OK | 15:16.17 |
| Robin_Watts | Thanks for spotting that. | 15:16.23 |
| kens | NP | 15:16.36 |
| | was waiting for new BP to finish WIndows setup | 15:16.44 |
| Robin_Watts | BP ? | 15:16.57 |
| kens | PC, sorry, I'm quite tired now | 15:17.05 |
| Robin_Watts | :) | 15:17.09 |
| kens | Oh god, which piece of paper had the product activation key on it :-( | 15:19.39 |
| Robin_Watts | The product activation key should be on a sticker attached to the side of the PC, right? | 15:20.29 |
| kens | Round the back,where its nice and dark :-( | 15:23.39 |
| sebras | Robin_Watts: btw, when you review my patches -- can you especially look for instances where I have function (arg1, arg2) instead of function(arg1, arg2)? this is an annoyance I have to adhere to at work, so I'm worried it will carry over to mupdf code... | 15:26.02 |
| Robin_Watts | sebras: Extra spaces? Sure. | 15:27.56 |
| sebras | Robin_Watts: yes, tnx. | 15:28.50 |
| Robin_Watts | tor8: Various patches on my git repo for you to look at. | 16:21.21 |
| | Ignore the function one and the pdfwrite one. | 16:21.32 |
| | In fact, I've pushed again without the pdfwrite one. | 16:22.04 |
| tor8 | Robin_Watts: you want *both* an abort and abort_ptr? I'd pick one and stick with it. | 16:29.27 |
| Robin_Watts | tor8: Well, if I had it to do over, I'd pick abort_ptr, I think. | 16:29.48 |
| | But I think there may be people out there using abort already. | 16:30.06 |
| tor8 | hmm. still don't like it :/ | 16:30.41 |
| Robin_Watts | tor8: Fair enough. | 16:30.57 |
| | Thinking about it, I'd like to have a run at persuading zeniko to try a different way of working. | 16:31.33 |
| | So let's leave that one for now. | 16:31.37 |
| tor8 | the other commits are fine, I haven't studied the fz_function in detail but I assume it's mostly a renaming shuffle and adding indirection by function pointers | 16:31.39 |
| Robin_Watts | tor8: It is. | 16:31.46 |
| | I haven't actually made use of the fz_function one yet, so until I do, it doesn't need to be pushed. | 16:32.02 |
| Robin_Watts | does the ray summoning dance (gently grins fresh coffee under the webcam). | 16:44.12 |
| kens | grins ? | 16:44.34 |
| mvrhel_laptop | grinds perhaps | 16:44.44 |
| Robin_Watts | bah. I blame my keyboar. | 16:44.55 |
| kens | AH, a Key Boar to look after your keys | 16:45.22 |
| Robin_Watts | Why are there 4 d's in "Edward Woodward" ? | 16:46.09 |
| kens | OK off now, goodnight all. | 16:55.46 |
| Robin_Watts | night kens. Have a good weekend. | 16:56.04 |
| mvrhel_laptop | Robin_Watts: did you want me to call Ray? | 17:39.31 |
| Robin_Watts | I'll text him,I guess. | 17:40.02 |
| henrys | everywhere I look I see mupdf ... nice ... http://www.youtube.com/watch?v=moiRFi6Ttp4 | 17:53.30 |
| tor8 | henrys: customer of ours or some hack? | 18:14.03 |
| Robin_Watts | They say GPL. | 18:14.19 |
| henrys` | no not a customer. | 18:14.47 |
| | his code is here:https://github.com/hwhw/kindlepdfviewer | 18:17.31 |
| Robin_Watts | I have texted ray, but no response. I guess he's busy on an errand or something. | 18:37.19 |
| mvrhel_laptop | Robin_Watts: if you need me to look things over I can try to help. | 18:44.26 |
| Robin_Watts | well, you may be able to point me in the right direction. | 18:46.29 |
| mvrhel_laptop | ok | 18:46.33 |
| Robin_Watts | I have a file that when run gives a SEGV due to the reader and writer clist getting out of step. | 18:46.53 |
| | The case in question is where we write a fill_trapezoid command. | 18:47.15 |
| | That has to write various color values, and it dutifully writes cldev->clist_color_info.num_components of them (14) | 18:47.44 |
| | When we come to read back, we have a pdf14cmykspot device in place. | 18:48.01 |
| | and it reads tdev->color_info.num_components (17) of them. | 18:48.17 |
| | hence getting out of sync and falling over. | 18:48.28 |
| mvrhel_laptop | is there transparency involved in this file? | 18:48.54 |
| Robin_Watts | No idea. | 18:49.04 |
| mvrhel_laptop | oh run with -dNOTRANSPARENCY | 18:49.14 |
| | or run with -Zv | 18:49.38 |
| | oh sorry | 18:49.47 |
| | I am a little slow | 18:49.51 |
| | missed your comment about pdf14cmykspot | 18:50.04 |
| | device | 18:50.05 |
| | so yes there is transparecy | 18:50.10 |
| | ok | 18:50.15 |
| Robin_Watts | right, yes, I figured there must be. I thought you were asking something more cunning that I was failing to understand :) | 18:50.35 |
| mvrhel_laptop | so for some weird reason the pdf14 device is in a different state during writing vs. reading | 18:50.40 |
| | this sounds like one you should dump upon me | 18:50.47 |
| Robin_Watts | ok, if you're happy for that. | 18:51.03 |
| mvrhel_laptop | well | 18:51.09 |
| | It does make sense. I have run across these before | 18:51.27 |
| Robin_Watts | Feel free to kick it back if you can give me something to try. | 18:51.31 |
| mvrhel_laptop | it will just be a couple weeks before I can get to it though | 18:51.45 |
| | The changes required for customer 330 are getting a bit more complex | 18:52.08 |
| Robin_Watts | So the problem is that we shouldn't have a pdf14cmykspot device at that point in the writer? | 18:52.46 |
| mvrhel_laptop | no we should have one | 18:52.57 |
| Robin_Watts | Does that mean that a remove compositor thing has been missed? | 18:53.00 |
| mvrhel_laptop | and they need to be in the same state | 18:53.04 |
| Robin_Watts | Right. at the moment we have one in the writer, but not in the reader. | 18:53.17 |
| mvrhel_laptop | that is the issue the | 18:53.26 |
| | then | 18:53.28 |
| Robin_Watts | Right, but is the problem that there isn't one in the writer, or that there is one in the reader ? | 18:54.04 |
| mvrhel_laptop | If you can narrow down the band and look at the pushes and pops that occur in that band prior to the filling vs. what happens in the nonclist case, that may help you | 18:54.29 |
| Robin_Watts | I will try reducing the file. | 18:54.43 |
| mvrhel_laptop | I don't understand how any drawing is going on in the writing side, without the pdf14 device in place | 18:54.56 |
| Robin_Watts | At the very least that will help you if I do have to leave it to you. | 18:55.02 |
| | and maybe it'll shrink it enough to fit it into my brain. | 18:55.30 |
| mvrhel_laptop | you can always put a break point in place to see when the pdf14clist writer device is getting installed | 18:55.55 |
| | hold on let me find that | 18:55.59 |
| | so get_pdf14_clist_device_proto | 18:57.14 |
| Robin_Watts | OK. I'm in the middle of android debugging at the moment. | 18:57.51 |
| mvrhel_laptop | that will be the writer clist device | 18:57.53 |
| | it will get the spot proto type | 18:58.02 |
| Robin_Watts | But I'll look at that when I turn my attention back to this on monday. Thanks. | 18:58.13 |
| mvrhel_laptop | during the reading you get the actual pdf14 device | 18:58.15 |
| | ok | 18:58.25 |
| | basically the thing is to see what is going on with the color components for the pdf14 clist writer device | 18:58.49 |
| Robin_Watts | gotcha. When the writer device gets put in place, does it update the clist_color_info.num_components then ? | 18:58.57 |
| mvrhel_laptop | and what is going on with the components for the real pdf14 device during reading | 18:59.09 |
| | Robin_Watts: normally yes. num_components may change | 18:59.54 |
| | however | 18:59.59 |
| | if we are going out to psdcmyk or tiffsep they should not be changing | 19:00.14 |
| | unless a softmask is present | 19:00.20 |
| Robin_Watts | OK. I really need to simplify the file I suspect. | 19:00.39 |
| mvrhel_laptop | when we go out to psdcmyk or tiffsep, blending color spaces are ignored and the data is kept in separations | 19:00.52 |
| | so it is actually even odder that this is failing with psdcmyk | 19:01.03 |
| | ok. I hoped I have you going in the right direction | 19:01.53 |
| | hope | 19:01.56 |
| | back to the salt mine.... | 19:02.03 |
| Robin_Watts | Yeah, that's certainly the kind of nudge I needed. Thanks. | 19:02.13 |
| mvrhel_laptop | fun with device link profiles | 19:02.15 |
| | cool. got the source device link profile working for the non clist case with minimal changes in the code | 19:44.57 |
| | now clist case.... | 19:45.09 |
| | which is when we have image data, since the device link profile will be pickled into the clist | 19:46.09 |
| | this should not be too painful either | 19:46.26 |
| | then I have to worry about the case where the device link is going to an NCLR space | 19:46.43 |
| | i.e. cmyk + spots | 19:46.50 |
| | and then a case, where the device profile is NCLR but the device link profile is going only to CMYK | 19:47.10 |
| | then that should cover the device link profile cases | 19:47.34 |
| | and the clist case works fine | 20:04.57 |
| | time to commit this and then look at the NCLR cases | 20:05.12 |
| | bbiaw | 20:13.15 |
| | nice the NCLR cases both work as is | 21:41.35 |
| | Forward 1 day (to 2012/10/06)>>> | |
IRC Logs