| <<<Back 1 day (to 2014/03/13) | 2014/03/14 |
Robin_Watts | new ups. reboot. | 13:54.18 |
| .quit | 13:54.29 |
kens | did the power cuts fry your UPS robin ? | 14:10.23 |
Robin_Watts | It wouldn't start up afterwards :( | 14:10.38 |
| It's been dodgy for a while. | 14:10.48 |
kens | Hmm not good for a UPS | 14:10.49 |
ray_laptop | Robin_Watts: mvrhel: Can one of you review the patch on my repo. It fixes the repeated vertical lines every 32 pixels with landscape images. http://git.ghostscript.com/?p=user/ray/ghostpdl.git;a=commitdiff;h=b98312c88836476c9ca1fa2293bb40ae0fee955c | 14:10.50 |
| Robin_Watts: mvrhel: If it isn't clear enough what I did, I moved the 'move_landscape_buffer to inside the j < spp_out loop, and took the movinf out of reset_landscape_buffer which follows the loop | 14:12.39 |
| Robin_Watts and mvrhel: If either of you have time today, I'd also like to have you look at the 4-bit halftoning approach. I used the same approach to use the threshold buffer as ex-cust 801 did, but it doesn't look as nice as the 'standard' slow approach | 14:16.09 |
mbrindamour | Hello all - Looking for some advice on an issue we're having with ghostscript. Have not found any other mentions of the specific problem on google/forums... | 14:17.06 |
| If anyone has any ideas, it would be much appreciated. Here goes: | 14:17.26 |
| We are converting a PDF (generated from Jasper Reports) to TIFF, using the following command: | 14:17.51 |
| gs -dPARANOIDSAFER -dNOPAUSE -dBATCH -sDEVICE=tiffg4 -r300 -sPAPERSIZE=letter -sOutputFile=BAS_TIFF_qa4_1.tiff BillAnalysisSummaryReport_MBrindamour_qa4_2014_03_13_1039.pdf | 14:18.17 |
| Using ghostscript 8.15.2 (which I realize is old)... | 14:18.44 |
chrisl | Not "old", but "really, very, very old"...... | 14:19.27 |
mbrindamour | On two particular systems, this PDF, which is in landscape, is output rotated 90 degrees in TIFF. | 14:19.31 |
| on another system, using the same PDF, same command, same version of ghostscript, it is being output properly (no rotation). | 14:20.01 |
kens | THen there is some difference in the two setups | 14:20.26 |
mbrindamour | The only difference I have come across is that the kernel version is a bit older on the two offending systems... but we're talking minor version older, not major. | 14:21.02 |
kens | By the way, -sPAPERSIZE won't do anything because the Media will beset from teh PDF MediaBox. You would have to set -dFIXEDMEDIA as well for that to have any effect | 14:21.15 |
mbrindamour | Any thoughts on what other differences may interfere? | 14:21.16 |
kens | To be honest, no, not really. | 14:21.35 |
| I'd have to see the PDF file at minimum | 14:21.50 |
mbrindamour | good to know on the PAPERSIZE setting @kens. Thank you | 14:21.50 |
ray_laptop | mbrindamour: that is such an old version that it's hard to recall, but there have been changes/improvements in the page rotation | 14:22.20 |
mbrindamour | I'd be happy to share it to you - have a preferred method? | 14:22.30 |
chrisl | mbrindamour: you should try a newer version of GS first | 14:22.48 |
kens | Any public means is fine if you;re happt to have it shared publicly | 14:22.50 |
ray_laptop | mbrindamour: please test with the current release, because even if you open a bug, that's the first thing we will do. Then if it works, we close the bug as FIXED | 14:23.20 |
mbrindamour | @chrisl - agreed - not sure what it will take to get these particular servers upgraded... I think I need to interface with a different team for that... will investigate - I would prefer to do that as well. | 14:23.36 |
ray_laptop | mbrindamour: but in general, the best thing to do is open a bug and attach the file | 14:23.48 |
mbrindamour | @ray_laptop - ok, thanks... | 14:23.56 |
chrisl | mbrindamour: well, assuming we find a problem, and fix it, you'd need to get a new version anyway...... | 14:24.16 |
ray_laptop | chrisl: unless the patch is simple enough for mbrindamour to back port to the old gs (*UNLIKELY*) | 14:24.59 |
mbrindamour | yeah... I was looking for suggestions on what might interfere in the environment first off - but I understand that's harder to pinpoint. | 14:25.17 |
chrisl | ray_laptop: and in the Postscript...... | 14:25.24 |
mbrindamour | (and/or cli option adjustments) | 14:25.41 |
chrisl | mbrindamour: possible different default page sizes? | 14:26.00 |
ray_laptop | mbrindamour: and unless the patch is in the PS code, you'd have to rebuild on the servers anyway, in which case you're back to interfacing with the same team that would upgrade | 14:26.01 |
kens | THere's nothing I can think of, but that version is so archaic its hard to be certain | 14:26.01 |
ray_laptop | chrisl: but the default pagesize doesn't apply if he has -sPAPERSIZE=letter | 14:26.35 |
mbrindamour | to each of you, right, to your respective points. Thank you. | 14:26.46 |
kens | The default page size is irrelevant (in current implementations) because the media size is set form the PDF MediaBox | 14:27.03 |
chrisl | ray_laptop: just trying to think of environmental things that might influence it | 14:27.15 |
ray_laptop | kens: right. And I can't see how the default page size would affect the orientation since that's set when the PDF does the setpagedevice | 14:27.45 |
kens | exactly... | 14:27.55 |
ray_laptop | but this was 8.15, so, /who knows what funky things might have been going on/ | 14:28.28 |
kens | Like I sai, the real problem is that the version is so old I can'r be certain it behaves the same way | 14:28.35 |
chrisl | mbrindamour: FWIW, you could probably build ghostscript yourself, and use it just for your specific user as a test | 14:28.37 |
ray_laptop | mbrindamour: or if you have a windows box around, download a pre-built binary to test | 14:29.12 |
mbrindamour | well, I have tested on windows (cygwin) | 14:29.32 |
| works perfectly. | 14:29.37 |
| it's specific to these systems | 14:29.43 |
kens | Our old binaries only go back to 8.62 (10 years ago) | 14:30.04 |
ray_laptop | kens: one could pull 8.15 from the git repo -- it goes back to 6. something. But that's not something we would do | 14:31.04 |
kens | ray_laptop : yes, but I'm not really inclined to do that, too much like hard work, and I bet I couldn't build it on Windows. In fact, given my and Chris' last experiences in this area, it might be difficult to build on a recent Linux too. | 14:31.49 |
mbrindamour | @chrisl I will look into building within my user to give it a shot. Thank you all for your suggestions! | 14:31.50 |
| (building latest) | 14:31.58 |
ray_laptop | ooh. Looking at the git repo, I see that the first release there is 5.55. On my birthday in 1998 (right after I started working for Artifex) | 14:32.57 |
chrisl | mbrindamour: what OS is this on (or did I miss that)? | 14:33.09 |
kens | boggles slightly | 14:33.14 |
mbrindamour | rhel | 14:34.30 |
kens | Hmm,I think 8.15.2 was *ESP* Ghostscript | 14:34.40 |
mbrindamour | Linux qatemp04.allegientsystems.com 2.6.18-53.1.4.el5 #1 SMP Wed Nov 14 10:37:27 EST 2007 x86_64 x86_64 x86_64 GNU/Linux | 14:34.48 |
chrisl | Eek, ESP GS all bets are off :-( | 14:35.15 |
ray_laptop | mbrindamour: can you do "gs -h" on that ? | 14:35.17 |
kens | chrisl yes, that was pretty much what I was thining | 14:35.28 |
| I don't think Artifex did a release with that version number | 14:35.41 |
chrisl | kens: probably 8.15 was the Artifex one, and 8.15.2 was the 8.15 release with the ESP patches applied | 14:36.22 |
ray_laptop | kens: right. Peter went to major.minor with 3.xx 2.x.x was the last Aladdin/Artifex release with 3 parts | 14:36.35 |
kens | The web page I found it on claims to date from 2007, and yes I expect you're right regarding the version number | 14:36.45 |
mbrindamour | gs -h ESP Ghostscript 8.15.2 (2006-04-19) Copyright (C) 2004 artofcode LLC, Benicia, CA. All rights reserved. Usage: gs [switches] [file1.ps file2.ps ...] Most frequently used switches: (you can use # in place of =) -dNOPAUSE no pause after page | -q `quiet', fewer messages -g<width>x<height> page size in pixels | -r<res> pixels/inch resolution -sDEVICE=<devname> select device | -dBATCH exi | 14:37.20 |
ray_laptop | my very first ghostscript was 2.5.2 | 14:37.25 |
chrisl | mbrindamour: if you are lucky, you might be able to try the prebuilt LInux binary here: http://www.ghostscript.com/download/gsdnld.html - but your system *might* be too old. | 14:37.39 |
kens | OK so ESP Ghostscript, we *really* can't help with that, except to say please don't use that ;-) | 14:37.47 |
mbrindamour | :-) | 14:38.06 |
kens | mbrindamour : ESP was a fork of Ghostscript, so we don't really know too much about it. Its no longer required, the forks were merged some time ago | 14:38.49 |
ray_laptop | chrisl: do those pre-built binaries not use shared libs? That would be handy for us to have for when people have strange problems (due to some so that is out of date) | 14:38.56 |
chrisl | ray_laptop: they use the minimum viable shared libraries. I leave out the X devices, libpaper, fontconfig etc. But there are still some linked in unforunately | 14:40.06 |
ray_laptop | chrisl: I was thinking about the ones that we've had problems with. openjpeg most notably | 14:40.51 |
| and maybe freetype | 14:41.16 |
chrisl | ray_laptop: they use the code we ship for our third party libs | 14:41.39 |
ray_laptop | chrisl: good. that's what I was concerned about | 14:42.10 |
mbrindamour | sorry, was discussing upgrading with my scrummaster, catching up | 14:43.17 |
chrisl | ray_laptop: it's still a bit of a crap-shoot because even the cut down list of linked libraries occasionally break the ABIs.... | 14:43.31 |
mbrindamour | @chrisl - downloading that url now... so I should just be able to untar this and run it in place, yes? | 14:45.57 |
chrisl | mbrindamour: yes, you'll find the executable name longer than the installed ones - I add version numbers to these binaries | 14:46.43 |
mbrindamour | ok, thanks... will let you know in a few minutes | 14:47.24 |
| (I know you're dying to find out the result. :-) ) | 14:47.46 |
chrisl | mbrindamour: I'm on the edge of my seat - this is better than watching sport! ;-) | 14:48.21 |
mbrindamour | depends which sport - I prefer hockey. :-) | 14:48.44 |
chrisl | by the fact I prefer this, it should be obvious I don't watch much sport...... | 14:49.43 |
mbrindamour | nor do I, except hockey. :-) | 14:53.13 |
| it works properly with that binary. | 14:53.21 |
chrisl | So, it may have been a bug or "feature" that's been fixed, or (unlikely) it may be related to some of libraries I leave out of that binary | 14:54.19 |
| mbrindamour: note that the entire TIFF output has been rewritten since 8.15, to use libtiff....... | 14:55.10 |
| s/TIFF output/TIFF output code | 14:55.34 |
mbrindamour | cool | 14:56.24 |
| Would extracting this binary be the standard installation (so I can advise our hosting team properly), or is there a full installer built for rhel that I should point them at? | 14:57.21 |
chrisl | mbrindamour: for Linux/Unix you normally build/install from source | 14:57.51 |
mbrindamour | ok | 14:58.25 |
chrisl | mbrindamour: I doubt an up to date Ghostscript RPM would install on something that old..... | 14:58.38 |
mbrindamour | is there any reason why using this tarball and changing the /usr/bin/gs link would be bad? | 14:59.40 |
chrisl | mbrindamour: the main thing is that the binary in there doesn't include the X display devices, so it won't let you view on screen | 15:00.52 |
mbrindamour | I don't think we need to. | 15:01.03 |
| this is run programattically only. | 15:01.10 |
chrisl | mbrindamour: personally, I would prefer to keep the original binary, and install this as a new one, called something like gs910, and tweak the calling program to use it | 15:01.59 |
mbrindamour | yes, I think that would likely be our plan. | 15:02.39 |
| Thanks so much for your help, all of you. | 15:03.38 |
chrisl | mbrindamour: glad to help | 15:03.48 |
mbrindamour | You never know coming into a support channel if people are going to call you a newb and make fun - was nice to see how willing you are... Will pass that around. :-) | 15:04.15 |
| much appreciated. | 15:04.28 |
chrisl | mbrindamour: we'll wait until you're gone, then we'll make fun ;-) | 15:04.42 |
mbrindamour | no doubt. | 15:04.50 |
| ;-) | 15:04.52 |
Robin_Watts | Please don't pass it around. We'll only get lots more newbs in here :) | 15:05.02 |
ray_laptop | mbrindamour: don't let too many people know or we'll be swamped (JK) ;-) | 15:05.05 |
| mbrindamour: and you can check to logs to see what we say about you after you leave ;-) | 15:05.39 |
mbrindamour | lol | 15:06.22 |
chrisl | Yeh, the logs do keep us honest - until we edit them...... | 15:06.27 |
ray_laptop | but we usually reserve our most acerbic comment for paying customers ;-) | 15:06.31 |
| mostly because we have to work on their problems, no matter how unreasonable they or their problems are | 15:07.22 |
mbrindamour | Hopefully I was at least reasonable. :-) | 15:08.02 |
ray_laptop | mvrhel_laptop: please see above for request for reviewing a patch (you or robin_watts) | 15:08.06 |
| mbrindamour: quite | 15:08.16 |
| our customers often refuse to upgrade and want a patch for an old version | 15:09.18 |
| not realizing that they are still shipping a version with myriad bugs that have been fixed, and any spot patching hasn't been thoroughly tested for side effects | 15:10.15 |
| chrisl: do you still put a count of bugs fixed and enhancement bugs that we've done since the last release in the release notes ? | 15:11.36 |
chrisl | ray_laptop: no, that was dropped before I took over the releases | 15:12.24 |
ray_laptop | in particular it would be nice to note how many segfault or other potential security bugs, how many other bugs, and how many enhancements. | 15:13.00 |
chrisl | ray_laptop: I have no clue how I'd get that information out of bugzilla | 15:13.36 |
Robin_Watts | ray_laptop: Looks plausible to me. | 15:13.44 |
| but I don't know the code well. | 15:13.53 |
chrisl | What code? | 15:14.08 |
ray_laptop | I may do something like that (with a history going back several releases) to give to scott-san for his customer newsletter. Or at least to have to use when customers drag their heels about upgrading | 15:14.30 |
| chrisl: the fast thresholding landscape code | 15:14.44 |
| gxht_thresh.c | 15:14.51 |
chrisl | Ah, I got confused, thought that was about the stats gathering | 15:15.05 |
ray_laptop | robin_watts: thanks. | 15:15.06 |
| chrisl: no, that's in response to my request for patch review | 15:15.28 |
chrisl | ray_laptop: the problem with gathering the stats like that is separating seg fault fixes from others | 15:16.29 |
ray_laptop | chrisl: well, there are 148 bugs with segfault in the description and 252 with segfault in a comment (114 with both) | 15:19.18 |
chrisl | ray_laptop: or crash, access violation, bus error, GPF......etc | 15:19.59 |
ray_laptop | bugzilla gives up at 500 with SEGV in the comment | 15:20.11 |
tor8 | ray_laptop: did you look at getting the mujs.com domain yet? | 15:21.13 |
ray_laptop | chrisl: oh, /only/ 416 if I don't include mupdf | 15:21.23 |
| tor8: no, I thought the "marketing department" was going to come up with a name. I guess it won't hurt to ask | 15:22.07 |
| as long as we don't sound too interested | 15:22.33 |
tor8 | ray_laptop: I suspect the "marketing department" is too busy worrying about smartoffice to do much else for the near future :) | 15:22.53 |
| ray_laptop: if we could get it, I think we could call it a day and settle for mujs as a name, if not then we'll have to pester the marketing department some more | 15:23.25 |
ray_laptop | tor8: probably true w.r.t SOT | 15:23.59 |
| tor8: OK, I have a price. I'll call Miles to see if he wants to spend the $,$$$ for the name. | 15:32.55 |
tor8 | ray_laptop: that was quick! | 15:33.08 |
ray_laptop | yeah. They even have special discounts for March and same day purchase. :-) | 15:33.54 |
| left message for Miles on his cell. Also told Scott in case he and Miles talk, he can make sure Miles calls me. | 15:41.41 |
| henrys: see my last two comments, FYI | 15:41.57 |
Robin_Watts | reboot. for a change. | 15:51.56 |
henrys | marcosw: are you about? | 16:01.33 |
| or Robin_Watts do you know how to edit the irc logs? | 16:02.28 |
Robin_Watts | henrys: i am, and I do. | 16:02.38 |
| What would you like me to edit? | 16:03.09 |
henrys | 01/07/2014 any sentence with smart office. I'm not sure why but I'll find out soon. | 16:03.52 |
| actually I don't even know if that is the problem I'm just guessing | 16:06.33 |
| Robin_Watts_: internet problems today? | 16:14.47 |
Robin_Watts_ | evidently :( | 16:14.54 |
Robin_Watts | henrys: OK. I've removed ALL the logs for 2014/01/07 for now. | 16:17.16 |
| When we know what was said that was offensive, I can put an edited version back. | 16:17.36 |
Robin_Watts | hopes it wasn't me | 16:17.49 |
henrys | turned out to be nothing. somebody searched for Smart Office + Artifex found that and asked Miles if it should be a private discussion and Miles thought it was upon brief reading. I've explained everything and we're good, sorry for the interruption. | 16:18.57 |
Robin_Watts | so I should put it back? | 16:19.10 |
henrys | Robin_Watts: don't bother | 16:19.34 |
Robin_Watts | but our logs will be incomplete! :) | 16:19.55 |
henrys | Robin_Watts: well if that will upset you by all means undo the changes. | 16:21.39 |
paulgardiner | Was it something that we might want to be careful of in the future in any case? | 16:22.29 |
henrys | paulgardiner: nope | 16:23.41 |
paulgardiner | Okay good. I have a vague recollection of what it might have been | 16:24.04 |
ray_laptop | Robin_Watts: sorry -- I hadn't noticed the tab pop up | 16:34.17 |
henrys | paulgardiner: where are you at with the printing stuff? I assume we can disable it if absolutely necessary. | 16:34.56 |
| paulgardiner: software imaging is off the table for now price-wise | 16:35.13 |
Robin_Watts | henrys: Did embedded-general get back with a revised price? | 16:35.39 |
henrys | yes and the price was too high and miles is still negotiating plan B is to use the expired contract until we can cobble something together. | 16:36.27 |
paulgardiner | henrys: I have the iOS app building as far as the signing stage. I believe I can make it complete once I have access to the keys. So I think I'm ready to go once we know for sure we need to pull it. | 16:37.37 |
| The android app we've already built so shouldn't be a problem. | 16:38.22 |
| I'm also fighting with the lib builds for a customer. | 16:38.43 |
| Joseph is looking at the iOS one. I'm looking at the Android one. It is giving up a good fight. | 16:39.22 |
henrys | paulgardiner:okay then I wouldn't worry about printing at all for now. | 16:40.00 |
paulgardiner | henrys: you mean concentrate on the customer for now until we have a definitive decision on printing? | 16:41.15 |
Robin_Watts | Embedded General do a ticprint app for ios. | 16:42.54 |
paulgardiner | s/you/do you/ | 16:43.16 |
henrys | paulgardiner: right customer and the bugs | 16:43.40 |
paulgardiner | Good. Will do | 16:44.14 |
ray_laptop | Robin_Watts: re the mujs.com, I'll leave the negotiations up to Miles. It's not my strong suit, and I don't know how much he wants the name (since we aren't committed to it). It just means he (and Scott?) have to come up with a different name. | 16:45.51 |
Robin_Watts | ray_laptop: Yeah. That seems silly money to me though. | 16:46.20 |
paulgardiner | Robin_Watts: I've seen ticprint in the build. We may be able to support it without including it in SO, in the same way the Android app can print to Sofware Imaging provided the user has bought it in addition to SO | 16:46.32 |
Robin_Watts | paulgardiner: Right, that's what I was thinking. | 16:46.50 |
| So we wouldn't be saying "You can't do printing any more", we'd be saying "In order to do printing in future, you will need this additional app installed" | 16:47.14 |
| which will probably annoy people, but hopefully not as much. | 16:47.21 |
ray_laptop | Robin_Watts: they probably value it based on the number of outfits that show up in a google search for mujs. Since there are so many disparate outfits, one of them might bite | 16:47.27 |
paulgardiner | ... although what I see in the build options seems to provide for either print or no print | 16:47.35 |
Robin_Watts | ray_laptop:http://www.sedo.com do selling of names. The valuations there seem hopeful. | 16:48.41 |
paulgardiner | Robin_Watts: presumably a company cannot submit an app that handles Android intents and then demand a licence for use of that interface | 16:48.50 |
Robin_Watts | paulgardiner: Why not? | 16:49.21 |
paulgardiner | It would break the whole intent/activity scheme. I'm just guessing, but I can't see how it would work. | 16:51.09 |
chrisl | marcosw: I have pushed the branch for the release, and the tag for the release candidate. | 16:51.26 |
ray_laptop | Robin_Watts: they have the same price for mujs.com (without the discounts) | 16:52.32 |
paulgardiner | We probably should check then. I was assuming we could enable SO talking to Software Imaging, but not bundle their print app so it would work only if the user had bought it in addition | 16:52.58 |
Robin_Watts | ray_laptop: According to themhttp://www.wss.co.uk was worth a lot, IIRC. | 16:53.25 |
ray_laptop | chrisl: so the next version after 9.12 is 9.14 ???? | 16:57.34 |
chrisl | ray_laptop: if we're skipping 9.11, why not 9.13? | 16:58.10 |
| ray_laptop: if it offends, I'll make it 9.13..... | 16:58.44 |
ray_laptop | chrisl: more like "why skip it". Are you planning on having to make a follow up "this version really works" release ? | 16:59.04 |
chrisl | ray_laptop: just for the 13..... | 16:59.59 |
henrys | ray_laptop: probably not I already bothered Robin_Watts with this once today. Where do the logs live and I'll fix it myself. | 17:03.03 |
| ? | 17:03.05 |
Robin_Watts | henrys: ~ghostbot | 17:03.37 |
| Fir today you'll need to edit several files. Let me. | 17:03.48 |
chrisl | ray_laptop: actually, maybe you've hit on a good idea: maybe Ghostscript versions should increment by two, automatically leaving us a "spare" fallback number if things go wrong! | 17:04.32 |
henrys | Robin_Watts: okay | 17:04.59 |
ray_laptop | chrisl: I don't see your ;-) | 17:05.01 |
| chrisl: but I hope you're kidding | 17:05.13 |
chrisl | ray_laptop: we'll see - if this release goes okay, then I'm kidding...... :-) | 17:05.58 |
ray_laptop | chrisl: we don't really need that since it isn't hard to bump numbers -- it's not like we ever ship X.XX PRE-RELEASE | 17:06.08 |
chrisl | ray_laptop: btw, your vertical lines on planar fix didn't make the RC, but I've cherry picked it onto the release branch | 17:07.06 |
ray_laptop | chrisl: OK. Thanks. | 17:07.22 |
| chrisl: I was counting on you not doing it until Monday ;-) | 17:07.44 |
Robin_Watts | henrys: OK, I think that's edited out. Give it 5 mins, then check it all :) | 17:08.08 |
chrisl | ray_laptop: no, I was kidding - building the expectation of failure into our release process wouldn't do much for the public image...... | 17:08.32 |
henrys | Robin_Watts: thanks will do | 17:09.30 |
Robin_Watts | henrys: For future reference, the current log is ~ghostbot/irclogs/20140101 (or the current date), but that's updated live, so you have to edit it in quiet moments. | 17:11.38 |
| The html is generated from that every 5 mins. | 17:11.48 |
| At midnight it gets copied into the 2014/01/01 dir. | 17:12.08 |
ray_laptop | Robin_Watts: looks sanitized to me | 17:12.32 |
henrys | Robin_Watts: I'll put it in my notes | 17:12.35 |
Robin_Watts | tor8: Ping! | 17:16.42 |
tor8 | Robin_Watts: pong! | 17:23.42 |
Robin_Watts | tor8: I thought I had my cleaning code working. | 17:24.02 |
tor8 | also: yeah, that name isn't worth what they're asking, by far. | 17:24.04 |
Robin_Watts | but alas, there still looks to be a problem with it. | 17:24.24 |
| BUT... it would be worth you having a look at some of the reviews on robin/master anyway. | 17:24.38 |
| In particular the fz_stream rework one. | 17:24.47 |
tor8 | Robin_Watts: with cleaning code, do you mean the op-filter stuff or the pdfclean rework? | 17:26.01 |
Robin_Watts | the op filter stuff. | 17:26.11 |
| the pdfclean rework has gone in. | 17:26.17 |
ray_laptop | well, that's disappointing :-( The CMYK planar subtractive fast 4-bit thresholding is SLOWER than the non-planar code. I guess I don't have to have anyone review that commit | 17:42.51 |
| Oh, maybe not. I think there's a bug in the switch to the detection in HEAD | 17:46.29 |
| yep. There's a bug. CMYK 4-bit is totally broken in HEAD. Maybe somebody can double check that for me. gs -sDEVICE=pkmraw -o x.ppm -dUsePlanarBuffer -dGrayValues=16 J8_landscape.ps | 17:57.10 |
Robin_Watts | ray_laptop: Just give me a mo. | 17:58.18 |
| running now. | 17:59.31 |
| crashes for me. | 17:59.45 |
ray_laptop | Robin_Watts: well that qualifies as broken. | 18:00.39 |
| Robin_Watts: please try with -dMaxBitmap=400m | 18:01.13 |
| (page mode) | 18:01.17 |
| I got some kind of low res, darkened output on Windows | 18:02.01 |
| I have to go now. I have a fix for this for HEAD | 18:02.18 |
Robin_Watts | Crashes in page mode too. | 18:02.26 |
ray_laptop | and with it, my 'fast' code is faster. | 18:02.33 |
Robin_Watts | but I get the darkened version. | 18:02.38 |
| Let me update. | 18:02.49 |
ray_laptop | If you can, can you pull from my repo and try that version ? | 18:02.51 |
| dr. appt. bbiaw | 18:03.06 |
Jogux | is the picsel wiki any closer to being accessable? (might possibly save me from making blind guesses as to what versions of tools the build/release process is expecting to find....) | 18:06.54 |
Robin_Watts | Jogux: Let's see if marcosw is here. marcosw ? | 18:08.12 |
| ray_laptop: OK, with the HEAD version in pagemode, I get bright colors, but it looks like the pixel columns down the page are reversed in bands. | 18:18.23 |
| or garbled somehow anyway. | 18:18.43 |
Jogux | gnnnhs as I find a second place in the build that falls over because it was assuming it was checked out from CVS | 18:21.48 |
henrys | Jogux: do you need a particular file from the NAS that marcosw set up. We can get to the machine and grab stuff until marcosw has things set up | 18:55.17 |
| ? | 18:55.18 |
Robin_Watts | henrys: I'm strongly of the opinion that none of us should venture anywhere near it until marcosw has it backed up. | 18:57.01 |
henrys | Robin_Watts: oh I thought he was going to do that as soon as he got back. | 18:58.06 |
Robin_Watts | henrys: He may have done, but he hasn't said he's done it. | 18:58.22 |
| If he's done that, then it's just a matter of copying Jogu's ssh key onto the NAS and he can hunt for stuff to his hearts content. | 18:59.10 |
henrys | bbiab | 19:04.53 |
Jogux | robin> that sounds awesome | 19:17.36 |
| I have actually got this build to work now, need to verify it against the previous release and check nothing looks overly broken... | 19:18.10 |
| [building ios release of smartoffice sdk for a customer] | 19:18.43 |
| am I meant to check the release spreadsheet back into CVS after I've changed it? | 19:28.19 |
Robin_Watts | Jogux: Into git? Yes. | 19:28.41 |
| You've added a new line at the bottom, right? | 19:28.56 |
Jogux | uh... it's not in git? | 19:30.00 |
Robin_Watts | Oh, duh. | 19:30.14 |
| Yes, then back into CVS would seem smart :) | 19:30.23 |
Jogux | cool | 19:32.17 |
| and, yes, newline at bottom (or at least, once I get that far) | 19:33.01 |
mvrhel_laptop | ok. the new gsview can now extract and save a small selected area on a page as ps, pdf, or eps. need to add svg | 20:37.39 |
ray_laptop | mvrhel_laptop: and save as text ? | 20:57.36 |
henrys | torn, Robin_Watts you guys still here miles wants to pull the trigger the mujs name or us come up with another name. | 20:57.40 |
| tor8 ^^^ | 20:57.55 |
| paulgardiner, mvrhel_laptop ^^^ | 20:58.11 |
mvrhel_laptop | ray_laptop: no text save yet. that will be a little different | 20:58.13 |
| henrys: I like mujs myself, but the domain price seems high to me | 20:58.52 |
henrys | mvrhel_laptop: seemed that way to me too, but I don't know what the domain market is like. | 21:01.34 |
mvrhel_laptop | henrys: I don't either | 21:01.47 |
ray_laptop | henrys: as long as it's not occamscript, and I think either keeping the 'mu' prefix, or something else that means small | 21:02.20 |
| my second choice would be mujscript but tor objected because MS made JScript something non-standard | 21:03.40 |
| mujavascript seems cumbersome, but it _is_ available as a .com domain | 21:05.21 |
henrys | ray_laptop: and how much does that cost, is the mujs price what we can expect to pay for anything similar? | 21:06.27 |
ray_laptop | and because it is such an undesirable name, google doesn't have any hits on it (it reverts to javascript hits) | 21:06.34 |
| note that mu-js is unique and mu-js.com is available | 21:07.51 |
henrys | ray_laptop: the only objection to mujs is price so if we aren't going to do much better price wise it doesn't matter | 21:08.57 |
mvrhel_laptop | are there objections to mu-js.com? | 21:09.41 |
ray_laptop | the mu-js.com is not taken | 21:09.53 |
| so it's just the normal registration cost | 21:10.12 |
| Miles controls the marketing budget :-) so whatever name he likes best (except occamscript which I don't think will fly with him, either) | 21:11.17 |
henrys | ray_laptop: oh okay I certainly don't object better let the the other "mooers" weigh in though | 21:13.08 |
ray_laptop | http://www.steps-to-make-your-own-website.com/hyphen-in-domain.html | 21:14.00 |
| henrys: Yes, I agree that others should weigh in | 21:14.19 |
| fortunately (or not) it's not hard to outweigh Tor ;-) | 21:14.59 |
| at least if we are using scales | 21:15.39 |
NoiseEee | just wondering if ghostscript will follow/parse symlinks? | 21:16.29 |
| in the following, i get "Could not open the file <pathToFile>" where the path involves a symlink: /usr/bin/gs <bunch of stuff> pathTofile | 21:17.33 |
ray_laptop | NoiseEee: no, it isn't a browser/viewer | 21:17.39 |
NoiseEee | i'd have thought everything would respect symlinks | 21:18.38 |
| i mean, the filesystem knows its a symlink? | 21:18.49 |
henrys | ray_laytop:he mean ln -s on uniz | 21:18.56 |
ray_laptop | but if the file is a symlink (or a hardlink) then, yes, GS can access the file -- assuming that the permissions are valid and you aren't running -dSAFER or -dPARANOIDSAFER | 21:18.57 |
NoiseEee | one of the directories in the filepath is a symlink... so it *should* be following it then, from what I understand? | 21:19.25 |
ray_laptop | and -dSAFER _might_ be OK | 21:19.30 |
NoiseEee | maybe it is a permissions issue, i'll check. thanks for the help, ghostscript is awesome | 21:19.50 |
ray_laptop | NoiseEee: that is up to the OS. gs just does an fopen on the name | 21:19.58 |
henrys | bbiab | 21:20.50 |
ray_laptop | NoiseEee: but -dSAFER requires that the path be specified on the command line (as -I___ or as the input file) or it won't be on the PermitFileReading list of paths | 21:21.58 |
NoiseEee | hmm not permissions. "**** Could not open the file /var/www/vhosts/xmeasures.com/fsheets/fsheetCheckout/xj5qymsddxpvsrsk7q94m/cs5sf52hx9jnqzyh_gs.pdf " where the "fsheetCheckout" is actually a symlink. I can see the file... i'm not using -dSAFER in the command | 21:22.20 |
| my bad ray_laptop, that was a permissions issue | 21:23.26 |
ray_laptop | NoiseEee: OK. Thanks | 21:23.48 |
NoiseEee | presumably writing should work the same / respect symlinks, or do I need to adjust an argument? | 21:24.21 |
| https://gist.githubusercontent.com/NoiseEee/75b205ac6e954ffa205b/raw/a9c87bf8f51e7515655fb7461a6d7ae558ab95ff/gistfile1.txt | 21:25.02 |
| the complaint now is: Error: /undefinedfilename in (/var/www/vhosts/xmeasures.com/fsheets/fsheetCheckout/xj5qymsddxpvsrsk7q94m/cs5sf52hx9jnqzyh.pdf) | 21:25.19 |
ray_laptop | NoiseEee: writing is pretty much the same. If the file is specified with -o ____ or -sOutputFile=___ it is added to the PermitFileWriting list of paths | 21:25.24 |
NoiseEee | (same idea, 'fsheetCheckout' is a symlink to a folder elsewhere) | 21:25.35 |
| the full error is https://gist.github.com/NoiseEee/8b04bf3825fb1cf355f1 | 21:28.00 |
| let me see about adding -o somewhere i suppose | 21:28.08 |
ray_laptop | NoiseEee: if permissions are OK, that should work. Try a simple C program: FILE *f = fopen("/var/www/vhosts/xmeasures.com/fsheets/fsheetCheckout/xj5qymsddxpvsrsk7q94m/cs5sf52hx9jnqzyh_gs.pdf", "r") ; printf("%p", f); | 21:28.33 |
NoiseEee | ray_laptop: can I just enter that at the linux prompt? sorry, not a linux master :P | 21:30.37 |
ray_laptop | NoiseEee: looks like it should work (unless there are permissions problems). The errno is what is used to give the message: Last OS error: No such file or directory | 21:31.04 |
| NoiseEee: no, you need the usual C boilerplate, need to put it in a file, and gcc it to get an executable | 21:31.49 |
| NoiseEee: I think you still have permission problems. | 21:32.23 |
| note that not just the file, but the directory has to be readable | 21:32.46 |
NoiseEee | right... i chmod'd "/var/www/vhosts/xmeasures.com/fsheets/fsheetCheckout/xj5qymsddxpvsrsk7q94m" to 777 for testing, but still no go | 21:33.01 |
| i'll see what i can figure out.... i appreciate the help ray_laptop | 21:33.13 |
ray_laptop | NoiseEee: try chmod -R 777 "/var/www/vhosts/xmeasures.com/fsheets/fsheetCheckout/xj5qymsddxpvsrsk7q94m" | 21:34.04 |
| NoiseEee: you made the directory readable, but not the contents | 21:34.28 |
NoiseEee | yeah, still no go. it seems to complain about the one to be written. | 21:35.32 |
| ie: before i had "could not open the file" (the one to be read) versus "no such file or directory' on the one to be written, /undefinedfilename | 21:36.32 |
| oh wait i think i have that mixed up | 21:36.57 |
ray_laptop | in the last post to gist.github, it was failing on the input (PDF) file. The output file was cs5sf52hx9jnqzyh_gs.pdf | 21:37.35 |
NoiseEee | yep. i'll mess around. thanks ray_laptop | 21:37.53 |
ray_laptop | NoiseEee: OK. I have to go now anyway. Good luck | 21:38.13 |
NoiseEee | thanks :) | 21:38.18 |
| message to future ray_laptop: all due to permissions. I appreciate the help! silly mistake. | 21:53.29 |
| Forward 1 day (to 2014/03/15)>>> | |