| <<<Back 1 day (to 2013/09/29) | 2013/09/30 |
kens | ducks | 14:07.24 |
tor7 | hi | 14:07.48 |
ghostbot | privet | 14:07.48 |
henrys | kens:doesn't compile - at least since the psi split off ... | 14:33.53 |
kens | well, kill it thernI say | 14:40.58 |
| cannot possibly be important | 14:41.07 |
Robin_Watts_ | tor7: 1 commit for review on robin/master (the last one) | 14:48.10 |
tor7 | Robin_Watts_: I'll need to study that one in more detail | 14:51.44 |
Robin_Watts_ | tor7: Sure. | 14:51.53 |
tor7 | but first off, why calculate s.rect twice? | 14:51.59 |
| you do the fz_gel_scissor, invert_matrix, transform_rect and += linewidth dance twice | 14:52.44 |
Robin_Watts_ | tor7: D'Oh. | 14:52.59 |
| Cut and paste error. I'll fix. | 14:53.09 |
| Better version pushed. | 14:55.56 |
zeniko | Hi all, I've got a Ghostscript question for a change: | 15:06.05 |
| Trying to convert http://tinyurl.com/Sample-ps into a PDF document | 15:06.15 |
| using a command line like | 15:06.27 |
| gswin32c.exe -q -dSAFER -dNOPAUSE -dBATCH -dEPSCrop -sOutputFile=Taller2.pdf -sDEVICE=pdfwrite -c .setpdfwrite -f Taller2.ps | 15:06.27 |
| On some systems, that gets me Letter sized pages, | 15:06.45 |
| even though the document should state that the pages are to be A4. | 15:07.00 |
Robin_Watts_ | zeniko: -sDEFAULTPAPERSIZE=letter | 15:07.07 |
| zeniko: How does the document state that? | 15:07.20 |
zeniko | %%BoundingBox: 0 0 595 842 | 15:07.40 |
Robin_Watts_ | There might be comments in the file saying A4? | 15:07.42 |
| That's a comment. | 15:07.45 |
zeniko | That should have been a double-% | 15:08.06 |
Robin_Watts_ | i.e. it's not a postscript command. | 15:08.12 |
zeniko | Then there's %%%%DocumentPaperSizes: a4 | 15:08.33 |
| and between %%%%BeginSetup and %%%%EndSetup there's %%%%PaperSize: A4 | 15:08.58 |
Robin_Watts_ | Again, anything that begins with a % is just a comment. | 15:09.07 |
zeniko | Who are these comments aimed at, then? | 15:09.45 |
Robin_Watts_ | There is a standard for the way such comments should be laid out (DSC) (Document Structure Comments or something?) | 15:10.08 |
| so some programs that process postscript files can understand them. | 15:10.29 |
| but gs (as standard) is not one of those programs. | 15:10.44 |
zeniko | Does that mean that I'll have to parse the file myself to determine the desired paper size and then pass that to Ghostscript through -sDEFAULTPAPERSIZE? | 15:11.36 |
Robin_Watts_ | I am not an expert in such things; kens ? | 15:12.12 |
kens | GS does parse DSC< a properly constructed PS file will contain postscript to request the media | 15:12.39 |
zeniko | kens: What if the PS file fails to request the media and only contains that information in the DSC? | 15:15.21 |
kens | then you get the default | 15:15.29 |
zeniko | What does Ghostscript parse the DSC for, then? | 15:15.46 |
Robin_Watts_ | zeniko: I think kens meant "GS doesn't parse DSC" | 15:16.13 |
kens | basically nothing, but you can ask it to parse stuff out | 15:16.14 |
| No I didn't, GS does parse comments | 15:16.25 |
| and it has a DSC parszer | 15:16.31 |
Robin_Watts_ | oh, ok, sorry. | 15:16.32 |
kens | But it doesn't *use* it (generally) itself | 15:16.52 |
zeniko | So I could ask GS to extract the desired page size from the command line and then feed that information back to it from a script? | 15:17.18 |
kens | No | 15:17.26 |
| well, sort of | 15:17.39 |
| See Language.htm | 15:19.21 |
| You can use ProcessDSCComment to do 'stuff' using PostScript, when the parer encounters a DSC comment | 15:22.08 |
zeniko | Thanks. So I could install a DSC hook on the command line (through -c) and then set the bounding box or paper format for the next page on the fly? | 15:24.16 |
kens | Yes | 15:24.33 |
| Or more accurately 'whenever that DSC commentis encoutnered' | 15:24.56 |
| Beware of PostScript with mulitple media comments in it | 15:25.08 |
zeniko | And now I know why GSView gets the paper size right for that file: it ships its own DSC parser. | 15:31.30 |
kens | quite so | 15:31.50 |
Robin_Watts_ | tor7: ping | 15:34.20 |
zeniko | kens: That's work for me, then. Thanks again for all the info. | 15:34.42 |
kens | NP | 15:34.48 |
zeniko | Robin_Watts: There's one new small patch for review on zeniko/mupdf (the pdf_write_document one) | 15:36.01 |
Robin_Watts_ | zeniko: Thanks. | 15:37.00 |
kens | Night all | 16:05.19 |
Robin_Watts_ | tor7: 2 commits on robin master now (both related to bug 694526). Together they solve it. | 16:07.17 |
| mvrhel_laptop: ping | 16:54.32 |
mvrhel_laptop | Robin_Watts_: pong | 16:55.05 |
Robin_Watts_ | Bug 693070 | 16:55.12 |
| Basically, psdrgb is broken at the moment. | 16:55.25 |
| And has been since we did the devn change, I think. | 16:55.46 |
| James Cloos has submitted a patch to fix it, and I have a slightly revised version of it here. | 16:56.04 |
| Let me upload it to the bug. Could I trouble you to give it a once over? | 16:56.15 |
mvrhel_laptop | Robin_Watts_: ok. I am neck deep in trans stuff now and will look at it today | 16:56.42 |
Robin_Watts_ | mvrhel_laptop: No problem at all. Just running through my old bugs trying to get easy ones out of the way. | 16:57.30 |
mvrhel_laptop | really that device should be tested on all the files to see if anything explodes. | 16:57.46 |
| I suspect that there may be several issues with psdrgb | 16:58.05 |
Robin_Watts_ | Well, at the moment, we get all black for tiger. | 16:58.12 |
| now we get plausible tigering :) | 16:58.19 |
mvrhel_laptop | ok. | 16:58.38 |
Robin_Watts_ | Hmm. We could do a run where we make psdcmyk and psdrgb and convert them both to png and compare or something? | 16:59.25 |
marcosw | henrys: you may have already noticed, but I've started entering some additional GhostPCL fuzzing issues. sorry. | 17:05.30 |
henrys | marcosw: so far they've been pretty easy | 17:06.22 |
marcosw | I started with the easy ones :-) (not true, I started with the ones that showed up a lot, this batch occurred much less frequently, some only showed up once in the hundreds thousands of fuzzing iterations I've run, so I'm not sure they worth fixing). | 17:08.14 |
Robin_Watts_ | marcosw: Hi. | 17:09.07 |
marcosw | Robin_Watts_: morning | 17:09.17 |
Robin_Watts_ | How hard would it be to do a test of psdcmyk vs psdrgb? | 17:09.21 |
| I believe that psdrgb is broken currently, and I have a patch (from James Cloos) to fix it. | 17:09.42 |
| actually, maybe it's enough to test psdcmyk against any other rgb device. | 17:10.19 |
marcosw | doesn't psdcmyk generate extra planes for the spot colors? Or am I confusing psdcmyk with a different device? | 17:11.15 |
| brb | 17:12.41 |
Robin_Watts_ | psdcmyk does indeed generate extra planes. | 17:13.08 |
| psdrgb does not. | 17:13.12 |
| and therein lies at least part of the problem with the device :) | 17:13.38 |
marcosw | Robin_Watts_: that will make comparing them a bit tricky... | 17:19.57 |
Robin_Watts_ | marcosw: Well, we'd convert both to rgb first? | 17:20.25 |
| but how about comparing psdrgb to ppmraw or something? | 17:20.39 |
| No extra components to worry about there. | 17:20.58 |
marcosw | comparing psdrgb to ppmraw would be easy. do you want to send me a patch to test or is there a branch? | 17:21.18 |
Robin_Watts_ | My patch from bug 693070 | 17:21.58 |
marcosw | Robin_Watts_: okay will do. | 17:23.45 |
Robin_Watts_ | Thanks. | 17:23.53 |
henrys | marcosw:maybe sans email would be best | 17:43.15 |
| marcosw: or not | 17:52.25 |
rayjj | marcosw: I asked last week.... What do I need to do with the PE server to let you get it up as a cluster machine ? | 18:17.40 |
henrys | anybody know what PCL-M is? Something from adobe I don't think it's related to PCL? Supposed to be a PDL for the mopria standard. | 18:18.19 |
rayjj | henrys: PTSD checklist for the Military ?? ;-) | 18:19.57 |
henrys | that's where my googling led also⦠mopria looks like something doomed to fail but I'm having a look at it anyway. | 18:21.19 |
| bbiab | 18:24.17 |
rayjj | henrys: BTW, have you ever checked out "verypdf" -- they have a PCL to PDF that claims to create high level PDF with searchable text. The interesting thing is they show "\"-J@PJL SET PAPER=A3\"" as the argument to their -pclcmd option | 18:26.20 |
| I find the use of -J interesting since that's what we use | 18:26.38 |
henrys | bastards | 18:26.46 |
| I'll have a look | 18:26.56 |
rayjj | henrys: http://www.verydoc.com/pcl-converter.html | 18:27.17 |
mvrhel_laptop | ugh. just when I thought I had the knockout non-isolated stuff working, ran into one more issue. | 18:28.52 |
| bbiab | 18:36.17 |
rayjj | marcosw: you should be able to log in to 64.183.45.165 -- I've put your keys there (from peeves) and have you in the sodoers | 18:47.18 |
| Robin_Watts_: henrys: (or anyone else): let me know if you want me to add you as a user on it as well, and copy your keys from peeves | 18:48.30 |
Robin_Watts_ | rayjj: Are you going to give it a more memorable name, and add it to the blah.ghostscript.com DNS stuff? | 18:49.19 |
| I'm assuming it's static IP, right? | 18:49.28 |
rayjj | Robin_Watts_: well, it's current host name is peeved | 18:49.50 |
| Robin_Watts_: and, yes, it is a static IP | 18:50.02 |
| it's on the same subnet as peeves, so it should be fast to copy files over to it from peeves (such as the regression files) | 18:51.23 |
Robin_Watts_ | I have to pop out. I will try to add peeved.ghostscript.com later tonight or tomorrow if you don't get to it first. | 18:52.02 |
rayjj | Robin_Watts_: and it has a fair amount of disk space; 1.3Tb made from 3 hybrid 750Gb in a RAID 5 | 18:52.30 |
| Robin_Watts_: OK. I'll add you as a user on it in the meantime | 18:52.56 |
henrys | rayjj:running strings on the exe - it is certainly our code. | 18:55.45 |
rayjj | henrys; for the log. If you fire it up with -pclcmd "\"-h"\" it should be interesting | 19:01.40 |
| Robin_Watts_: you should be able to log in, but I can't test it because the "identity" you have on peeves is not in the the authorized_keys on peeves :-/ | 19:02.40 |
| Robin_Watts_: and there is not identity.pub so I can't just add that to authorized_keys | 19:03.25 |
| now I have to run an errand (get a supplementary AC for the office for when the building AC goes to "night" mode). The theromstat's not in my office so when the guys next door aren't working it gets stuffy | 19:06.07 |
| has for a while, but the PowerEdge makes it noticeably hotter | 19:06.28 |
| I'm going to vent the AC into the office next door since the tubes for their AC are above my drop ceiling. That may cause the building AC to kick on sooner | 19:07.47 |
henrys | rayjj:are you sure they aren't a customer with a new name? the name rings a bell | 19:17.29 |
mvrhel_laptop | yeah! I have the knockout stuff working with all the crazy files that Robin_Watts_ had given me long ago. | 19:54.29 |
| time for a cluster push to see what explodes | 19:54.35 |
Robin_Watts_ | mvrhel_laptop: Woo Hoo! | 19:54.37 |
mvrhel_laptop | Robin_Watts_: you had some very good files for this | 19:54.48 |
| thanks | 19:54.50 |
Robin_Watts_ | I made them when I was looking at mupdf. | 19:55.01 |
| So, if it passes, you can look at removing group alpha ? | 19:55.17 |
mvrhel_laptop | I am a little surprised what all was broken in gs | 19:55.18 |
| Robin_Watts_: yes. that would be the next step | 19:55.44 |
| the group alpha is currently used in the recomposing when the current group is not isolated, but it appears to me that the group alpha values are always the same as the alpha values. however there may be a few optimizations that make use of it. | 19:56.41 |
| I will tread carefully on this | 19:56.49 |
| great. lots of segvs..... | 20:49.43 |
| need to run an errand. bbiab | 20:50.38 |
rayjj | Robin_Watts_: have you tried peeved (64.183.45.165) yet ? | 21:26.27 |
| Robin_Watts_: can you log in ? | 21:26.34 |
| marcosw: same question for you ? | 21:28.13 |
| have my AC. It works. Now to rig the exhaust (the second hardest part -- the hardest part was lugging 97 lbs up the stairs) | 21:29.48 |
| it's quite a bit larger than I expected (I hadn't really looked at the dimensions on the spec sheet) | 21:31.02 |
Robin_Watts_ | rayjj: Gimme a sec and I'll try. | 23:28.02 |
| rayjj: And I'm in. | 23:28.46 |
| rayjj: And I've installed emacs, so I'm happy :) | 23:34.59 |
| Forward 1 day (to 2013/10/01)>>> | |