| <<<Back 1 day (to 2013/11/17) | 2013/11/18 |
henrys | quiet today | 13:33.25 |
kens | All busy :) | 13:33.34 |
sebras | kens: _and_ I'm not annoying any of you. | 14:08.45 |
kens | as if you would... | 14:11.27 |
punknroll | hi all, is there already a release date for 9.11? | 14:19.17 |
chrisl | punknroll: feb or march next year | 14:26.06 |
punknroll | ah ok thanx... | 14:39.59 |
Micha` | It'll always be too soon for 9.11 related things. | 14:40.20 |
| \o/ | 14:40.22 |
punknroll | i need to compile the current master and i wonder if i should compile in in a separate directory or system wide...I think the first approach is better...if something goes wrong I can use the system wide | 14:41.36 |
Robin_Watts | punknroll: We specifically recommend you avoid sharing system libraries. | 14:42.09 |
punknroll | i don't know exactely what that means...i compiled it locally with the option --prefix in a custom directory...that worked well | 14:45.41 |
Robin_Watts | punknroll: OK, that's just building it, so that installing it will go to the specified directory, AIUI. | 14:46.11 |
punknroll | yes this is what i meant...after make install i have two versions of ghostscript on my machine right? Or let me ask more detailed: is the master branch stable enough for production or should i wait for 9.11 and install a second version in another directory? | 14:49.36 |
Robin_Watts | punknroll: The master version is stable enough for production, I believe. | 14:50.16 |
punknroll | ok thanx | 14:53.04 |
| by they way...thanx kens for the segfault bugfix!!! | 14:53.30 |
kens | Hmm, I *hope* its a fix, I'm not totally confident.... | 14:53.55 |
| tkamppeter, I believe we did remove the psgray device when we removed pswrite, yes | 15:00.37 |
tkamppeter | kens, thanks, I will tell it to the original reporter. | 15:03.13 |
kens | NP | 15:03.18 |
ray_laptop | morning, all | 15:38.30 |
kens | Hi Ray | 15:38.50 |
ray_laptop | Robin_Watts: I saw your reply to cust 801 | 15:39.00 |
Robin_Watts | ray_laptop: Ah, yes, I meant to talk to you about that. | 15:39.30 |
| should they be able to do ./autogen.sh ? | 15:39.42 |
ray_laptop | Robin_Watts: I looked at the .raw output and the values were different on the black plane | 15:39.49 |
Robin_Watts | .raw output ? | 15:40.09 |
| I've said many times, that I've only tested the pgm output. | 15:40.46 |
ray_laptop | Robin_Watts: yes, if I packaged up enough stuff. Since I was working (and thinking) windows, I hadn't done a linux build, so the first .zip may not have had the configure.ac and Makefile.in needed for autogen.sh | 15:41.12 |
Robin_Watts | ray_laptop: Right. | 15:41.22 |
ray_laptop | Robin_Watts: I looked at a hex dump of the .raw files of tiger. The numbers are "similar" in range, but are not quite identical sequences, as if the BNM is being applied differently | 15:42.39 |
Robin_Watts | .raw? | 15:43.28 |
ray_laptop | Since the original .zip, the patches I sent you and mvrhel (and cc'ed tech) make the corrections/additions to make something that also builds and runs on linux | 15:43.42 |
| Robin_Watts: initially I was going to compare the .pgm's but I have to "fix" yours or theirs so they are comparably (in range, etc.) e.g. theirs has maxval == 15, yours is 255 | 15:45.08 |
Robin_Watts | ok. | 15:45.36 |
ray_laptop | and your .pgms are inverted compared to theirs (at least when viewed with XNView | 15:45.37 |
Robin_Watts | (255-x)>>4 for each byte then. | 15:45.54 |
| Oh, no, it's worse than that of course. | 15:46.11 |
| Oh! I might know what it is. | 15:47.10 |
| You're running at 600dpi, and it's the black plane that differs ? | 15:47.24 |
| Give me 10 mins to park what I am doing and check. | 15:49.40 |
ray_laptop | Robin_Watts: I have to take my youngest to school now, bbiab (25 min or so) | 15:59.28 |
Robin_Watts | I've found a problem in the CMY plane dithering stuff, but not in the K. | 16:29.02 |
| hi ray_laptop | 16:47.05 |
| I've found a problem in the CMY planes I think, but not in the K plane. | 16:47.17 |
ray_laptop | Hi, Robin_Watts: soory I took so long | 16:47.20 |
Robin_Watts | no worries. | 16:47.25 |
ray_laptop | Robin_Watts: I'm not _sure_ that it's a problem, just a *difference* | 16:47.51 |
Robin_Watts | this is a problem in that I'm running through the noise tables twice as fast as I should. | 16:48.17 |
| So only the even lines are used, and they repeat every 64 rather than every 128. | 16:48.36 |
ray_laptop | the numbers look *similar* (in the .raw data). Just the sequence isn't identical | 16:48.47 |
| Robin_Watts: right, I can see how that could happen with the downscaled data | 16:49.09 |
Robin_Watts | What resolution are you using? And what planes were you seeing differences in? | 16:49.20 |
ray_laptop | although I didn't spot it in the source :-( | 16:49.23 |
| I was using -r600 | 16:49.37 |
| and I was looking at the start of the data, which I am pretty sure is the K plane | 16:49.57 |
Robin_Watts | @600dpi I would expect differences in C/M/Y but not K. | 16:50.08 |
ray_laptop | Robin_Watts: I see (at the head of the file) the data starts with: | 16:53.52 |
| 00000050: 00 0A 00 00 F0 13 00 00 21 12 12 21 21 11 12 11 |........!..!!...| | 16:53.53 |
| 00000060: 22 21 22 22 11 21 22 11 12 12 22 11 11 22 11 21 |"!"".!"..."..".!| | 16:53.55 |
| 00000070: 21 22 12 21 21 11 12 11 11 22 12 21 21 11 12 21 |!".!!....".!!..!| | 16:53.56 |
| but with your code (the header is different, but ignoring that): | 16:53.58 |
| 00000020: 00 74 40 00 F6 09 00 00 00 74 40 00 F6 09 00 00 |.t@......t@.....| | 16:53.59 |
| 00000030: 12 21 21 12 12 11 21 11 22 12 22 22 11 12 22 11 |.!!...!.".""..".| | 16:54.01 |
| 00000040: 21 21 22 11 11 22 11 12 12 22 21 12 12 11 21 11 |!!".."..."!...!.| | 16:54.02 |
| Robin_Watts: so the original began 2,1,1,2,1,2,2,1 and yours is: 1,2,2,1,2,1,1,2 | 16:54.49 |
| thus, the distribution of 1's and 2's is about the same 50% | 16:58.47 |
| Robin_Watts: I don't know if it's anything to be concerned about or not, and I've been working on the new issues from cust 532, so I haven't stepped through the original and your pms_blueNoiseMask to see where the difference comes from | 17:03.23 |
| nice thing is that function is only used for the K plane :-) | 17:03.49 |
Robin_Watts | ray_laptop: I shall investigate more, thanks. | 17:03.49 |
ray_laptop | Robin_Watts: well, I'm glad you spotted an issue with the CMYB plances | 17:04.20 |
| planes | 17:04.24 |
Robin_Watts | ray_laptop: Oh! Could the difference be that I am packing to 8 bits different? | 17:07.18 |
| HL rather than LH ? | 17:07.24 |
ray_laptop | Robin_Watts: looks plausible from the few lines I've looked at (it'd be easier to see if the headers where the same size :-) | 17:09.32 |
Robin_Watts | how are the headers differently sized I wonder... | 17:09.46 |
ray_laptop | Robin_Watts: I haven't looked into what's in their header | 17:10.06 |
Robin_Watts | as I say, I will investigate. | 17:10.21 |
ray_laptop | Robin_Watts: OK. Thanks. No word from mvrhel ? Do you know which day they are going to cust 801 ? | 17:11.09 |
ray_laptop | forgot to ask | 17:11.19 |
Robin_Watts | I think he said wednesday? but I could be wrong. | 17:11.30 |
| no word from mvrhel, no. | 17:11.36 |
ray_laptop | back to fixing the hacks in cust 532 code... | 17:13.10 |
| Robin_Watts: SMS me if you want to chat and I'm not here. | 17:13.33 |
Robin_Watts | ray_laptop: Thanks. | 17:13.44 |
ray_laptop | my stoopid phone keeps powering off :-( Plenty of battery, but I go to call and find that its powered off | 17:14.20 |
Robin_Watts | ray_laptop: Yes, they pack into HIGH first, and evidently I pack into LOW first. | 17:14.51 |
| I think it's gs policy to pack high first, so I'm clearly wrong. | 17:15.07 |
| should be an easy enough fix though. | 17:15.35 |
ray_laptop | OK. That's in the section I wrote, and I thought that I had it for the first byte in the high 4 bits. Sorry | 17:17.12 |
Robin_Watts | no worries, I just need to tweak it a bit. | 17:17.37 |
ray_laptop | Robin_Watts: right. | 17:17.44 |
| going for (more) coffee. bbiab. | 17:18.27 |
Micha` | Robin_Watts: For PDF annotation related questions, is it with Paul that I should deal? | 17:20.57 |
Robin_Watts | Micha`: Probably. He's away today. | 17:21.13 |
Micha` | Good, thanks! | 17:22.55 |
Robin_Watts | ray_laptop: I have fixes here. | 17:57.22 |
| but I can't make their device work to output PGMs for me to compare. Possibly because the version of their device I have here has been hacked about. | 17:57.53 |
henrys | Robin_Watts: are you just compiling the system they uploaded? That has a couple problems. | 18:00.55 |
Robin_Watts | henrys: yeah. I think I had it working at some point, but something has happened to it. | 18:01.34 |
| I forgot to put it into a git repo when I first unpacked it though :( | 18:01.46 |
| Ah. Me being dim. | 18:02.56 |
henrys | Robin_Watts: let me know if you want my diffs | 18:06.36 |
Robin_Watts | no, I have it working now. I was forgetting to run it with -dNumRenderingThreads | 18:06.54 |
henrys | right it won't work without that. | 18:07.18 |
Robin_Watts | yes, it died with a divide by zero in somewhere unexpected, and I thought "ah, that's cos I fiddled with that code" :( | 18:07.51 |
| Gah. Their device and our device give different page sizes. | 18:10.10 |
| Is there an easy way to force the device page size from the command line ? | 18:10.58 |
henrys | isn't it PaperSize | 18:11.32 |
| ? | 18:11.33 |
Robin_Watts | -dDEVICEWIDTHPOINTS=w -dDEVICEHEIGHTPOINTS=h | 18:11.44 |
henrys | Robin_Watts: I assumed they were just using a4, right? | 18:12.33 |
Robin_Watts | yes, but somehow they are ending up with a size of 4960x7018 where we have 4958x7016 | 18:13.27 |
henrys | so -sPAPERSIZE=a4 won't work? | 18:14.07 |
Robin_Watts | henrys: No, both versions are a4 already. | 18:14.22 |
| I bet they are calculating their size from the raster. | 18:14.30 |
henrys | I wonder if that won't cause other problems | 18:14.43 |
Robin_Watts | it'll mean they get 2 random pixels at the edge. | 18:14.58 |
| but I suspect they don't care. | 18:15.08 |
henrys | but you aren't testing that | 18:15.10 |
Robin_Watts | I was hoping to just be able to diff the two... | 18:15.26 |
ray_laptop | Robin_Watts: -g forces the pagesize to pixel level | 18:16.17 |
| Robin_Watts: -gWWWWxHHH | 18:16.27 |
Robin_Watts | ray_laptop: Right, but I fear that ain't gonna help either. | 18:16.35 |
henrys | Robin_Watts: lop off the last 2 pixels? imagemagick can do that but it's deathly slow on large images | 18:18.26 |
Robin_Watts | ray_laptop: If I do -gXxY will that affect the ctm? i.e. will tiger get distorted ? | 18:19.32 |
henrys | Robin_Watts: clipped | 18:20.48 |
ray_laptop | Robin_Watts: that doesn't affect the CTM, but tiger fills the clippath which _is_ affected by the -g | 18:21.17 |
| so if you make it too small, part of the tiger face will get clipped | 18:21.38 |
Robin_Watts | It looks to me like we have matching output now, at least for the first line. | 18:22.03 |
ray_laptop | otherwise it just changes how much gray background there is to the right | 18:22.03 |
| Robin_Watts: great. You are looking at the raw or the PGM's ? | 18:22.32 |
Robin_Watts | ray_laptop: PGMs. | 18:22.39 |
| but the raw should match too now. | 18:22.48 |
| cos the PGMs are generated from the raw data in memory. | 18:22.57 |
ray_laptop | Robin_Watts: yes, I would expect so. So did you change your PGM format to match theirs ? (maxval = 15) ? | 18:23.20 |
Robin_Watts | I did. | 18:23.26 |
| I now have a 'VIEWABLE_PGMS' define that chooses between their format and something that photoshop will open. | 18:23.47 |
ray_laptop | Robin_Watts: can you send me a patch so I can keep my branch up to date ? | 18:23.49 |
Robin_Watts | ray_laptop: Absolutely. | 18:23.55 |
ray_laptop | Robin_Watts: XnView (on windows, maybe other platforms) is free, opens faster and will display their PGM;s | 18:24.24 |
Robin_Watts | convert out.pgm out.png && start out.png works fine too. | 18:24.47 |
| BUT... viewing their PNGs that way is still bad. | 18:25.02 |
ray_laptop | Robin_Watts: I know 'convert', what is 'start' ? | 18:25.19 |
Robin_Watts | because their PNGs only use values from 8..15 rather than 0..15, so they all appear too faint. | 18:25.37 |
| start is a windows command. | 18:25.43 |
| It means "pretend I just double clicked on this file in explorer" | 18:25.59 |
ray_laptop | Robin_Watts: there you go with TLA confusion again. ;-) | 18:25.59 |
Robin_Watts | :) | 18:26.38 |
ray_laptop | Robin_Watts: since I have .pgm extensions linked to XnView, I don't need to 'convert' | 18:26.48 |
| Robin_Watts: but you are right that they are light, rather than expanded to 0..15 | 18:27.15 |
Robin_Watts | right. so 'start blah.pgm' will save you mouse clicks. | 18:27.26 |
ray_laptop | Robin_Watts: I just use the command line, and have a 'display' script that invokes XnView :-) | 18:28.12 |
Robin_Watts | fair enough. start has the benefit of working with any filetype though. | 18:28.34 |
ray_laptop | Robin_Watts: right. Thanks. | 18:29.07 |
Robin_Watts | ray_laptop: So, I'll send Nomura a message about my mistakes. | 18:29.42 |
| Should we do him another drop? This time with all our patches to date in? | 18:30.04 |
| and with the configure stuff? | 18:31.05 |
ray_laptop | Robin_Watts: Yes, I think so. Do you want me to build on linux with your patches first ? | 18:31.06 |
Robin_Watts | ray_laptop: I haven't pulled your patches in yet. | 18:31.18 |
ray_laptop | Robin_Watts: the configure stuff is in the patches I sent | 18:31.23 |
Robin_Watts | so it may be easier (if you wouldn't mind) for you to generate the source drop. | 18:31.37 |
ray_laptop | Robin_Watts: no problem. | 18:31.47 |
| have you sent the patch (or uploaded it) ? | 18:32.05 |
Robin_Watts | OK. I will write him a mail explaining what I have found, and say that you will be in touch about a new source drop later. | 18:32.21 |
| I have. | 18:32.25 |
| emailed it. | 18:32.28 |
ray_laptop | Robin_Watts: OK. Thanks | 18:32.35 |
Robin_Watts | but I discovered that my mail server is on the spamcop blacklist at the moment. Had a shout at my hosting company about that earlier. | 18:33.12 |
| so if the mail does not arrive, please say. | 18:33.24 |
ray_laptop | I have it. | 18:34.41 |
Robin_Watts | ah, my server appears to have cleared the blocklist now. | 18:34.59 |
ray_laptop | My machine is SLOW. I need to reboot. | 18:35.01 |
| Robin_Watts: for some reason, git is telling me "Patch format detection failed." | 18:51.30 |
| I guess I'll just apply them manually ? unless you have an idea what might be wrong | 18:51.56 |
Robin_Watts | ray_laptop: There is an argument you can use to fix that... | 18:51.58 |
| ray_laptop: You might want to check to see if the mail sending has altered the 'From' line. | 18:53.19 |
ray_laptop | Robin_Watts: I even got rid of all that | 18:53.37 |
| cut it down to just the 'diff' line on | 18:53.55 |
| when it had the From line it said >From: ... | 18:54.40 |
Robin_Watts | git am -p0 0001-.... ? | 18:55.37 |
ray_laptop | Robin_Watts: nope. Let me try saving the entire message including both attachments (that seems like what git-mailinfo is set up for) | 18:56.33 |
| Robin_Watts: oh, with just the patches, I used 'git apply'. That seems to work | 18:57.16 |
Robin_Watts | ray_laptop: Except git apply doesn't generate commits. | 18:57.51 |
ray_laptop | a bunch of trailing whitespace issues | 18:57.54 |
| what does it do ? git st doesn't show me any changed files (unless it skipped doing changes because of the trailing whitespace issues) | 18:59.18 |
Robin_Watts | ray_laptop: Try: git diff --cached | 19:00.10 |
| With git you have 3 conceptual areas. The database of stored snapshots (all your commits), a 'staging' area (known as the index), and your working set of code. | 19:01.05 |
| Often when you do operations, they work 'via' the index. | 19:01.21 |
| so when you: "git diff" afterwards you don't see any changes as the index and the working set are the same. | 19:01.55 |
| "git diff --cached" shows what's different between the working set and the index (cache). | 19:02.15 |
ray_laptop | Robin_Watts: OK. I saved the entire email, including the attachments as a file 'Patches.eml' and git am Patches.eml said "Applying: Patches" now my git log has the patches and the commit message has first line "Patches" and the body of the email as the rest of the log message up to the "Subject [PATCH 1/2]..." line | 19:05.43 |
Robin_Watts | ray_laptop: OK, so you can tweak that using git rebase -i HEAD~2 if you want, but it's probably OK as it is. | 19:06.24 |
ray_laptop | Both patches seem to have been applied | 19:06.31 |
| AS A SINGLE PATCH | 19:06.45 |
| Robin_Watts: I'll fix up the commit message locally to include both your comments | 19:09.29 |
Robin_Watts | ray_laptop: What a pain in the ass. | 19:09.53 |
| ray_laptop: Maybe we should put a private git repo on casper that we can share ? | 19:10.08 |
ray_laptop | Robin_Watts: that might be a good idea :-) | 19:12.01 |
Robin_Watts | ray_laptop: Where is your repo currently? | 19:12.25 |
ray_laptop | can I get mine up there somehow ? | 19:12.25 |
Robin_Watts | ray_laptop: Let me create a private repo, and talk you through adding it as a remote to yours. | 19:12.49 |
ray_laptop | Robin_Watts: currently it is a branch on my laptop, but I also push that branch to my directory on peeves | 19:12.55 |
Robin_Watts | Then you can push yours up. | 19:13.07 |
ray_laptop | Robin_Watts: OK. In the meantime, I'm going to build and test on Windows and peeves | 19:13.42 |
Robin_Watts | ray_laptop: OK. I just did: git remote add private robin@ghostscript.com:/home/robin/private_repos/ghostpdl.git | 19:18.19 |
| you should be able to tweak that to what you need. git remote -v should show you your current remotes so you can see what they look like. | 19:18.48 |
ray_laptop | Robin_Watts: OK, now my git remote -v also shows: | 19:22.54 |
| robin ray@ghostscript.com:/home/robin/private_repos/ghostpdl.git (fetch) | 19:22.55 |
| robin ray@ghostscript.com:/home/robin/private_repos/ghostpdl.git (push) | 19:22.57 |
Robin_Watts | Hmm. Not sure that having it called 'robin' is ideal, but... | 19:23.18 |
| git push robin cust801 | 19:23.31 |
| or whatever your branch is. | 19:23.36 |
ray_laptop | Robin_Watts: why not call it robin ? | 19:23.43 |
Robin_Watts | ray_laptop: Well, i have a public repo for ghostscript. | 19:24.01 |
ray_laptop | Robin_Watts: I shouldn't do git push robin cust801:cust801 ? | 19:24.09 |
Robin_Watts | and I have 'mvrhel' and 'ray' and 'tor' etc as remotes for their public repos. | 19:24.39 |
ray_laptop | AIUI, that would create a branch on that repo | 19:24.40 |
Robin_Watts | git push robin X is shorthand for git push robin X:X | 19:24.59 |
ray_laptop | OK | 19:25.13 |
| Robin_Watts: I'll change the remote to robin_private | 19:26.09 |
Robin_Watts | I have to shoot into banbury to pick Helen up. back later. | 19:26.41 |
ray_laptop | Robin_Watts: I should have the build and tests (basic) done by then | 19:27.11 |
Robin_Watts | ray_laptop: Fab. Thanks. | 19:27.29 |
ray_laptop | Robin_Watts: (for the logs) the PGM's on linux are all empty. Their device with: -r600 -dBandHeight=128 -o x. . Windows does the same thing, so I'll debug it there. | 19:40.33 |
| Robin_Watts: nm. Somehow my git repo still had #define NO_OUTPUT (which _still outputs the PGM headers, BTW) | 19:57.15 |
| works on windoze now. Building on linux... | 20:05.19 |
| OK. Runs on linux. Cleaning up and packing up a password protected zip for them | 20:06.59 |
Robin_Watts | ray_laptop: oh ass. | 20:19.00 |
| ray_laptop: What is your branch called? | 20:19.12 |
ray_laptop | Robin_Watts: OK. All seems to work, and I just sent them an email with instructions to get the .zip (protected) snapshot. | 20:44.13 |
| Robin_Watts: going to push it now. | 20:44.29 |
| Robin_Watts: error: insufficient permission for adding an object to repository database ./objects | 20:45.43 |
| Robin_Watts: it seems that although the .git directory has group gs-priv, all of the contents have group "robin" | 20:49.44 |
| Robin_Watts: I hope you don't mind. I did chgrp -R gs-priv * in your ghostpdl.git and it works OK now | 20:52.15 |
| Robin_Watts: and the 'xxxx' branch (you have to guess the 4 letters ;-) ) is there now when I do git branch | 20:53.23 |
| I have to run an errand. I'll check logs later to see if you were able to grab it | 20:54.29 |
Robin_Watts | ray_laptop: perfect. I forgot the -R. | 23:44.01 |
ray_laptop | Robin_Watts: OK. Thanks | 23:45.18 |
| wow! my Skype just said that Takane-san is online. I wonder if that is just his office CPU. I thought they were visiting customers today (or traveling) | 23:48.50 |
Robin_Watts | laptop? | 23:49.03 |
| wifi on the train? or in the hotel? | 23:49.17 |
ray_laptop | Robin_Watts: maybe. I think that some of the "high end" trains have wi-fi | 23:49.33 |
Robin_Watts | UK trains do these days. | 23:49.42 |
| or he might be on a cellphone. | 23:49.50 |
ray_laptop | even us laggards in the US have some commuter buses and trains with wi-fi. Some cities were motivated whey Google started doing that 5+ years ago. | 23:52.04 |
| s/whey/when/ | 23:52.21 |
| Robin_Watts: BTW, I continually forget, but it seems that mvrhel, you and I all have birthdays coming up (facebook is good for something after all) | 23:54.04 |
Robin_Watts | Indeed. Apparently today is Mvrhel's in fact. | 23:54.40 |
| Forward 1 day (to 2013/11/19)>>> | |