| <<<Back 1 day (to 2015/01/12) | 20150113 |
mvrhel_laptop | henrys: if you can look over the commit in my repos for xpswrite that would be great | 00:08.52 |
| rayjj: might want to take a look too | 00:09.09 |
| hopefully this will fix things for gsview printing in windows for me. | 00:09.47 |
| now I just need to wrap up a couple things in gsview, and get the beta on the web site this week then I will get back on SOT | 00:10.38 |
| bbiab | 00:12.32 |
henrys | mvrhel_laptop: sure | 00:16.00 |
mvrhel_laptop | thanks henrys | 00:35.36 |
rayjj | mvrhel_laptop: I've looked at the rather massive commit, and it seems OK. It's really hard to spot problems in refactored code since the "diff" is pretty much worthless (showing a massive 'cut' in one file and a 'paste' in the other) | 05:03.41 |
| mvrhel_laptop: I just have to assume that there were no changes and it was just "moved" | 05:04.00 |
| (unless I were to have the original source file and new destination with something llike vimdiff that would let me see if there are any actual diffs | 05:05.17 |
| mvrhel_laptop: did you have a reason for naming the image_decode_t "decode_st" instead of just "decode" ? | 05:11.38 |
| mvrhel_laptop: Also in the local 'tiff_set_values' (copied from gstiffio.c ?) it looks like you still have TIFFSetField(tif, TIFFTAG_COMPRESSION, COMPRESSION_LZW); | 05:16.30 |
| mvrhel_laptop: is this really useful if it is going to be zipped in the XPS LZW is not as effective as zip deflate, so the LZW is mostly just a time waste in that zip of the LZW is almost the same as just the zip (just guessing) | 05:18.38 |
| mvrhel_laptop: if you really want me to dig more into it, let me know and I'll set up for side by side files, but I'm inclined to let regression testing make the call | 05:20.24 |
kens | chrisl don't know if this is important or not, but I just ran a bmpcmp and got some 'odd' errors returned which appear to be build errors | 14:30.44 |
| misc errors: | 14:31.27 |
| ../main/pcl6_gcc.mak:153: ../config.mak: No such file or directory | 14:31.27 |
| make[1]: *** No rule to make target `../config.mak'. Stop. | 14:31.27 |
| make: *** [xps-clean] Error 2 | 14:31.27 |
| Each one repeated 5 times | 14:31.40 |
chrisl | kens: that's odd - can you try it again? | 14:33.05 |
kens | Sure, one minute | 14:33.14 |
| It did seem to work correctly though, I got the diffs I expected | 14:33.42 |
chrisl | I've run quite a few cluster runs since I made changes to that stuff yesterday | 14:34.48 |
kens | THe cluster run was fine, it was when I did a bmpcmp | 14:35.06 |
| Same again | 14:37.29 |
| I wonder if the bmpcmp compile is somehow different to the cluster push | 14:38.16 |
chrisl | It suggests the cluster is doing a make (xps-)clean before running autogen.sh - but nothing I've done should change that behaviour | 14:38.39 |
kens | Well it does seem to work OK, but I've not seen that set of errors before. I admit I haven't done a bmpcmp since before Xmas | 14:39.17 |
chrisl | I really, really want to ban Hin-Tak from bugzilla!!! | 14:40.51 |
kens | Oh good grief, Hin-Tak is incapable of leaving it alone..... | 14:40.57 |
chrisl | I guess the bmpcmp warnings only appear in the e-mails | 14:44.53 |
kens | Seems to be the case, yes, do ytou want a copy of the mail ? | 14:45.08 |
chrisl | I'm not sure it will tell me much, but sure.... | 14:45.26 |
henrys | minimally he needs to not use ghostscript.com in his email. | 14:46.25 |
kens | OK | 14:47.00 |
| Mail on its way | 14:47.04 |
chrisl | Well, as far I could tell, the ghostscript.com addresses are not longer active, but that's why I mailed Marcos to ask about it | 14:47.25 |
kens | Hmm, then we should probably unsubscribe anyone using a ghostscript.com email address from Bugzilla too | 14:48.00 |
chrisl | I guess I should say *my* ghostscript.com address is not longer active | 14:48.29 |
kens | Hmm, I'm guessing Hin-Tak is only getting these as email because he's on gs-bugs. | 14:49.10 |
| He hasn't added himself as a recipient, so its not going to his ghostscript.com email address. | 14:49.31 |
| Of course, I don't know any other email address for him. | 14:49.49 |
chrisl | Er, well, in the cluster code, "build.pl" does not actually do the builds...... | 14:52.19 |
kens | THat seems odd | 14:52.39 |
chrisl | I don't really understand what the cluster is doing, but it does look like it is possible for it to run "make xps-clean" before it runs autogen.sh - which is bad | 14:57.19 |
kens | Yes, that sounds like a bad plan :-) Maybe one for Marcos to look at ? | 14:57.51 |
chrisl | I'm baffled as to why it would arise now, hoever | 14:58.33 |
| however | 14:58.36 |
| Oh, I wonder if it's because you had only PS and XPS diffs and no PCL diffs...... | 14:59.48 |
kens | Boggles | 15:00.07 |
Robin_Watts | For Henrys: https://www.indiegogo.com/projects/the-level-a-deck-designed-to-set-you-in-motion | 15:00.40 |
chrisl | If the bmpcmp code only builds the products for which it needs to generate diffs, there might be a problem | 15:02.52 |
kens | Umm, I guess I can see that | 15:03.06 |
henrys | Robin_Watts: wow cool I have been thinking treadmill but that might be better. | 15:03.51 |
chrisl | Aha, there is a bug in the xps case...... | 15:06.11 |
kens | Well I was more concerned that it might be a build problem, rather than a cluster build problem | 15:06.45 |
chrisl | No, definitely a cluster problem, and seems to have been around for a while | 15:07.11 |
kens | Strange that I should just trip over it now, oh well. | 15:07.23 |
| It doesn't actually seem to cause a problem though | 15:07.34 |
chrisl | It needs XPS to be built and *not* PCL - possibly a rather rare occurence | 15:07.57 |
kens | Hmm, I guess so, xps doesn't change so much | 15:08.20 |
chrisl | It very clearly does: "nice make xps-clean ; nice ./autogen.sh........." | 15:09.01 |
kens | ROFL | 15:09.16 |
chrisl | I'll point marcos at it when he appears (if I remember) | 15:10.20 |
kens | I'll try and remember too. I assume we're having a meeting in 20 minutes ? | 15:10.38 |
chrisl | I would think so - good time for a beverage, I reckon...... | 15:12.21 |
kens | Good plan | 15:12.32 |
henrys | meeting at the 1/2 hour | 15:13.13 |
| not sure if this broke when chrisl redid the ghostscript.com but source links in the documentation are broken. For example try accessing an .[ch] file from http://ghostscript.com/doc/current/Develop.htm | 15:31.26 |
chrisl | Those have been broken since we moved to git | 15:32.21 |
mvrhel_laptop | hi rayjj: I changed decode_st to decode in the commit. not sure what I was thinking when I did that. Also, the tiff content is getting compressed right now as the zlib contents are not. I agree that we don't want to double compress. when we turn on the compression for the zlib contents then we will want to have libtiff create content uncompressed | 15:32.33 |
Robin_Watts | A spot of apache mod_rewrite would fix that. | 15:33.00 |
henrys | three cheers the us folks seem to have health insurance now. You can now resume dangerous activities. | 15:33.09 |
Robin_Watts | Redirect to within the git repo. | 15:33.13 |
henrys | chrisl: oh never notices. | 15:33.29 |
Robin_Watts | Henrys: Like living in the US? :) | 15:33.34 |
henrys | s/notices/noticed | 15:33.37 |
kens | is finished dangerous pursuits for the time being | 15:33.44 |
rayjj | mvrhel_laptop: I think I understand that, except for "zlib contents are not [compressed]" | 15:34.01 |
mvrhel_laptop | right now there is no compression going on | 15:34.13 |
| not sure how else to say it | 15:34.30 |
henrys | chrisl, Robin_Watts: would be nice to fix. | 15:34.53 |
rayjj | you are generating a zlib with no compression ??? | 15:34.55 |
mvrhel_laptop | I am not | 15:34.59 |
kens | marcosw chrisl has diagnosed a bug in the bmpcmp cluster scripting. See earlier in the logs | 15:35.03 |
mvrhel_laptop | but xpswrite seems to be | 15:35.29 |
chrisl | henrys, Robin_Watts I can problem fix it with a tweak to the cron script that updates the "current" docs | 15:35.37 |
henrys | mvrhel_laptop: I've reviewed but it's too large for me to be confident... I wish we could have things like this be 10 commits and a merge... | 15:36.07 |
marcosw | kens: okay, I'll take a look at the logs | 15:36.22 |
mvrhel_laptop | henrys: it would be hard to do that | 15:36.25 |
| there was just too much wrong | 15:36.30 |
kens | marcosw at around 15:00 | 15:36.44 |
chrisl | marcosw: if you check line 1500 in run.pl it does a "make xps-clean" before running autogen.sh | 15:36.45 |
rayjj | mvrhel_laptop: I agree -- most re-factors are difficult to sub-divide | 15:36.50 |
mvrhel_laptop | I did a ton of testing though | 15:36.58 |
| on a variety of image types | 15:37.14 |
| and it fixes at least 3 bugs | 15:37.20 |
| not to mention several that were never reported | 15:37.30 |
henrys | mvrhel_laptop: okay... but I don't agree, note you can break the tree with a merge just the merge point will be tested. Not sure if folks are aware of that. | 15:37.55 |
rayjj | mvrhel_laptop: it would be best (but not essential, AFAIC) to have the re-factor and fixes be separate commits, but I don't want to make more work. | 15:38.19 |
mvrhel_laptop | I suppose, I could have pulled the stuff out of the clist as a separate commit | 15:39.08 |
henrys | mvrhel_laptop, Robin_Watts we are about to own ETS. Miles is buying it out from Raph - the patent and all. Do we want to do something with it? I know you guys improved it should we try to extend the patent with the new changes. | 15:39.21 |
chrisl | You can always keep lots of commits in your own repo for review, then only squash them for inclusion into the main repo | 15:39.25 |
mvrhel_laptop | yes | 15:39.36 |
| I always find stuff wrong though and keep amending the current commit :) | 15:40.00 |
Robin_Watts | henrys: Not sure the changes are patentable. | 15:40.07 |
rayjj | part of the problem is that the code only ever partly followed the patent, and now even less | 15:40.07 |
mvrhel_laptop | until it gets away from me | 15:40.08 |
chrisl | mvrhel_laptop: that is a hard habit to break! | 15:40.33 |
rayjj | chrisl: agreed. | 15:40.55 |
Robin_Watts | mvrhel_laptop: I often end up chiselling fixes off the monolith at the end. | 15:41.10 |
rayjj | Robin_Watts: me, too | 15:41.24 |
mvrhel_laptop | I suppose. This is not that big of a commit though. Its relatively isolated | 15:41.49 |
Robin_Watts | Look at the commit in gitk, spot things that are separable, 'exclude' them from the commit. Add them into a new one, then rebase that before the monolith. | 15:41.53 |
| sure. | 15:41.57 |
rayjj | which I usually do by copying the file I want to end up with off somewhere "safe" | 15:42.15 |
henrys | mvrhel_laptop: anyway I'm fine with I think I saw some refactoring that can be done writing to the zip file but we can do that later. | 15:42.16 |
mvrhel_laptop | henrys: yes. we will want to do some refactoring there, as well as fix an issue | 15:42.40 |
| we can't really have all the temp file handles open as we create the document | 15:42.50 |
| we will run into issues with documents that have a lot of pages | 15:43.06 |
henrys | I'd prefer we save the small commits and do a merge but I won't argue. | 15:43.16 |
chrisl | henrys: that could make bisect even more precarious | 15:44.01 |
rayjj | mvrhel_laptop: how many open files per page ? is it a fixed number, or dependent on the contents? | 15:44.03 |
Robin_Watts | henrys: I'd prefer it if the fixes were on master and therefore testable. | 15:44.09 |
| but in the case where that is not feasible, a merge is preferable. | 15:44.26 |
marcosw | chrisl: not sure why doing a 'make xps-clean' before the 'autogen.sh' is wrong (haven't had my morning coffee, so be gentle when you explain it). | 15:44.37 |
mvrhel_laptop | I only have 1 image temp file at a time now, which originally I did not have have as I was allowing the zip up at the end to close them. Now there is just 1 image file open but it is put in the zip file as soon as end_image occurs | 15:44.46 |
henrys | chrisl: curious how the library stuff is coming, we did have a shared language customer "nibble" which alway makes me think of this project. | 15:44.56 |
mvrhel_laptop | the icc profiles are stuffed directly from memory into the zip file | 15:45.02 |
chrisl | henrys: I did very little work on it while I was on holiday........ | 15:45.19 |
mvrhel_laptop | other files, like the xml page contents are still left open until the end. this will need to be fixed | 15:45.41 |
chrisl | marcosw: autogen.sh creates (part of, in this case) the makefile - without the complete makefile, running make is a bad idea..... | 15:45.42 |
henrys | chrisl: oh of course. I'll harass you in 2 weeks ;-) | 15:45.48 |
chrisl | henrys: thanks! | 15:45.58 |
henrys | tor8: what became of the js comparision for scott and miles? | 15:46.17 |
mvrhel_laptop | at least that is the way it appears to me | 15:46.21 |
chrisl | henrys: I'm girding myself to tackle the Windows stuff next | 15:46.25 |
Robin_Watts | marcosw: 3 days to viva, right? | 15:46.33 |
henrys | chrisl: we can't have bisect ignore branches? | 15:46.44 |
chrisl | henrys: I thought you were saying you wanted all the interim commits on master | 15:47.25 |
marcosw | Robin_Watts: yeah, submitted updated dissertation draft on Sunday and have my practice defense talk today. | 15:47.30 |
Robin_Watts | marcosw: That's where you get to say things like "well, your momma might think that"... | 15:47.55 |
rayjj | chrisl: I understood it as just wanting 'fixes' separate from the re-factor commits (I don't like having "interim" commits that don't work) | 15:48.36 |
kens | marcosw you need us to look after support for a few days ? | 15:49.02 |
chrisl | Ah, misreading on my part, then | 15:49.03 |
Robin_Watts | marcosw: best of luck etc. | 15:49.04 |
marcosw | chrisl: right. I see the pcl case used to do that and was changed to a 'rm -fr main/obj' before the autogen.sh. | 15:49.09 |
henrys | chrisl: no the branch should exist though, the merge point and branch point shouldn't break the build but anything on the development branch is fine. Are we agreeing with that? | 15:49.11 |
tor8 | henrys: I've got numbers assembled, but I want to track down some more different benchmarks to try out | 15:49.34 |
marcosw | Robin_Watts: thanks. | 15:49.36 |
chrisl | henrys: I generally keep all my commits on my own branch, and only squash on master | 15:49.55 |
tor8 | henrys: christmas got in the way... | 15:50.08 |
mvrhel_laptop | rayjj: so 695754 is fixed. I tested each of the files with xpswrite through gxps | 15:50.49 |
marcosw | kens: probably would be a good idea; i've been busy the last few days and so have been letting things slide a bit and it's going to get worse before it gets better. | 15:51.09 |
henrys | The squash is what I don't like. That is why we look at code and have no idea how it got to where it is. | 15:51.12 |
kens | marcosw no problem | 15:51.23 |
| henrys but if you load all the commits that were done on a branch you stand a pretty good chance of a bisect fcailing because one or more of those commits doesn't build/work | 15:51.58 |
chrisl | henrys: without the squash, you get all the interim commits on master | 15:52.04 |
tor8 | henrys: any word from URW about those fonts we ordered? | 15:52.10 |
marcosw | henrys: don't say bad things about squash; how would you feel if chrisl said he didn't like triathlons? | 15:52.22 |
henrys | tor8: nothing ... thanks for reminding me I'll ping again. | 15:52.47 |
| marcosw: ;-) | 15:52.54 |
chrisl | The way my legs feel, you can say what you want about squash..... :-( | 15:53.14 |
henrys | is this a customer or a person using Luratech code with GPL ghostscript: http://bugs.ghostscript.com/show_bug.cgi?id=695768 | 15:53.19 |
| ? | 15:53.20 |
chrisl | Nobody I've ever heard of | 15:53.41 |
rayjj | kens: not if you use commit --amend as you apply fixes to the commit in work on a branch | 15:54.03 |
kens | huh ? | 15:54.23 |
marcosw | henrys: no idea about the bug you mentioned. | 15:54.41 |
chrisl | henrys: is there any restriction on using Luratech the AGPL release - as long as it's not for distribution, of course | 15:54.55 |
rayjj | kens: I try to keep each commit runnable, even if it isn't "done" yet | 15:54.59 |
marcosw | am now curious what license keys are in the context of JP2. | 15:55.18 |
chrisl | Should have been: "using Luratech with the AGPL release" | 15:55.20 |
Robin_Watts | rayjj:Right, and that's specifically what henrys is asking us not to do. | 15:55.32 |
henrys | marcosw: and how can that be broken, wouldn't customer be hollering | 15:55.38 |
rayjj | so that when the commits are pushed from a branch, each one builds and runs, even if it still has bmpcmp issues | 15:55.46 |
Robin_Watts | kens is correct. If we allow development branches to have unrunnable or otherwise broken steps in, and we use merges on the branch, it DOES cause bisect problems. | 15:56.28 |
rayjj | Robin_Watts: no, each commit has a single "reason" -- just mistakes have been squashed by using commit --amend along the way | 15:56.45 |
henrys | rayjj: no the merge is tested not the commits on the branch. | 15:56.46 |
kens2 | Can't we simply review the branch then squash for commit ? | 15:57.12 |
Robin_Watts | henrys: The cluster will only test the merge, this is true. | 15:57.17 |
rayjj | henrys: generally it's rare for me to have more than one commit on a branch anyway, so it doesn't happen often | 15:57.27 |
Robin_Watts | But subsequent bisects will be broken by it. | 15:57.30 |
| kens: That is a reasonable suggestion. | 15:57.38 |
chrisl | kens2: on the rare occasions it seems necessary, that's what I do | 15:58.02 |
Robin_Watts | rayjj: If we all only had 1 commit on a branch this question would be moot :) | 15:58.05 |
kens2 | I have 41 commits on my current branch | 15:58.34 |
rayjj | Robin_Watts: right, but I want to avoid having the trunk have commits like "DeviceN" | 15:58.37 |
mvrhel_laptop | I had so many incarnations of this xpswrite commit, I think a pile of commits would be even harder to review | 15:58.38 |
henrys | chrisl: are we inbetween little cms releases. Worried about forks? | 15:58.44 |
rayjj | kens2: 41 different reasons ? | 15:59.07 |
kens2 | Yes | 15:59.14 |
marcosw | henrys: that's what I was wondering. seems like it should be wrong for everyone. unless JP_LICENSE_NUM_1 is usually not defined... | 15:59.14 |
Robin_Watts | rayjj: The need to write decent commit messages for each commit is completely orthogonal to this discussion. | 15:59.20 |
chrisl | henrys: we're, I think, a release behind with LCMS... forks? | 15:59.33 |
Robin_Watts | chrisl: We should be tracking lcms really. | 15:59.48 |
marcosw | kens2, et.al.: I think I've fixed the xps cluster build issue. it only happened if xps needed to be built but not pcl. | 15:59.55 |
Robin_Watts | Certainly (for a while at least), it was us driving it. | 15:59.58 |
| I think we're on 2.6, which is the one we funded. | 16:00.06 |
kens2 | marcosw tht's what chrisl said yes | 16:00.07 |
mvrhel_laptop | forks. its a town in the Olympic peninsula | 16:00.10 |
henrys | chrisl: I don't like the busines of forking our little cms if we can take a release, is what I meant> | 16:00.11 |
mvrhel_laptop | lots of vampires | 16:00.14 |
rayjj | Robin_Watts: maybe, but it's hard to imagine having 41 different commits adequately described (with some detail) with a single log message | 16:00.24 |
chrisl | henrys: what I committed isn't really a fork | 16:00.32 |
kens2 | Michael's xps commit failed to compile | 16:00.33 |
mvrhel_laptop | crap | 16:00.37 |
| I did a cluster push before | 16:00.42 |
| :( | 16:00.43 |
Robin_Watts | rayjj: Who said anything about 41 different commits having a single log message? | 16:00.52 |
mvrhel_laptop | grrrr | 16:01.06 |
kens2 | Robin_Watts : If I squash them up for commit, they will have a single log message | 16:01.13 |
marcosw | mvrhel_laptop: your recent clusterpush didn't compile on the cluster due to: | 16:01.20 |
| ../gs/devices/vector/gdevxps.c:2047: error: 'for' loop initial declarations are only allowed in C99 mode | 16:01.20 |
rayjj | Robin_Watts: if kens squashes his 41 commits, then that's one log message | 16:01.21 |
marcosw | apparently whatever version of visual studio you use is more forgiving :-) | 16:01.55 |
rayjj | kens2: can't you put some of them together on the branch with interactive mode | 16:02.01 |
Robin_Watts | Right, so we have a single commit message that contains all the existing 41 commit messages, plus an introduction paragraph? | 16:02.08 |
paulgardiner | Are we sure merging breaks git bisect? | 16:02.08 |
mvrhel_laptop | wtf. how did that pass the cluster push then? | 16:02.11 |
marcosw | mvrhel_laptop: no idea. presumably a bug in the cluster code :-) | 16:02.42 |
chrisl | Robin_Watts: it was agreed some time ago that mvrhel_laptop would handle our tracking of LCMS releases (although I think you've done most of the leg work on it). So I ignore it | 16:02.46 |
henrys | I'm just saying do little commits on your own branch and merge the branch so we can look back a year from now and see what though process brought us from commit n to n+1 on master. If each little commit is going to be tested by bisect this won't work at all. | 16:02.54 |
Robin_Watts | IME msvc is the most likely thing to barf on that. | 16:02.58 |
rayjj | kens2: that means that which files go with which part of the change is lost to git | 16:02.59 |
henrys | s/though/thought | 16:03.16 |
Robin_Watts | henrys: Yes, every little commit is going to be tested by bisect. | 16:03.28 |
mvrhel_laptop | ok. sorry about that. is there an easy way to back out the commit, or should I just do the fix and commit that? | 16:03.31 |
kens2 | rayjj I don't know what you mean. The way I see it I am doing a pile of work for a specific reason. When I commit the code I'm happy for that to say 'work done for this reason'. But the *many* intermediate steps that I took, I would like to keep as separate commits, so I know what I did and why. But I *don't* want to commit them as 41 cvommits, because for sure some of them will not work | 16:03.32 |
Robin_Watts | And there is no easy way about that. | 16:03.38 |
paulgardiner | Oh it's not merging that's the problem, it's that the branch goes through broken states. Is that what you're saying? | 16:04.13 |
Robin_Watts | mvrhel_laptop: If the commit that broke it is the last one, force push a fix, and tell me and I'll fix the cluster. | 16:04.17 |
| paulgardiner: Yes. | 16:04.27 |
mvrhel_laptop | Robin_Watts: ok. I will hurry then... | 16:04.34 |
kens2 | paulgardiner : yes, for me at least, I don;t cluster test each commit on a branch while I'm working, so its piossible some will not work on all interpreters | 16:04.43 |
mvrhel_laptop | and I found 4 spots where I did this | 16:05.08 |
| bad habit. not sure where I picked that up. I think SOT does this quite a bit | 16:06.19 |
| so still curious why the cluster push did not catch it | 16:07.05 |
marcosw | mvrhel_laptop: i'm worried that your clusterpush didn't fail. there shouldn't be any differences in the options used to compile user tests and commit tests. | 16:07.08 |
Robin_Watts | SOT does not do that at all. | 16:07.17 |
henrys | Robin_Watts: wow that really sucks. | 16:07.19 |
mvrhel_laptop | hmm | 16:07.21 |
Robin_Watts | All SOT code is C89. | 16:07.27 |
henrys | but you seem to be right. | 16:07.28 |
rayjj | kens2: I guess as long as the changes needed for the squashed commit are discussed, then I guess the "all one reason" are OK, which is different to a squashed commit with all of the design change details gone with a log message like "DeviceN" | 16:07.31 |
Robin_Watts | henrys: Had to happen eventually. | 16:07.47 |
kens2 | rayjj I find the intermediate commit messages are often only useful to me (and not always even then, some are cryptic). | 16:08.16 |
mvrhel_laptop | Robin_Watts: yes it does in spots | 16:08.44 |
henrys | I guess squashing is needed ... sigh. | 16:08.48 |
Robin_Watts | Writing better commit messages is something we can fix :) | 16:08.50 |
rayjj | kens2: I'm more concerned with describing the nature of the changes. | 16:08.50 |
kens2 | If we want to discuss branches merging etc, it would be easier to do round the table in March | 16:08.53 |
henrys | kens2: right | 16:09.15 |
rayjj | Robin_Watts: yeah, but we can't fix it once we've pushed the commit, so pre-review (at least of the log message) is critical | 16:09.31 |
kens2 | That's odd, I just got a bmpcmp message | 16:09.33 |
mvrhel_laptop | but I guess the dumb c# code is the culprit | 16:09.38 |
henrys | anything else for the meeting? Sorry we went over. | 16:09.49 |
mvrhel_laptop | I did that type of thing all the time in the gsview code | 16:09.56 |
Robin_Watts | mvrhel_laptop: All SOT code is supposedly C89 compliant. In particular I had to fix some code that crept in last week with variable declarations not being at the top of blocks because it was breaking autobuilds. | 16:10.01 |
marcosw | kens2: I ran a bmpcmp job to test the make xps-clean fix. | 16:10.05 |
mvrhel_laptop | Robin_Watts: do a search on for (int | 16:10.17 |
kens2 | Ah, well it gave me the same error messages marcosw | 16:10.17 |
marcosw | apparently I didn't fix it :-( | 16:10.31 |
henrys | marcosw: you do test the luratech build right? | 16:10.32 |
Robin_Watts | rayjj: We can fix any commit message we like before it hits golden. | 16:10.36 |
rayjj | I'm having to develop a 'dump' for a clist. I'll probably put it in toolbin | 16:10.38 |
marcosw | henrys: yes, during the weekly regression tests. | 16:10.43 |
Robin_Watts | This is why reviews are good :) | 16:10.46 |
rayjj | Robin_Watts: hits golden ??? | 16:10.51 |
marcosw | also I'm pretty sure customers would have compiled about this before now... | 16:11.03 |
| ^compiled^complained | 16:11.13 |
jogux | mvrhel_laptop: what Robin means is you guys all refuse to drop windows or move to a decent toolset that actually supports C99 ;-) | 16:11.24 |
Robin_Watts | mvrhel_laptop: I see hits in platform/tgv/apps/activex and platform/tgv/apps/iphone-alien | 16:11.33 |
| so those are platform specifics. | 16:11.42 |
mvrhel_laptop | ah ok | 16:11.45 |
Robin_Watts | and in the gsoap test code. | 16:11.51 |
marcosw | C99 is a bad idea, we should go back to K&R :-) | 16:11.59 |
mvrhel_laptop | :) | 16:12.03 |
henrys | marcosw: ah do you have a test that uses encoding though? I think the user has found a problem but he is using luratech illegally. | 16:12.39 |
mvrhel_laptop | Robin_Watts: what if I amend my commit here and then do the force | 16:12.44 |
| will that be fine? | 16:12.53 |
Robin_Watts | mvrhel_laptop: I'll need to prod the cluster, but yes, that's fine. | 16:13.06 |
chrisl | henrys: illegally how? | 16:13.06 |
henrys | kens2: are we using luratech for any pdfwrite encoding? | 16:13.10 |
mvrhel_laptop | Robin_Watts: ok | 16:13.15 |
kens2 | henrys, absoltely not, unless you specifically change some compilation flags | 16:13.28 |
henrys | chrisl: he's linking luratech with gpl ghostscript. | 16:13.29 |
kens2 | Isn't he OK to do that if he has a Luratech licnece ? | 16:13.48 |
rayjj | henrys: a licensed customer ? | 16:13.49 |
henrys | chrisl: had luratech gone free? | 16:13.52 |
rayjj | henrys: not that I'm aware of | 16:14.06 |
chrisl | henrys: As long as he's not distributing it, that's legal, isn't it? | 16:14.09 |
henrys | rayjj: I just asked is he was a customer and you guys said no. | 16:14.14 |
marcosw | henrys: no, there isn't any jpeg2000 output tests. it would be work to add those. | 16:14.16 |
mvrhel_laptop | Robin_Watts: ok I did the forced push | 16:14.27 |
| how do you start up the cluster? | 16:14.40 |
rayjj | henrys: I missed who we are talking about (that is not a customer) | 16:14.56 |
kens2 | rayjj its a bug report | 16:15.07 |
| 659768 | 16:15.32 |
rayjj | which bug # | 16:15.47 |
kens2 | ^^ | 16:15.52 |
henrys | chrisl: yes I was assuming he was distributing the code or using it. | 16:16.05 |
kens2 | http://bugs.ghostscript.com/show_bug.cgi?id=695768 | 16:16.15 |
rayjj | sorry I was up looking at the logs to see if I could find it | 16:16.17 |
Robin_Watts | mvrhel_laptop: git magic :) | 16:16.42 |
chrisl | henrys: well, "using it" yes, but it could be a company's internal use only, or whatever. | 16:16.50 |
kens2 | AIUI if he has a Luratech licence, and is not distributing GS, then he can build it in if he wants. | 16:16.51 |
| I'm not sure about the SaaS usage with AGPL though, he may be in contrvention if he does that | 16:17.30 |
rayjj | henrys: I agree with kens. If he distributed it, he'd have to provide source under GPL, but he can't do that for Luratech | 16:17.55 |
marcosw | mvrhel_laptop: this is scary, I just tried a clusterpush with your broken code and it compiled. i'm worried that clusterpushing is broken. | 16:17.57 |
henrys | I'm not sure that would make the luratech code GPL by the virality of the license. | 16:18.09 |
kens2 | If he's using the code internally he's pretty much clear whatever he does, as far as I know | 16:18.52 |
rayjj | henrys: how else would someone be able to build the end product if it required luratech in the link | 16:18.53 |
chrisl | As long as no (re-)distribution takes place, I believe it's legal | 16:19.18 |
Robin_Watts | mvrhel_laptop: looks to be sorted. Any problems with the cluster not testing commits, please tell me. | 16:19.25 |
rayjj | chrisl: agreed | 16:19.27 |
henrys | anyway back to his issue, if we do use luratech for encoding anything we should test it. Do we? | 16:19.54 |
marcosw | i'm going to disable the cluster for a bit, to try and figure out why clusterpush isn't doing the right thing. | 16:19.56 |
kens2 | I thikn if he's making the app available as a SaaS he would be in trouble, because he can't make the source available | 16:20.07 |
jogux | chrisl: well, plus the AGPL stuff, if he's put it on a server and letting people [outside his company?] use it, he can't use non-GPL code with it without a commercial licnese. | 16:20.08 |
chrisl | jogux: yes, that too..... | 16:20.27 |
jogux | yeah, what kens2 said :) | 16:20.36 |
mvrhel_laptop | Robin_Watts: ok grea. | 16:20.37 |
| great | 16:20.39 |
| marcosw: Thanks for checking that. I was worried I had testing the wrong thing | 16:21.16 |
chrisl | Erm, I didn't think JPX encoded anything...... | 16:21.22 |
| we JPX encoded anything | 16:21.34 |
| (I'm tired...) | 16:21.41 |
kens2 | We have some hooks in pdfwrite to permit JPX encoding | 16:21.57 |
| But they don't do anything at all usually | 16:22.05 |
chrisl | Oh, I thought we only used the JBIG2 encoder | 16:22.06 |
henrys | kens2: well then I bet his bug is good and it won't work. | 16:22.23 |
kens2 | THere are some hooks for both | 16:22.24 |
rayjj | marcosw: I wonder if that relates to my not getting any "new warnings" when I did an oopsey with xps code last week (bare return from a function that is supposed to return a pointer) | 16:22.24 |
kens2 | chrisl I have no idea whether either of JIB2/JPX encoding actually *works*, with any encoder library..... | 16:23.00 |
mvrhel_laptop | Yes. I also had no warnings | 16:23.05 |
| I bet there will be a pile from the commmit :( | 16:23.24 |
chrisl | kens2: I think we have a customer using JBIG2 encoding | 16:23.31 |
kens2 | RTeally ? I guess that does work then..... | 16:23.46 |
rayjj | kens2: I think Igor tested it with luratech when it was put in, but not sure if the cluster luratech test would show it up | 16:23.47 |
mvrhel_laptop | at least it builds now though. | 16:23.50 |
kens2 | rayjj I'm certain the cluster test won't test it | 16:24.02 |
rayjj | I am pretty sure that Raed's outfit (cust 580) is using luratech encoders | 16:24.44 |
henrys | kens2: can we create a test specimen easily? | 16:24.52 |
chrisl | I thought we tested luratech JBIG2 encoding, just not in pdfwrite | 16:24.53 |
kens2 | henrys, certainly, any input file with an image will do, but you need to set the flags correctly. | 16:25.15 |
chrisl | marcosw: don't we test JBIG2 output with luratech? | 16:25.24 |
kens2 | chrisl No idea if we test JBIG2, what I meant is that the cluster won't test it with pdfwrite | 16:25.44 |
henrys | chrisl: what would need jbig compression other than pdfwrite? | 16:26.08 |
kens2 | Or at leats I don't believe it will | 16:26.09 |
chrisl | henrys: we have a jbig2 device | 16:26.41 |
henrys | nvm I'll create a test and have a look. | 16:26.42 |
rayjj | kens2: we'd have to set the Filter distillerparams to force JBIG2 or JPX, right ? | 16:26.52 |
kens2 | correct | 16:26.59 |
| ColorImagerFilter and MonoImageFIlter | 16:27.13 |
marcosw | chrisl: I don't think we do. The only output we test is ppm, tiff, psdcmyk, pdf, ps, and xps. | 16:27.14 |
henrys | if we have a customer using jbig2 in pdfwrite we ought to test it. marcosw has fixed the cluster so we can have parameters right? | 16:27.28 |
rayjj | kens2: so a PS input file that did setdistillerparams and did images _could_ be constructed to test pdfwrite + Luratech | 16:27.57 |
marcosw | henrys: not sure if you could test the luratech tech code with the cluster the way things are currently set up. | 16:28.04 |
chrisl | marcosw: We talked about it, because I fixed the device..... | 16:28.06 |
kens2 | rayjj yes, or you could just send a command line | 16:28.15 |
henrys | based on the user description I don't understand how it would work unless you happened to decode jbig2 before encoding it. | 16:28.23 |
rayjj | kens2: I was thinking about a test that would run as part of the weekly run with luratech | 16:28.42 |
kens2 | We could do that yes | 16:28.50 |
| Turkey would work | 16:29.06 |
henrys | chrisl: that missing 'i' looked odd in that mupdf doesn't fail. That was on my list of discussion items but we've run way over. | 16:30.48 |
mvrhel_laptop | kens2: so I did not get a chance to figure out skiing possibilities before the staff meeting. I can go, but I had not looked into getting a place to stay, car etc | 16:31.07 |
chrisl | henrys: mupdf does fail, in the same way | 16:31.07 |
mvrhel_laptop | oops kens2 left | 16:31.12 |
rayjj | mvrhel_laptop: car should be no problem. | 16:31.43 |
henrys | rayjj, marcosw based on the users report I don't see how encoding would work at all with luratech so this does deserve investigation. | 16:32.13 |
rayjj | mvrhel_laptop: since kens didn't say good bye, he'll probably be back | 16:32.20 |
mvrhel_laptop | ah all kinds of warnings to clean up that did not show up on the cluster push | 16:32.23 |
| tor8: would you like to join us? | 16:32.45 |
marcosw | henrys: I agree, assign the bug to me and I'll look into it (but possibly not until next week). | 16:32.48 |
rayjj | henrys: AIUI, our luratech lib doesn't need keys | 16:32.53 |
mvrhel_laptop | we need to have a head count before we look for a place | 16:33.08 |
rayjj | I think just the three of us so far | 16:33.28 |
mvrhel_laptop | there he is | 16:33.34 |
kens | Stupid network :-( | 16:33.40 |
| Did I miss something ? | 16:33.44 |
rayjj | kens: you're in for skiing, right? | 16:33.54 |
mvrhel_laptop | kens: we were chatting a bit about CO | 16:33.57 |
| skiing | 16:33.57 |
kens | absolutely, yes | 16:34.04 |
kens | waits for logs to catch up | 16:34.22 |
mvrhel_laptop | ok so we need to know if tor8 is also interested and then we need to pick a resort and find a place | 16:34.27 |
marcosw | mvrhel_laptop: your laster clusterpush does not match the commit code that wouldn't build (but it wasn't unmodified either, it had changes from the previous master), so I'm not sure how this happened. | 16:34.35 |
kens | SOunds good to me | 16:34.37 |
marcosw | is it possible you clusterpushed the wrong code (or perhaps the windows clusterpush is broken in some way, I haven't tested it). | 16:34.58 |
Robin_Watts | mvrhel_laptop: I think henry suggested that Miles may be interested in skiing (just not paying for everyone to go skiing :) ) | 16:35.03 |
marcosw | ? | 16:35.04 |
mvrhel_laptop | marcosw: yes I had cleaned up a few minor things | 16:35.05 |
kens | As long as it has snow, lifts, a bed and somewhere to eat I'll be haoppy (oh and a rental place too) | 16:35.06 |
mvrhel_laptop | oh let me check with miles today | 16:35.29 |
chrisl | henrys: having argued that he may, in fact, be doing nothing wrong, I do think it might be good to get Miles to send one of his "nice" letters to the guy on Bug 695768 | 16:35.35 |
kens | mvrhel_laptop : rayjj assume I'm happy with anything you can come up with. I'm good for before or after the staff meeting, don't mind how many days (will adjust flights to match you guys). Don't mind what resort. | 16:36.23 |
rayjj | mvrhel_laptop: kens: what days are we thinking? staff meeting is Fri/Sat | 16:36.52 |
kens | I'm happy with any days. Tue/Wed/Thurs ? | 16:37.15 |
rayjj | I'm happy with skiing Wed/Thur, but to ski Tue, I'd have to fly out Monday | 16:37.47 |
henrys | chrisl: yea that's what I'm doing. | 16:37.57 |
marcosw | At this point I'm very confused as to what the cluster is doing, unfortunately I have to run out in a bit. Will fix the bmpcmp issue kens found and the clusterpush problem mvrhel_laptop is having later today. | 16:37.57 |
kens | Yes, me too, flights arrive Denver in the afternoon | 16:38.03 |
chrisl | henrys: great | 16:38.10 |
rayjj | I haven't checked into red-eyes | 16:38.16 |
chrisl | marcosw: did you see my mail about the ghostscript.com addresses? | 16:38.31 |
kens | There's no possible way for me to get there in the morning, but I'm happy to stay in the airport a night if needed | 16:38.38 |
marcosw | chrisl: yup, will get to that this afternoon as well. | 16:38.55 |
kens | Can we mail Hin-Tak and tell him before we cut it off ? And tell him to reregister on Bugzilla too | 16:39.20 |
chrisl | marcosw: no rush, I just wanted to check - but thanks | 16:39.28 |
rayjj | kens: yes! | 16:40.07 |
kens | rayjj is that re Hin-Tak ? | 16:40.33 |
rayjj | marcosw: I assume that if hin-tak registers with a non ghostscript.com we can zap the database to change all of his stuff over ??? | 16:40.44 |
| kens: yes, re hin-tak | 16:40.56 |
jogux | kens: he could just change his email address, then all the old comments will move to the new email too... | 16:40.57 |
kens | jogux, that's what I was assuming yes | 16:41.11 |
jogux | changing email address can be done by admin or the user. | 16:41.19 |
chrisl | Maybe we should just do it: > /dev/null | 16:41.56 |
kens | Seems Hin-Tak really p*ssed off chrisl today | 16:42.13 |
rayjj | kens: the least expensive non-stop (only $150 RT) gets me in at past midnight | 16:42.16 |
kens | O.O | 16:42.26 |
| I have little choice, its the afternoon or nothing basically | 16:42.37 |
chrisl | kens: between being quite rude to the reporter, and then doing an "I told you so...." post - yeh, pretty much | 16:43.06 |
kens | Not many carriers direct to Denver and I don't want to change (especially with checked baggage) | 16:43.07 |
rayjj | but then I'd be OK with coming Monday to ski Tue/Wed/Thur | 16:43.18 |
Robin_Watts | Change it to hintak@newemailaddressrequired.com and that'll make him take the hint. | 16:43.26 |
rayjj | kens: when are you booked coming in ? | 16:43.32 |
kens | I'd be happy to arrive Monday and ski Tuesday | 16:43.34 |
| rayjj I have not made a booking yet | 16:43.41 |
| Waiting to see what would suit you and Miichael | 16:43.52 |
| and tor8 of course | 16:43.58 |
| and Mile sif he's interested :-) | 16:44.04 |
rayjj | kens: OK. We'll try to get Miles in or out today. | 16:44.28 |
kens | THat would be good. Tor8 are you still interested in ski-ing at all ? | 16:44.48 |
Robin_Watts | IIRC henrys reason for not skiing was because he was looking after those of us who were coming to visit after the meeting. | 16:45.21 |
| If you're skiing before the meeting, potentially Henry could ski too ? | 16:45.34 |
kens | Fine by me, I could use an experienced skier to pick me up :-) | 16:46.06 |
henrys | I don't ski much lately got into it for a while but I snow shoe now with Sabrina. | 16:46.20 |
kens | Never tried that, maybe I should give it a bash | 16:46.35 |
rayjj | kens: you are considerably more experienced than me (totally out of practice) | 16:46.41 |
kens | Well I've just had my yearly practice and managed not to fall off the mountain, so not all bad :-) | 16:47.06 |
| I must go and enter my lift pass and see how far I went | 16:47.23 |
henrys | I don't think shoeing is as fun but it's okay. | 16:47.39 |
rayjj | kens: oh, cool. It tracks it. I always have to keep track myself | 16:47.58 |
kens | The Skiwelt is all linked up and internetted, it tracks which lifts you use | 16:48.30 |
| So you can get a time, distance and height change, as well as a little animation showing where you wnet | 16:48.51 |
rayjj | kens: but then you can't exaggerate it like I do (like most fishermen do with the size of their catch) ;-) | 16:49.06 |
kens | This year Melanie and I skied to the other end of the liked area, 4 mountains over | 16:49.17 |
| I'd like to know how far that was, it took 6 hours to do the round trip | 16:49.32 |
rayjj | kens: how many lifts does that take (from end to end) ? | 16:49.50 |
kens | Hmm, not certain, I'd have to count it up. We also did 2 extra due to making wrong tunrs | 16:50.19 |
rayjj | kens: here (at Mammoth/June) it takes about 7 lifts iirc | 16:50.22 |
kens | I think it would be more for us (round trip) | 16:51.00 |
| One of them was a funicular railway :-) | 16:51.08 |
henrys | the nice thing about shoeing is you frequently see elk sometimes moose and other critters... | 16:52.45 |
rayjj | henrys: that and there is no lift ticket cost. | 16:54.12 |
| henrys: the down side is there are no lifts and no sliding down, so it's a LOT of work | 16:54.39 |
kens | There were a *lot* of Nordic skiers out on the mountain this year, saw 16 on one run..... | 16:55.20 |
rayjj | when my kids were learning snowboard we just went to a hill and trudged up and came down. That was *exhausting*. Luckily, they learned pretty fast (at least enough to take the bunny slope lifts) | 16:56.19 |
| I've never tried nordic, but it is interesting (primarily because there are lots of places to go that don't need passes) | 16:57.24 |
mvrhel_laptop | sorry stepped away. catching up now | 16:57.39 |
rayjj | mvrhel_laptop: you didn't miss anything. We are waiting for Tor to let us know | 16:58.07 |
| and then Miles. | 16:58.22 |
| mvrhel_laptop: oh, and is going up early Tue AM OK with you ? | 16:59.21 |
mvrhel_laptop | rayjj: good question let me check on thing | 17:01.22 |
rayjj | mvrhel_laptop: there's a flight late Monday that gets in Tue past midnight for me. We could either go up that night (and avoid traffic) or early Tue AM and ski Tue/Wed/Thur | 17:01.26 |
mvrhel_laptop | early Tues AM in and getting any skiing in that day would be really difficult | 17:05.20 |
| Tuesday past midnight is a bit late | 17:06.05 |
| rayjj: It would be better get in during the early evening at the latest | 17:06.30 |
| oh I see what you are saying | 17:06.59 |
| sorry rayjj | 17:07.03 |
| I am a little slow this morning | 17:07.07 |
henrys | oh hi fredross-perry | 17:08.27 |
fredross-perry | hi there. | 17:08.36 |
henrys | we have a tuesday meeting 7:30 your time. | 17:08.47 |
mvrhel_laptop | rayjj: though I think driving up after midnight is not going to work so well | 17:08.56 |
fredross-perry | yes I remember now. Thanks. | 17:08.58 |
mvrhel_laptop | rayjj: if we drove up tuesday morning early, we would just need a place for 2 nights at the resort | 17:09.49 |
henrys | fredross-perry: did you have anything for the meeting? You can read the irclogs online to see what we discusssed but there wasn't anything specifically for you. | 17:09.50 |
| fredross-perry: http://www.ghostscript.com/irclogs/current.html | 17:10.14 |
fredross-perry | not just yet. Thanks. | 17:10.29 |
mvrhel_laptop | kens: so you would arrive monday late? | 17:12.18 |
kens | mvrhel_laptop : Monday afternoon | 17:12.30 |
mvrhel_laptop | ok | 17:12.33 |
kens | I cna either hang around if you guys are getting in late, or stay the night if you get in early in the am | 17:12.46 |
Robin_Watts | It strikes me that waiting for ray to get in at gone midnight is going to make life hard for you all. | 17:14.08 |
mvrhel_laptop | that is not going to work | 17:14.27 |
Robin_Watts | If ray gets a (possibly slightly more expensive flight) and lands in the afternoon you could all travel up together. | 17:14.38 |
mvrhel_laptop | I am thinking we may just spend the night at the airport and drive up very early tuesday though | 17:15.03 |
kens | Well I can make any variety work, maximum is I spend a night in an airport hotel | 17:15.06 |
Robin_Watts | Then it's just 1 more night on the rental at the ski resort rather than having to stay in a hotel near the airport. | 17:15.07 |
henrys | fredross-perry: need to give you an artifex address we do first name period last name at artifex.com so it's fred dot ross-perry at artifex.com for you? | 17:15.18 |
mvrhel_laptop | lifts open at 9am at copper (if that is where we are going) | 17:15.18 |
kens | Copper is fine by me | 17:15.30 |
fredross-perry | yes thatâs right. | 17:15.45 |
kens | I have no experience of the area of course :-) | 17:15.49 |
mvrhel_laptop | I have only been to vail | 17:16.02 |
| copper is closer | 17:16.07 |
fredross-perry | and also an official git location perhaps. | 17:16.11 |
Robin_Watts | fredross-perry: Do you have an account on casper? | 17:16.37 |
fredross-perry | no sir | 17:16.54 |
Robin_Watts | OK, so we need to get one of those made. | 17:17.15 |
| Let me have a go. | 17:17.33 |
rayjj | mvrhel_laptop: kens: if you guys are fine with going up early Tue AM (staying at the hotel Tue PM) that's best for me | 17:18.18 |
kens | For me its not a problem | 17:18.39 |
mvrhel_laptop | rayjj: you mean Mon PM? | 17:18.39 |
chrisl | Robin_Watts: don't forget ssh keys etc... | 17:18.46 |
rayjj | mvrhel_laptop: I meant Mon, yes. Sorry | 17:18.56 |
mvrhel_laptop | not a problem for me either. it may actually be a bit less $ wise | 17:18.58 |
| we would have 2 nights at the resort then | 17:19.14 |
Robin_Watts | fredross-perry: OK, so I've created you an account on casper. | 17:19.38 |
mvrhel_laptop | 75 miles to copper so if we left by 6:30 or so we would be there in plenty of time | 17:19.47 |
| of course that is 5:30am for you and me ray... | 17:19.53 |
henrys | fredross-perry: is "perry" capitalized? | 17:19.55 |
rayjj | Tue pm, Wed pm at the ski area, then ski Thur and drive back to airport, right ? | 17:19.56 |
mvrhel_laptop | yes | 17:20.01 |
Robin_Watts | In order for you to be able to log in, I need an ssh key. | 17:20.02 |
kens | mvrhel_laptop : and late morning for me :-) | 17:20.16 |
fredross-perry | both âRossâ and âPerryâ are caps. like this Fred Ross-Perry | 17:20.26 |
Robin_Watts | fredross-perry: Are you a unixer or a windowser? | 17:20.30 |
kens | A Mac I suspect | 17:21.06 |
| From the weird characters | 17:21.10 |
rayjj | Robin_Watts: which am I ??? I *hate* windows, but it's OK with a real shell | 17:21.15 |
mvrhel_laptop | best to get lift tickets on line i see | 17:21.21 |
kens | Yes, cheaper that way | 17:21.28 |
rayjj | mvrhel_laptop: yep. | 17:21.32 |
fredross-perry | Mac OS X all the way. | 17:21.46 |
mvrhel_laptop | 3 days in march is $227 | 17:21.57 |
| no the end of the world. same as my season pass here though | 17:22.18 |
kens | Really ? I saw it as 3 x $80 | 17:22.20 |
Robin_Watts | fredross-perry: ok. So I don't know how you generate ssh keys on macos. | 17:22.21 |
rayjj | mvrhel_laptop: are you in for the 3 days ? Just two is OK with me as well. | 17:22.39 |
fredross-perry | nor do I. But I can find out. | 17:22.42 |
kens | 6 day pass in Austria was less than a 1 day pass in Vail...... | 17:22.42 |
| Admittedly the Euro is weak atm | 17:23.03 |
Robin_Watts | https://help.github.com/articles/generating-ssh-keys/#platform-mac | 17:23.07 |
mvrhel_laptop | kens: tuesday, wed, and thursday for a total of $227 | 17:23.07 |
| walk up single day rate is $128 | 17:23.18 |
kens | NP | 17:23.18 |
chrisl | Robin_Watts, fredross-perry: I suspect the key generation will be the same as Unix | 17:23.21 |
kens | mvrhel_laptop : I'll book pass in advance when I have flights and stuff sorted | 17:23.45 |
fredross-perry | what is casper, and how/why does one access it? | 17:23.52 |
mvrhel_laptop | rayjj: if I am going to leave my wife hanging with the kids while I am skiing, I am going to ski | 17:23.58 |
| all three days | 17:24.03 |
kens | :-D | 17:24.07 |
Robin_Watts | yeah, probably. fredross-perry, if you generate a key, and then email the public part to me, I'll get you set up. | 17:24.17 |
kens | Melanie was not hugely impressed when I said I was hoping to ski at Denver in March as well as Austria :-) | 17:24.30 |
fredross-perry | ok then. | 17:24.33 |
Robin_Watts | fredross-perry: casper is the server that runs ghostscript.com, git.ghostscript.com etc. | 17:24.33 |
fredross-perry | thanks | 17:24.50 |
mvrhel_laptop | now, I would like to bring my board with me. question is, will if fit in the rental vehicle.... | 17:24.52 |
henrys | Robin_Watts: let me know when you have an account for him I'll put his gmail password in his casper directory. | 17:25.01 |
mvrhel_laptop | can I charge miles for the checked bag ;) | 17:25.09 |
kens | Well I'm bringing my boots, I'd have thought your board would fit | 17:25.11 |
Robin_Watts | Once you can ssh into there, we can set you up with access to the official git repos etc. | 17:25.20 |
| henrys: /home/fred exists. | 17:25.27 |
kens | Unless you have a *really* big board | 17:25.27 |
mvrhel_laptop | no | 17:25.40 |
kens | Should have thought it would fit in the trunk of a US car :-) | 17:26.04 |
| I'm going to have to go cook, I'll be back later. | 17:26.28 |
mvrhel_laptop | ok, so the question is finding a place to stay | 17:26.30 |
| ok kens | 17:26.33 |
| ttyl | 17:26.34 |
fredross-perry | I assume I should use my artifex email address when creating a new key? | 17:26.39 |
mvrhel_laptop | but, first let me check with miles | 17:26.47 |
| and we have not heard from tor8 | 17:26.52 |
Robin_Watts | fredross-perry: Not sure that matters. | 17:27.04 |
fredross-perry | ok | 17:27.11 |
chrisl | Robin_Watts: did you set the groups for fred on casper? | 17:29.31 |
henrys | fredross-perry: your email address and password are in your home directory on casper. | 17:29.47 |
mvrhel_laptop | rayjj and kens, just fired off an email to miles to see if he was interested | 17:30.21 |
henrys | fredross-perry: I made it only visible to you but you should change the password when you have a chance. | 17:30.39 |
Robin_Watts | chrisl: Not yet. | 17:30.41 |
mvrhel_laptop | once we hear from him and tor8 we will figure out the lodging and car | 17:30.45 |
henrys | fredross-perry: go to gmail.com and login ... then you can forward the email or whatever you want to do. | 17:31.21 |
chrisl | Robin_Watts: okay, the groups are set | 17:32.09 |
Robin_Watts | chrisl: Ta. | 17:32.15 |
rayjj | mvrhel_laptop: thanks. so we are waiting to hear from him and tor? (and did henrys say no?) | 17:32.52 |
chrisl | I wonder if there is any other havoc I should wreak while I'm logged into casper...... | 17:33.04 |
mvrhel_laptop | henrys: did not say no or yes | 17:33.09 |
| henrys? | 17:33.15 |
| should be a lot of fun | 17:33.22 |
rayjj | I have to run an errand. bbiab | 17:33.31 |
henrys | mvrhel_laptop: no I'm not going thanks though. | 17:34.08 |
mvrhel_laptop | ok | 17:34.18 |
henrys | bbiab | 17:36.05 |
mvrhel_laptop | ok. until we hear from miles and tor8, I will get back to work now on gsview printing. should have this wrapped up shortly for a beta release | 17:50.46 |
fredross-perry | mvrhel_laptop: isthat checked in so I can test-drive the UI? | 17:57.20 |
mvrhel_laptop | fredross-perry: not yet. let me get it in today. it will not all be working today though | 18:00.08 |
fredross-perry | thanks, let me know | 18:01.44 |
| for example, youâve got âsubsetâ under âprint rangeâ, but I donât know what it does. | 18:02.43 |
Robin_Watts | Ok, fred, you should now be able to ssh to fred@ghostscript.com: using that ssh key. | 18:06.25 |
fredross-perry | Robin_Watts: please email me any specifics of the casper account so I can try it. Thanks. | 18:08.41 |
| hahaha. ignore last message | 18:08.54 |
| I get âPermission denied (publickey)â | 18:10.13 |
Robin_Watts | fredross-perry: I can't see why offhand. | 18:12.17 |
rayjj | kens: There is a suspicious line of code in pdf_main.ps: /Orientation 0 def This affects having the FitPage option allow for 90 and 270 degree. It's part of a massive commit af4b67d2 | 18:12.21 |
| kens: The part of the comment that seems related is (F) /Orientation now always 0 which mentions making pdfmark work. What to do you think about making this ONLY for pdfwrite ? (Other devices don't use pdfmark) | 18:12.22 |
Robin_Watts | My key is a ssh-dss one, and yours is a an ssh-rsa one. | 18:12.31 |
| And mine looks longer than yours (oh, err) | 18:12.39 |
fredross-perry | I might have to direct ssh to the correct key; I have more than one | 18:12.47 |
Robin_Watts | fredross-perry: Your key looks akin to chrisl's though. | 18:13.18 |
rayjj | Robin_Watts: recall that linux SSH is picky about .ssh directory and file permissions in it | 18:13.21 |
Robin_Watts | rayjj: aye, that's all sorted, I hope. | 18:13.30 |
| fredross-perry: As long as ssh has been told about yours it should be OK. On macosx there is a keyring thing, maybe. | 18:15.31 |
| Try ssh -i <pathtokey> fred@ghostscript.com: | 18:16.22 |
fredross-perry | thanks, thatâs it. | 18:17.06 |
Robin_Watts | OK, let me do some git magic. | 18:17.58 |
rayjj | kens: BTW, that doesn't show any pdfwrite regressions | 18:18.17 |
Robin_Watts | fredross-perry: Anything you put in public_html will appear on http://ghostscript.com/~fred/ | 18:18.42 |
fredross-perry | sweet | 18:18.51 |
rayjj | chrisl: as the other PDF interp maintainer, can you look at the above messages to kens about Orientation 0 def in pdf_main.ps | 18:19.19 |
fredross-perry | The friendliest ghost, you know. | 18:19.53 |
Robin_Watts | fredross-perry: OK, in 'repos' in your home dir there are ghostpdl.git and mupdf.git bare git repos. | 18:22.53 |
| Those appear on git.ghostscript.com | 18:23.01 |
fredross-perry | nice | 18:23.09 |
Robin_Watts | Any other repos you create in that dir will automatically appear there too. | 18:23.21 |
chrisl | rayjj: Perhaps we should add a .special_op case for pdfmark compatibility - unless there's already one | 18:23.59 |
Robin_Watts | Generally the way we work is that we push any changes into our repos on casper. Then we get a fellow developer to review the commits before we push them to the main repo. | 18:24.06 |
| (For MuPDF we have a policy of requiring the review stage. For GhostPDL we recommend it, but it's not strictly required) | 18:24.34 |
fredross-perry | gotcha | 18:24.43 |
Robin_Watts | Feel free to force push etc into your own repos as much as you like. | 18:24.57 |
| basically, they are yours to do what you want with. | 18:25.07 |
| We try to avoid ever force pushing into the golden repo though. | 18:25.17 |
| (cos it upsets other people and the cluster) | 18:25.30 |
fredross-perry | ok | 18:25.52 |
Robin_Watts | Ah, the cluster, I should talk about that... | 18:26.24 |
| just a mo... | 18:26.25 |
| fredross-perry: OK, go here: http://ghostscript.com/regression | 18:30.29 |
| it'll ask you for a username/password | 18:30.42 |
tor8 | kens: mvrhel_laptop: rayjj: sorry, stepped out for a while and missed the discussion. | 18:31.03 |
mvrhel_laptop | np | 18:31.12 |
Robin_Watts | I've just given you them in a private tab. | 18:31.16 |
mvrhel_laptop | miles is not going | 18:31.19 |
| so now it is ray, ken, me and tor8? | 18:31.29 |
tor8 | skiing tue afternoon, all wed, and thursday morning? | 18:31.53 |
fredross-perry | âprivate tab?â | 18:32.21 |
mvrhel_laptop | at min. I am thinking all day tuesday, all day wed. and all day thursday ;) | 18:32.21 |
Robin_Watts | fredross-perry: irc private message. | 18:32.33 |
mvrhel_laptop | lifts are open 9 to 4 | 18:32.35 |
tor8 | mvrhel_laptop: I'd collapse of exhaustion before I made it to thursday :) | 18:32.40 |
mvrhel_laptop | oh you would be fine | 18:32.46 |
chrisl | rayjj: actually, we already have a conditional in there for whether we're handling pdfmarks, so we can just use that - ".writepdfmarks" | 18:33.03 |
mvrhel_laptop | ray said he might just do 2 days | 18:33.08 |
tor8 | mvrhel_laptop: I'm in, I can come before the meeting as well | 18:33.16 |
mvrhel_laptop | ok cool! | 18:33.21 |
| I will start looking for a place | 18:33.29 |
chrisl | tor8: you already practically sleep through the staff meetings..... | 18:33.30 |
mvrhel_laptop | hehe | 18:33.35 |
tor8 | I think 2 days would be enough for me, but if you and ken want to do 3 days then I'll cope somehow :) | 18:33.42 |
Robin_Watts | fredross-perry: let me know if that gets you in. | 18:33.59 |
mvrhel_laptop | ken seems a bit hard core. I have to try to keep up with him | 18:34.02 |
fredross-perry | yes, iâm in | 18:34.14 |
tor8 | and I guess it'll be easiest for me to make it onto kens flight if possible, depends on how early kens wants to depart from hearthrow | 18:34.17 |
| I need to find a connecting flight that makes it in time | 18:34.24 |
Robin_Watts | fredross-perry: OK, so a quick rundown on the cluster... This is the cluster dashboard. | 18:34.32 |
fredross-perry | yes | 18:34.45 |
tor8 | mvrhel_laptop: I haven't been on a pair of skis in almost 10 years now... | 18:34.52 |
mvrhel_laptop | right. I think we will stay that monday night at an airport hotel and then leave early in the morning for the resort | 18:34.54 |
Robin_Watts | On the left hand side is a list of the cluster machines, and what they are currently doing. | 18:34.58 |
mvrhel_laptop | tor8: its like riding a bike... | 18:35.05 |
Robin_Watts | All idle at the moment. | 18:35.10 |
mvrhel_laptop | plus your swedish | 18:35.15 |
tor8 | mvrhel_laptop: and the time before that was 10 years earlier as well. like riding a bike :) | 18:35.19 |
mvrhel_laptop | you were born with them on | 18:35.22 |
Robin_Watts | At the top is the current/most recent job - so ray was the last one to run something. | 18:35.36 |
| at the bottom there is a list of jobs. | 18:35.47 |
| red jobs have finished. | 18:35.51 |
tor8 | mvrhel_laptop: no mountains in sweden... I didn't learn skiing until I spent time in switzerland | 18:35.56 |
Robin_Watts | amber jobs are waiting to run, green jobs are running now (I like traffic lights) | 18:36.07 |
mvrhel_laptop | nothing at all there? not even x-country? | 18:36.14 |
Robin_Watts | At the moment, no queued jobs. | 18:36.21 |
fredross-perry | yes | 18:36.32 |
Robin_Watts | Top right are some 'useful' links. | 18:36.48 |
tor8 | mosty people here do cross country, but we don't get a lot of snow in southern sweden | 18:36.51 |
Robin_Watts | bottom right are some user links. Your name will appear there at some point. | 18:37.13 |
tor8 | the mountains are all up north bordering on norway, and they don't compare favorably with the alps | 18:37.21 |
mvrhel_laptop | ok. I see. | 18:37.45 |
| well this should be a good time. I will see what I can find lodging wise | 18:37.53 |
Robin_Watts | fredross-perry: Whenever anyone commits a change to mupdf or gs, a job is queued (or possibly several jobs) that run several thousand files through the code to see if anything breaks. | 18:38.22 |
| we collect the md5 sum outputs and see if they differ from before. | 18:38.36 |
fredross-perry | nice system. Whatâs the protocol for creating a new etst and adding it to the suite? | 18:39.14 |
Robin_Watts | The cluster then sends a report out to our emails. | 18:39.19 |
fredross-perry | âtest" | 18:39.20 |
kens | regarding pdfmark, check for IsDistiller, not the device name | 18:40.00 |
Robin_Watts | fredross-perry: We have an svn repo of test files; the cluster runs selections from that. | 18:40.05 |
| If you have a new file you want tested, you just check it into the svn repo in an appropriate place. | 18:40.32 |
chrisl | kens: see above - I think it would be better to use ".writepdfmarks" | 18:40.39 |
kens | THat works too | 18:40.48 |
fredross-perry | so, like, ârender this, the result should look like thatâ? | 18:40.51 |
kens | There is a -dNOPDFMARKS and -dDOPDFMARKS so it makes sense | 18:41.01 |
Robin_Watts | fredross-perry: We never tell the cluster what the result should look like. | 18:41.10 |
kens | I'm trying to catch up in the logs...... | 18:41.14 |
Robin_Watts | The first time it runs a file it remembers the md5 sum of the output, and it checks for differences each time. | 18:41.27 |
| Now, as well as jobs running whenever things are committed, users can submit jobs. | 18:42.04 |
| So if I have a code change, I can run it through the cluster before I commit it. | 18:42.22 |
| (as otherwise we'd only ever find out we'd broken something after we broke it). | 18:42.36 |
fredross-perry | So simply a regression, making sure nothing has changed. What do you do when something *should* change? remove the file and re-introduce it? | 18:43.14 |
Robin_Watts | These user jobs get queued in the same list as the commit jobs, but they get run at a higher priority. | 18:43.20 |
| fredross-perry: Exactly. It's just regressions. | 18:43.48 |
kens | tor8 ping | 18:43.54 |
Robin_Watts | If a file *should* change, then just commit the code change. You'll be told that the file has changed in the report. | 18:44.08 |
| Then it'll remember that it's changed and only report changes from the new version. | 18:44.27 |
fredross-perry | cool. | 18:44.36 |
Robin_Watts | hence it's important that we actually watch for changes :) | 18:44.42 |
rayjj | chrisl: kens: actually the entire reason for setting Orientation to 0 is suspect. There is no specific discussion in the log or source other than "it is a long story" | 18:44.53 |
fredross-perry | Does Artifex actually own the machine farm? Or som service? | 18:45.04 |
Robin_Watts | Now, suppose you run a user test, and you get told that 1000 files changed. | 18:45.05 |
| We own the machine farm. | 18:45.17 |
chrisl | fredross-perry: it's conventional to note expected cluster changes in your commit message | 18:45.17 |
rayjj | fredross-perry: it's machines in various places | 18:45.29 |
chrisl | rayjj: I noted that useful part of the commit message..... | 18:45.49 |
Robin_Watts | Most of it is in marcosw's garage, but some of the servers are in other places (like ray's office, miles' office etc) | 18:45.51 |
kens | OK I Guess tor8 is away. For the logs, tor8 my flight looks like it departs Heathrow at 12;20, arriving at 15:10 local time. That's a direct flight with BA on BA0219. Not booked it yet though. | 18:45.59 |
rayjj | fredross-perry: many are in either the HQ (Miles' office) or marcosw' house. I have peeves and peeved | 18:46.04 |
kens | rayjj I'll look at the commit, was it one of mine ? | 18:46.14 |
chrisl | It was Alex's | 18:46.37 |
Robin_Watts | OK, so suppose you run a test and you're told that 1000 files changed. | 18:46.41 |
kens | Oh great :-( | 18:46.44 |
rayjj | kens: no, it was one that Alex commited -- looks like the work was actually from SaGS, however | 18:46.47 |
Robin_Watts | It's often nice to be able to *see* what changed. | 18:47.01 |
fredross-perry | well, my first act will be to move everything Iâve got form GitHub. | 18:47.07 |
tor8 | kens: okay, I'll look see what options I get | 18:47.09 |
Robin_Watts | So we can run a different type of user job, called 'bmpcmp'. | 18:47.20 |
kens | Aha, OK tor8, as before if it helps to fly into Gatwick and travel round with me the next day, we can do you a bed for hte night | 18:47.36 |
rayjj | kens: so that is coming in on Monday ? | 18:47.37 |
Robin_Watts | That runs the files that changed on your most recent test through your new version of the code, and through the base version of the code, and generates a helpful page full of differences. | 18:47.57 |
kens | rayjj My only options are afternoons, so to ski Tuesday I need to fly in Monday yes | 18:47.58 |
| As I said to tor8, not booked anything yet | 18:48.13 |
tor8 | kens: march 2nd through march 8th? | 18:48.15 |
rayjj | kens: tor8: and mvrhel_laptop: Arapahoe Basin is also good. It's more "basic", but also cheaper and I had fun there | 18:48.27 |
Robin_Watts | For example, click on 'bmpcmp' next to 'chrisl' and you'll see an example. | 18:48.35 |
kens | tor8 yes, fly out Monday 2nd Back Sunday 8th | 18:48.47 |
| rayjj Like I said I don't mind which resort, it'll be fun for me to go somewhere different | 18:49.09 |
Robin_Watts | On those pages, the diffs are new/old/diff | 18:49.26 |
kens | tor8 return flight (BA 0218) departs Denver 18:40, arrives Monday 09:30 | 18:49.43 |
| COmes in a t £666 :-D | 18:50.11 |
tor8 | kens: first option I get from BA is on that flight (BA219 at 12:20) with a 3 hour layover in heathrow | 18:50.16 |
kens | Is that OK ? | 18:50.26 |
tor8 | kens: yes, it's as good as they get | 18:50.37 |
kens | Fair enough then :-) | 18:50.42 |
Robin_Watts | fredross-perry: AIUI all your current stuff is independent of the main ghostscript core, right? | 18:50.55 |
mvrhel_laptop | kens, rayjj, and tor8: there are a ton of lodging available. Looks like most run 400 or so a night. That means we would each only pay $200 | 18:50.58 |
rayjj | I've skiied all 4: Breckenridge, Keystone, Copperr and A-basin. I liked A-basin because it was less crowded, but we are mid week, so it should be OK anywhere. Keystone has some good intermediate runs that are l..o..n..g.. | 18:50.59 |
mvrhel_laptop | rayjj: we need to zero in on one place | 18:51.11 |
Robin_Watts | so it won't be exercised by the cluster tests. | 18:51.13 |
fredross-perry | Robin_Watts: yes, itâs all independent. | 18:51.33 |
mvrhel_laptop | if you want to pick it and find a place I am fine with that | 18:51.37 |
kens | mvrhel_laptop : That's a much nicer rate than staying at Vail in a hotel :-) | 18:51.40 |
rayjj | mvrhel_laptop: we will have a car -- we can ski any of them (or at least 3 of the 4) | 18:51.50 |
chrisl | rayjj: I guess the Orientation thing might cause problems with things like setting the CropBox, and other pdfmarks that take coordinates..... | 18:51.58 |
tor8 | kens: all the other options route me through dallas or a 2nd layover so I doubt I can get better than this | 18:51.58 |
Robin_Watts | fredross-perry: OK, a question worth asking... should we have your stuff in a separate git repo? | 18:52.01 |
mvrhel_laptop | rayjj: I would prefer to stick to one place | 18:52.07 |
Robin_Watts | We could have a gsview.git with all your stuff in. | 18:52.14 |
kens | tor8 yes I see many going via Dallas, I'[m avoiding all those..... | 18:52.15 |
rayjj | mvrhel_laptop: are you looking at 3 or 4 br ? | 18:52.17 |
Robin_Watts | And that could have ghostpdl.git and mupdf.git as submodules of it. | 18:52.25 |
| That seems like a nice solution to me. | 18:52.32 |
mvrhel_laptop | I was looking at 2 br sleeps 8 to 10 kind of thing. each room as 2 beds | 18:52.40 |
tor8 | kens: yeah, I don't want to do that! | 18:52.45 |
mvrhel_laptop | s/as/has/ | 18:52.46 |
rayjj | mvrhel_laptop: sticking at one place is fine with me. Just curious, why ? | 18:53.02 |
kens | 2 br is fine by me, I don't *think* I snore..... | 18:53.02 |
mvrhel_laptop | rayjj: I would rather ski than drive | 18:53.14 |
Robin_Watts | cos it means that a given SHA of gsview.git would capture the exact versions of ghostpdl and mupdf that you used too. | 18:53.21 |
mvrhel_laptop | Getting up and walking to lift and skiing back to room is nice | 18:53.35 |
| not to mention better deals on lift tickets for multi-day | 18:54.06 |
kens | Yes, in Austria we had to get the bus in the morning, but could ski back to the hotel. | 18:54.07 |
rayjj | mvrhel_laptop: I usually stay a drive away anyway, so a few extra minutes to drive further gets variety | 18:54.21 |
kens | mvrhel_laptop : I htink 3+ day passes can be used in multiple resorts | 18:54.21 |
fredross-perry | All my stuff happens in mupdf/platform/Qt | 18:54.42 |
mvrhel_laptop | ok. if you guys want to do a multi-resort type of trip I can adapt | 18:54.50 |
rayjj | kens: those tend to be pricier (but not that much) | 18:54.51 |
| I haven't shopped the 3-day packages. Copper 1 day was $78, A-basin was $56 | 18:55.40 |
Robin_Watts | fredross-perry: OK. Fair enough. | 18:55.45 |
| I thought maybe you had both ghostpdl and mupdf dependencies. | 18:56.04 |
mvrhel_laptop | rayjj: you are right, it is not that big of a savings | 18:56.27 |
kens | rayjj chrisl without a lot of head-scratching and investigation I can't be sure why Alex fixed the Orientation at 0. I suspect its because the Orientation gets 'baked in' to the media. Whereas the annotation form the pdfmark is independent of the /Orientation of the PDF page. SO we need to take account of any CTM change when generating hte annotation from the pdfmark. | 18:56.29 |
fredross-perry | My build process however involves building mupdf as static libs and linking them. Then, I build gs and gxps from ghostpdl and *copy* those executables. | 18:56.32 |
tor8 | kens: hm, economy plus is only 80% extra... a shame about the currency rates | 18:56.35 |
kens | Well, I'm not paying Euros :-) | 18:56.53 |
tor8 | the swedish crown is in free fall... dollar costs 26% more than 6 months ago | 18:57.17 |
kens | Wow ! | 18:57.24 |
tor8 | good for my pay! bad for everything else... | 18:57.30 |
kens | chrisl rayjj making the Orientation 0 for pdfmarks seems the simplest solution. | 18:57.45 |
tor8 | last summer it was down at 6.4 crowns per dollar, now it's over 8 | 18:57.51 |
Robin_Watts | fredross-perry: Right. That's why I was wondering about a separate repo. That way you could have a build system that called the submodules to do the building/copying, so it'd all be a one shot thing. | 18:58.14 |
kens | I thought Sweden was in the Euro Zone, but I think that may be even worse | 18:58.23 |
Robin_Watts | Also, are we decided that freds stuff is all going to be open source? | 18:59.05 |
tor8 | kens: well, the sek/euro ratio is relatively unchanged so I guess it's a general thing | 18:59.43 |
| in the euro zone | 18:59.54 |
kens | I guess, we got a good exchange rate for sterling this time. | 19:00.30 |
fredross-perry | One-shot. Is that a requirement, or simply a laudable goal? Right now I am writing up some build instructions that have the user using Qtâs IDE for some steps. | 19:01.40 |
Robin_Watts | a laudable goal. | 19:02.22 |
fredross-perry | thanks | 19:02.31 |
kens | OK I'm going to run off again. AIU the plan (for me and Tor) is to fly out Monday, stay in the airport Monday night. Travel to <insert ski resort here> early Tuesday, ski Tuesday, Wednesday and Thursday, then drive back to the airport for the meeting. I'll book flights and stuff tomorrow unless I hear differently from Michael/Ray. I'l try to be back later again to read hte logs and catch up | 19:02.53 |
Robin_Watts | For android we have 2 build commands. I could have written a shell script, but :) | 19:03.13 |
tor8 | fredross-perry: Robin_Watts: given that way of building, I'd suggest a separate repo with ghostpdl and mupdf as submodules would be the smoothest way to work | 19:03.28 |
| kens: ping me when you decide to book your flight, and I'll do the same | 19:04.10 |
| and then we'll need to shop for an airport hotel I guess | 19:04.34 |
Robin_Watts | certainly sounds best to me. | 19:05.50 |
fredross-perry | thanks all. | 19:06.02 |
Robin_Watts | fredross-perry: At some point in the future, we're going to want to look back and build a given version of your app. | 19:06.16 |
| and at that point, we'll really wish we had it tied together so that we know what exact version of gs needs to be built, and what exact version of mupdf is required. | 19:06.53 |
| a separate repo and git submodules achieves that really nicely for us. | 19:07.15 |
fredross-perry | how do submodules acheive this? | 19:07.39 |
Robin_Watts | fredross-perry: Take mupdf as an example. | 19:09.35 |
| It relies on libjpeg and freetype etc. | 19:09.48 |
chrisl | fredross-perry: you put a tag in your repo, it applies to the submodule versions in use at the time | 19:09.53 |
Robin_Watts | We have git submodules for each of those. | 19:10.05 |
| as far as git is concerned, a given 'commit' contains the exact SHA for each of those submodules too. | 19:10.40 |
| So when you checkout to mupdf-version-1 (or whatever), you can also pull the submodules along to exactly the state they should be in. | 19:11.14 |
| Then doing a build gets us exactly the right versions of the libs etc. | 19:11.37 |
| If a lib is updated (say we get a new freetype release), then we 1) update the freetype git module with the changes from upstream, and 2) do a commit in the main mupdf repo that brings the freetype submodule up to the new version. | 19:13.02 |
mvrhel_laptop | http://www.vrbo.com/493433 | 19:16.21 |
Robin_Watts | kens, chrisl: Random thing... I've just gone onto contract with three, for a tenner a month, (2GB data, 200minutes, unlimited texts), 12 month contract. | 19:19.07 |
| The attraction of that is that when I go to the states, I get to use the phone there with the same costs/allowances as here. | 19:19.36 |
chrisl | Interesting..... I wonder if Virgin will follow that lead at some point | 19:20.35 |
Robin_Watts | (It's not just the states, it's 16 countries worldwide, raising to 18 soon) | 19:22.00 |
chrisl | It's just, I do like my unlimited everything contract | 19:23.09 |
Robin_Watts | chrisl: For how much a month? | 19:23.19 |
chrisl | 15ukp | 19:23.27 |
Robin_Watts | yeah, that'd be 25 quid a month on three. | 19:24.15 |
chrisl | I get it a little cheaper because I have cable/tv/land line with virgin | 19:24.52 |
Robin_Watts | all you can eat data/texts, and 200 minutes would be 15 quid a month on three. | 19:24.52 |
chrisl | I'll look at three when my year is up with virgin | 19:25.42 |
| Being able to use data in the US is attractive | 19:28.27 |
mvrhel_laptop | ok so we have a van for denver now | 19:31.59 |
chrisl | henrys: We *are* up to date with lcms, that commit that I pulled over isn't in an official release yet | 19:32.13 |
mvrhel_laptop | lunch bbiab | 19:32.17 |
chrisl | I think I'm done for the day - 'night all | 19:32.52 |
iw2evk | hi | 19:37.56 |
ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 19:37.56 |
iw2evk | someone have a makefile for compile gs 9.15 with Djgpp (gcc for dos)? thanks | 19:38.57 |
Robin_Watts | iw2evk: The guy you'd need to speak to has just gone for the night. Can you come back in about 16 hours time? | 19:39.36 |
iw2evk | unfortunly i'm at my wokplace and can't access to chat | 19:40.15 |
chrisl_away | iw2evk: no, we haven't supported any dos builds for quite some time. I don't know enough about djgpp to judge whether it would be difficult to make it work | 19:40.30 |
iw2evk | last build for dos is 8.55 | 19:41.14 |
Robin_Watts | djgpp *was* just a gcc for DOS, but god knows what the restrictions on that are these days. | 19:41.19 |
iw2evk | i've tried to compile using sh ./configure -CC=gcc AWK=gawk , compilation start , goes to the end then hang up.. | 19:42.35 |
| i suppose need a patch for djgpp , but i don't find this patch on net :(( | 19:43.01 |
chrisl_away | 8.55 was a *long* time ago | 19:43.24 |
| iw2evk: you might want to try using the --disable-compile-inits configure option - compiling the resource files in results in a *huge* object object file for the linker to deal with | 19:45.01 |
iw2evk | someone have ported to dos mupdf but don't have print support .. gs it's the only one tools for manipolate pdf in dos.. | 19:45.30 |
| whath kind of graphic libraries use gs ? i 've GRX libraries enable.. | 19:46.46 |
Robin_Watts | iw2evk: mupdf does have print support, for certain printers. | 19:47.12 |
| it's not as extensive as gs, certainly. | 19:47.23 |
iw2evk | i don't view the command for ptinting invoking f1 .. | 19:47.44 |
Robin_Watts | what kind of graphics libraries use gs, or what kind of graphics libraries does gs use? :) | 19:47.49 |
iw2evk | use actually.. | 19:48.01 |
Robin_Watts | iw2evk: The core of MuPDF supports output to various different bitmap formats. And we have code to take those bitmap formats to printer streams for certain printers. | 19:48.28 |
chrisl_t530 | gs doesn't use any external graphics libs (unless you count x11 for the x11 devices) | 19:49.08 |
iw2evk | this is the mupdf porting :https://nanox-microwindows-nxlib-fltk-for-dos.googlecode.com/files/mupdf.zip | 19:49.20 |
Robin_Watts | Just because the port you are using hasn't hooked up the printing stuff, doesn't mean it's not there. | 19:49.30 |
| iw2evk: Which particular printer do you want to drive? | 19:50.08 |
iw2evk | hp deskjet 610 c on lpt1 | 19:50.29 |
| they are latest gs porting for dos : http://ftp.ibiblio.org/pub/micro/pc-stuff/freedos/files/util/file/gs/ | 19:50.44 |
chrisl_t530 | dos isn't really on our radar..... | 19:51.47 |
iw2evk | but i thinks someone on gs devels can wrote some rows of code for help dos users.. | 19:52.43 |
chrisl_t530 | Yes, long time ago | 19:53.19 |
Robin_Watts | iw2evk: MuPDF has code for various HPs. | 19:53.41 |
| Dunno about the 610c though. | 19:53.47 |
| see include/mupdf/fitz/output-pcl.h | 19:54.08 |
iw2evk | if exist deskjet only (like in gs) it's enought.. | 19:54.32 |
| ok guys , it' s a pleasure listen your suggestions, but i due shuthdown..i hope tho meet again on this channel.. | 19:57.15 |
| i tr | 19:57.34 |
| Ciao and good night! | 19:57.42 |
Robin_Watts | the code from mupdf was lifted and modified from gs. | 20:00.00 |
kens | For the logs, Michael that place on vrbo looks fine to me. I'll book flights tomorrow then look for a Holiday Inn or something at the airport. | 20:07.11 |
| Night all | 20:07.14 |
| rayjj fwiw apparently Melanie and I skied 37 km on the Tuesday, 5,571 vertical metres, 18 lifts | 20:08.06 |
rayjj | darn. My git seems to still be stuck at commit f02d0a5 (which doesn't build on Windoze due to variables not at the start of the block) | 21:40.24 |
| but that's the hash I see on http://git.ghostscript.com/?p=ghostpdl.git;a=summary | 21:41.14 |
| which I thought was supposed to be the "fixed" one | 21:41.44 |
| Robin_Watts: (if you are still really here) Does mvrhel's "git push -f origin" mean I have to do something different to: git up ??? | 21:43.25 |
| hmm... *NOT* a good time to be having discussions on IRC (during maintenance) | 21:44.00 |
| :-( | 21:44.06 |
fredross-perry | mvrhel_laptop: I sent you a couple of email queries re: your print dialog design. | 22:59.57 |
| Forward 1 day (to 2015/01/14)>>> | |