| <<<Back 1 day (to 2012/08/22) | 2012/08/23 |
sebras | believe it or not I need some help parsing output from gs... | 11:04.16 |
| http://pastebin.com/4g2UZQSK how do I read this to find out what the problem is? | 11:04.54 |
| this is from gs mynicedoc.ps | 11:05.01 |
chrisl | The postscript appears to be broken ;-) | 11:06.15 |
sebras | chrisl: ok. how? why? | 11:07.10 |
| since Vera is mentioned I looked inside the ps and found this, which I guess might be wrong in the spacing: | 11:07.30 |
chrisl | Well, from the error, "Vera" is undefined | 11:07.37 |
sebras | 0 0 0 nodecolor | 11:07.38 |
| 9 /Bitstream Vera Sans set_font | 11:07.38 |
| 489 603.4 moveto 52 (application) alignedtext | 11:07.38 |
chrisl | Yeh, names declared like that can't have spaces in them | 11:08.07 |
sebras | would it make sense to enclose it in parentheses? | 11:08.21 |
kens | (Vera Sans) cvn | 11:08.25 |
chrisl | If that really is the font name, then: (Bitstream Vera Sans) cvn | 11:08.48 |
sebras | looks up cvn. | 11:09.03 |
chrisl | "convert to name" | 11:09.11 |
sebras | aha! | 11:09.19 |
chrisl | It can be hard/impossible to decipher the actual problem with just the Postscript error - you usually need to see the error and the PS causing it to find the "real" problem | 11:11.02 |
sebras | that solved it perfectly. I might file a bug against graphviz for not using cvn... :-/ | 11:11.53 |
| thanks! | 11:12.01 |
chrisl | np, they probably assume that font names won't include spaces, hence the problem | 11:12.29 |
chrisl | ponders why we have to different ways of converting a string to utf-8.......... | 11:12.44 |
| kens: interesting new piece of UFST bizarreness...... | 11:14.38 |
| In UFST "Vertical Writing Mode" means "rotate through 90 degrees", and not what PS/PDF terms "vertical writing" | 11:15.27 |
Robin_Watts | chrisl: OK, I can regenerate the problem with 39fda0d on windows, but not with d527ead. | 11:36.49 |
| So I think it's fixed. | 11:36.58 |
| How were you regenerating the problem? | 11:37.05 |
chrisl | I used the command line the in the bug, 64 bit Linux - but I occurred to me after your post that I might have used an old exe | 11:39.23 |
kens | chrisl that's understandable, but weird :-) | 11:39.27 |
chrisl | kens: I can't see any reason for it, given that you have to pass in a matrix anyway | 11:40.34 |
Robin_Watts | mvrhel: Morning. | 11:44.08 |
paulgardiner | Robin_Watts: there's a fix for those segvs on paulg/forms | 11:45.46 |
Robin_Watts | 2 commits ? | 11:46.41 |
paulgardiner | The second is unrelated | 11:47.09 |
Robin_Watts | You've removed an fz_rethrow. | 11:48.02 |
| Should you have an fz_warn in there instead? | 11:48.09 |
paulgardiner | Not sure. I did consider it, but it would probably occur frequently and isn't necessarily the sign of anything problematic. | 11:49.37 |
Robin_Watts | Does return NULL from this routine indicate an error ? | 11:49.46 |
| or is that a valid response? | 11:50.02 |
| Cos if it can be a valid response it looks like we're silently folding an error case down onto a success case. | 11:50.21 |
paulgardiner | I believe it's an error in a file's javascript that can cause it. | 11:51.44 |
| Where's NULL turning into success? Is there a val?val:"" somewhere? | 11:53.04 |
| Hang on. I thought I was passing NULL back to v8 and letting the javascript that invoked this C call decide what to make of it. | 11:54.35 |
Robin_Watts | chrisl: Is that bugs reporter a commercial or support contract user? | 12:33.45 |
chrisl | Robin_Watts: I think he's just a GPL user - I think we refered him to Scott, but heard nothing more | 12:41.00 |
Robin_Watts | OK, so I feel letting him source his own build is appropriate. | 12:41.33 |
chrisl | Well, we're *really* trying to only build for Gemma and co now | 12:42.08 |
Robin_Watts | ok. Time for me to look at zenikos commit list. | 12:45.28 |
paulgardiner | Robin_Watts: just pushed an update, addressing that problem you identified. | 12:50.40 |
Robin_Watts | paulgardiner: You can do: if (!utf8) break; | 12:53.35 |
| (I deliberately tweaked the fz_try() macros so you can break out of them) | 12:53.59 |
paulgardiner | Ah ok. Useful to know... but then you know how much I hate breaks :-) | 12:54.15 |
Robin_Watts | Possibly we should be using fz_caught(ctx) rather than ctx->error->message directly ? | 12:55.51 |
| Hmm. Only pdf_form and pdf_js use ctx->error->message directly. | 12:56.45 |
paulgardiner | The one I keep wishing we had is fz_always_with_rethrow or something like that. | 12:56.53 |
| Oh. fz_caught already exists? | 12:57.16 |
Robin_Watts | paulgardiner: You mean fz_catch_and_release() ? | 12:57.20 |
| #define fz_catch_and_release(ctx) fz_catch(ctx) { fx_rethrow(ctx); } ? | 12:57.55 |
paulgardiner | Robin_Watts: hmmm, no I mean the case where I have something to do always but nothing to do except rethrow in the catch, but I still have to type the catch and rethrow. | 12:58.20 |
| catch and release would be handy too | 12:58.49 |
Robin_Watts | paulgardiner: Right. There is no way to do stuff after the always without having some syntax after the always. | 12:58.50 |
| limitations of macros etc. | 12:59.03 |
paulgardiner | But could we have an alternative to fz_always that rolls in the catch and release... or is that exactly what you are saying cannot be done. | 12:59.57 |
| ? | 13:00.00 |
| I may as well change to fz_caught in a separate commit as there are other places that need it. | 13:00.31 |
Robin_Watts | yes, sure. | 13:00.48 |
| Sorry: separate commit - sure. | 13:01.12 |
| alternative to fz_always that rolls in the catch and release - no. can't be done. | 13:01.25 |
paulgardiner | Right | 13:01.32 |
| Of course. I can see that now. | 13:02.17 |
Robin_Watts | The problem with fz_catch_and_release is that I've never convinced myself that any name I could come up with was clearer, and shorter than just typing it out in full :) | 13:03.33 |
paulgardiner | Yeah quite. :-) | 13:05.05 |
| fz_calease. Sounds like a Dothraki queen | 13:06.13 |
Robin_Watts | :) | 13:06.31 |
| so, those other commits, are they ready for pushing too ? | 13:10.32 |
paulgardiner | Yep | 13:10.40 |
marcosw | Robin_Watts: any insight why the cluster is stuck in a loop running the same three git commits over and over? | 13:12.13 |
Robin_Watts | It is ? | 13:12.36 |
| marcosw: I'm being called for lunch by SheWhoMustBeObeyed, so I'll look when I get back. | 13:13.35 |
marcosw | yeah, it has run a1cf28884ebfad593f70141cef681c4e5d994291 eb62043f520985f28ccaf265258ddb31cee79e1e and 2740bef445c51680d831ec40753436291f01760a four or five time each | 13:13.46 |
Robin_Watts | I'll have a look at the state of git when I get back then. Probably something stupid I've done. | 13:14.24 |
marcosw | no hurry, it's beeing going on for 6+ hours | 13:14.35 |
| Robin_Watts: it appears to be commit a1cf28884ebfad593f70141cef681c4e5d994291 | 13:16.14 |
kens | I thought it was doing that.... | 13:16.28 |
marcosw | if I run 'git log a1cf28884ebfad593f70141cef681c4e5d994291' it appears as a normal commit | 13:16.29 |
| but if I run "git log | grep a1cf28884ebfad593f70141cef681c4e5d994291" it doesn't appear in the list of commits | 13:17.07 |
Robin_Watts | Right. I have no idea how it managed to get onto a1cf288 | 13:46.28 |
| actually, maybe I do. | 13:48.13 |
| I think rays commit has been lost somehow - maybe alex did a force push ? | 13:48.45 |
| No, ray created a "HEAD" branch somehow. | 13:49.13 |
kens | That sounds like a bad thing | 13:49.36 |
paulgardiner | Could that be checkout without the -b option, commit and then push... not that I'm sure git will allow that. | 13:52.52 |
Robin_Watts | I suspect ray did a git push origin hash:HEAD | 13:54.33 |
| when he meant to do a: git push origin hash:master | 13:54.43 |
| should be sorted now. | 13:55.11 |
| Now I just need to reapply rays commit correctly. | 13:55.22 |
| oh, damn and blast. I broke the memento build. | 13:59.46 |
marcosw | Robin_Watts: would this be the correct time for me to suggest that our staff isn't smart enough to use git and we should revert to svn or cvs, I mean, really, what was wrong with cvs? :-) | 14:18.31 |
Robin_Watts | marcosw: no. | 14:18.57 |
| :) | 14:18.59 |
| Our staff is plenty smart enough. Evidence would suggest that they have been insufficiently motivated to read the docs though. :) | 14:19.43 |
marcosw | so "too lazy to use svn" then? | 14:20.07 |
Robin_Watts | although if ray fell foul of what I suspect he did, it was a simple typo, and could have happened to anyone. | 14:20.25 |
| sebras, tor8, paulgardiner: Various patches on my mupdf master branch. 6 so far, of which the first 5 are ready to go, I think. | 14:36.38 |
| actually, let me fix the whitespace issues. | 14:45.23 |
sebras | Robin_Watts: :D | 14:45.46 |
alexcher | I've got Unresolved external symbol _cmsChannelsOf on Windows | 14:49.27 |
Robin_Watts | alexcher: What sort of build? I'm not seeing that here. | 14:50.39 |
alexcher | nmake -f psi\msvc32.mak | 14:51.54 |
| I can debug if there's no oops moment. | 14:54.07 |
Robin_Watts | I'm doing a clean rebuild of gs now. | 14:54.24 |
paulgardiner | Robin_Watts: Checked through "Update Memento to match the version in gs" best I can, but have to admit to not understanding a lot. | 14:57.07 |
| Robin_Watts: Just checking, changing MEMENTO_LEAKONLY to be the default was intentional? | 14:57.47 |
| Robin_Watts: And MEMENTO_GS_HACKS has to be explicitly edited, rather than being picked up from the build? | 14:58.42 |
Robin_Watts | It's explicitly set true within the gs source, yes. | 14:59.12 |
paulgardiner | Robin_Watts: "Mupdfinfo: Cope with zero page pdf files" I understand that one. :-) Fine | 14:59.32 |
Robin_Watts | Too many makefiles for me to go through and add it explicitly in all of them with any degree of confidence. | 14:59.37 |
paulgardiner | Robin_Watts: "Silence some warnings" looks fine | 15:00.44 |
| Robin_Watts: "Rename fz_new_name to pdf_new_name" looks fine | 15:01.43 |
Robin_Watts | new versions (with fewer whitespace errors) pushed | 15:04.43 |
paulgardiner | Robin_Watts: "Make pdf resource localisation procedure non-static" looks fine | 15:07.47 |
mvrhel | good morning | 15:08.08 |
Robin_Watts | Morning | 15:09.12 |
| paulgardiner: Thanks for that. | 15:11.51 |
| paulgardiner: http://ghostscript.com/cgi-bin/clustermonitor.cgi?report=46aa12efced7578b466b2c49bf8e2c512ebf0b49&project=mujstest | 15:11.59 |
| 2 new SEGVs ? | 15:12.21 |
paulgardiner | Yep. I noticed that. I'll take a look | 15:12.24 |
Robin_Watts | mvrhel: Did you have any thresholding stuff left over? | 15:13.19 |
| If you'd like me to look at anything please just say,. | 15:13.42 |
mvrhel | Robin_Watts: I am waiting for Ray to commit and/or share his fix so I can check things out | 15:14.33 |
| getting ready to leave in an hour to go to Oregon though on vacation | 15:14.55 |
Robin_Watts | You're off on your bike ride, right? | 15:14.55 |
mvrhel | yes | 15:14.58 |
| should be fun | 15:15.29 |
| weather is looking great | 15:15.36 |
| bike is loaded down though | 15:15.40 |
| 6100 feet on day one (Friday) | 15:16.06 |
Robin_Watts | I was chatting to the physio this morning and she does triathlons. I mentioned that I'd seen a thing on the net about the fact that by far the largest demographic for amateur cyclists was men between the ages of 35-45. | 15:16.31 |
| And she said "Yes, they are called MAMILs". | 15:16.50 |
| Middle Aged Men In Lycra. | 15:16.57 |
mvrhel | ha | 15:17.00 |
Robin_Watts | That's a big climb. | 15:17.17 |
mvrhel | I have mt. biking shorts on. Lycra is on underneath my shorts... | 15:17.24 |
Robin_Watts | I know that Bradley Wiggins used to train by climbing from sea level to 7500 feet up a volcano each day. | 15:18.00 |
| yeah, I have a similar setup when I run. | 15:18.11 |
paulgardiner | Robin_Watts: might be better for tor8 to check over the editing one, but I guess I can if tor8's not available. What's the long term aim? Full editing, or is there some application I don't know about? | 15:18.23 |
Robin_Watts | (shorts, not 6100 feet ascents) | 15:18.30 |
| paulgardiner: The editing one is not ready to go yet, sorry. | 15:18.41 |
| It was something I was playing with on holiday. | 15:18.50 |
paulgardiner | Ah good. | 15:18.57 |
Robin_Watts | We were asked a couple of weeks ago if mupdf could support generation of PDF files. | 15:19.13 |
| and currently, it can't. But it would be nice if it could. | 15:19.27 |
| So the plan was to write an fz_device that would take the device calls and turn them back into streams of pdf operators which could then be put back into pages etc, | 15:20.18 |
| So you can now create a 'blank' pdf page, create a page, open a device on it, draw through that device, close the device, and then save the pdf file back. | 15:20.52 |
paulgardiner | Wow. Nice | 15:21.21 |
Robin_Watts | Currently it's sufficient for tiger.pdf to be rendered OK, but it'll ignore any bitmaps or text or transparency etc. | 15:21.30 |
alexcher | Please disregard my post about undefined _cmsChannelsOf on Windows. Full rebuild has fixed it. | 15:24.34 |
Robin_Watts | alexcher: Fab. | 15:24.41 |
| mvrhel: lcms v2.4 is out, but it's lost all the chameleonic header stuff for fast transforms :( | 15:25.17 |
| I may put that back in, and resubmit it to marti. | 15:25.31 |
mvrhel | ugh | 15:25.50 |
| Robin_Watts: sounds like a plan | 15:26.01 |
Robin_Watts | actually, maybe I mean 2.3 | 15:26.25 |
| ah, no 2.4 - it's in release candidate. | 15:32.49 |
henrys | mvrhel:can't wait to hear about this ride. | 15:40.42 |
mvrhel | henrys: frantically trying to get last minute things done and trying to figure out if I have everything I need. need to plan for just about every scenario | 15:41.27 |
henrys | I always make a checklist for that kind of thing. | 15:42.09 |
mvrhel | yes. I had one and I thought it was complete but then a few items were pointed out that I might want | 15:42.33 |
| like soap :) | 15:42.37 |
| by my wife | 15:43.00 |
Robin_Watts | Helen has a holiday checklist. It's about 3 pages long now. | 15:43.27 |
mvrhel | every pound is an issue here | 15:43.56 |
henrys | mvrhel:toothbrush? | 15:45.35 |
Robin_Watts | passport, credit card, bike. Addresses of 5 star hotels and taxi services along the route... can't think of anything else required... | 15:46.42 |
henrys | laptop to work on gemma's problems | 15:47.22 |
| ;-) | 15:47.35 |
Robin_Watts | Are there problems outstanding from Gemma? | 15:47.36 |
henrys | I don't think there are. | 15:48.00 |
| actually | 15:48.09 |
kens | she does | 15:48.13 |
Robin_Watts | next.pdf ? | 15:48.20 |
mvrhel | ha | 15:48.24 |
| yes just grabbed toothbursh | 15:48.30 |
henrys | tor8:can I send the list to zeniko or did you guys want to talk to him first? | 15:50.14 |
Robin_Watts | I haven't looked over the list yet (except very superficially). I'm happy to defer to tor8 though unless you want a double check done. | 15:51.06 |
tor8 | henrys: I can't think of anything else, did Robin_Watts want look through the list first? | 15:51.18 |
henrys | I don't really see the need to actually create bugs. | 15:51.46 |
Robin_Watts | ok. | 15:51.54 |
| So we're offering him 6 bounties effectively? | 15:52.17 |
tor8 | henrys: you may want to look through the commits and see if the scope is what we expect of bountiable bugs | 15:52.19 |
henrys | I did and it seems reasonable, the only thing I find a bit puzzling is I thought there would be more stuff than is in the list I guess you haven't pulled much from that large patch. | 15:53.56 |
Robin_Watts | henrys: The patch is, by definition, stuff that we haven't pulled. | 15:54.30 |
tor8 | henrys: most of the stuff from the large patch has been on the order of one-line diffs | 15:54.31 |
| the stuff we've taken on, that is | 15:54.41 |
| and the rest are from this list, or older (and by kowalczyk) | 15:54.57 |
henrys | okay I'm good, when Robin_Watts is okay we'll start the discussion. | 15:57.29 |
Robin_Watts | ok, I'll look through the lists now. | 15:57.44 |
henrys | no hurry | 15:57.54 |
| Robin_Watts:my understanding is we have taken stuff from the patch, as tor8 says. I think sebras has done a lot of that | 15:58.55 |
Robin_Watts | henrys: Right, but the moment we take stuff from the patch, the patch no longer shows what we've taken. | 15:59.39 |
henrys | tor8:we'll compensate him separately maybe he can give me an hours report? | 15:59.39 |
Robin_Watts | The patch is a diff between their code and ours; if we we take stuff, the patch no longer shows it. | 16:00.27 |
henrys | When I say "the patch" I mean http://software.zeniko.ch/sumatrapdf/SumatraMuPDF.patch | 16:02.35 |
Robin_Watts | Yes. | 16:02.42 |
| That's an auotgenerated diff between our code and theirs. | 16:02.56 |
henrys | oh I had no idea. | 16:03.17 |
Robin_Watts | (If it's not live autogenerated, it's updated whenever he runs a script or something) | 16:03.44 |
chrisl | 0.+ | 16:08.38 |
henrys | so strictly "the patch" should mean the url and the date of his email that requests money for the work for billing purposes. | 16:09.00 |
Robin_Watts | henrys: Oftentimes (in fact most of the time) we don't take his code *exactly*. We take something that's based upon his code. | 16:10.53 |
henrys | understood | 16:11.44 |
Robin_Watts | I don't know how you're planning to word the legalities, but it'd probably want to cover "code supplied or derived from patches supplied". | 16:12.02 |
henrys | he's getting paid for Tor's list and nothing else. | 16:12.53 |
Robin_Watts | There are things we have taken that aren't in tor's list. | 16:15.04 |
| many of them are 1 line fixes. | 16:15.13 |
| (and as such so small that once you've spotted the issue, the fix is obvious) | 16:15.32 |
| others are slightly bigger. | 16:15.42 |
henrys | well that was my concern above, can they be bundled? | 16:15.52 |
Robin_Watts | I'm making a list with that in mind now. | 16:16.50 |
henrys | okay | 16:17.42 |
raul_ | i have meny problems whit gs9.06 command pdfopt, may you help me, | 16:20.21 |
kens | OK I'm off, goodnight all, enjoy your bike ride mvrhel | 16:20.23 |
henrys | bye kens | 16:20.34 |
| raul_:the guy most likely to be of the most help has just exited kens, you can catch him GMT work hours. | 16:21.58 |
| but feel free to ask your question | 16:22.22 |
Robin_Watts | tor8: http://git.ghostscript.com/?p=mupdf.git;a=commitdiff;h=6e3e5bef0e262152ca71866e0059d0089b317a69 | 16:23.02 |
raul_ | -5 GMT Bogota Colombia | 16:23.04 |
Robin_Watts | second chunk of the patch; should that not be "bits not bytes" not "bytes not bits" ? | 16:23.20 |
raul_ | yes i has been aplyed patch, but i dont work, when i produced to pdf is 0 bytes | 16:24.21 |
henrys | what patch? | 16:25.07 |
mvrhel | see you all later | 16:25.41 |
Robin_Watts | have fun! | 16:25.46 |
mvrhel | thanks | 16:25.49 |
raul_ | one moment please i will to review | 16:26.00 |
henrys | raul_:normally we need an exact command line and a test file, the test file is not needed if it happens with any file. | 16:27.39 |
| tor8:btw with whom did you make your informal agreement or was it just assumed given the dependency of sumatra on mupdf? | 16:29.38 |
raul_ | i has been aplied the next patch http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=535d11e8 | 16:29.42 |
chrisl | raul_: that was *way* before 9.06 was released - that fix is already in 9.06 | 16:30.24 |
raul_ | i download pdfopt.ps and replace in my folder gs/gs9.06/lib/ | 16:30.35 |
Robin_Watts | henrys: AIUI, tor8 had an agreement with Kriswotsit. And we assumed that was fine until march this year, when we realised that zeniko had been doing most of the recent changes, and the agreement didn't cover his code. So we came to an informal agreement with him about his code whereby we could use his code, while he sorted something formal out with miles. | 16:31.45 |
| (That sentence seemed clearer while I was typing it, sorry) | 16:32.42 |
henrys | I think I understand it. | 16:33.52 |
| what was your informal agreement with zeniko, exactly? | 16:34.48 |
| this does change my understanding I thought an informal arrangement was reached with sumatra unbeknownst to zeniko | 16:36.13 |
Robin_Watts | "As for version 1.0: Until Miles gets the signed CLA back, you may proceed as if I'd already done so back when you started taking my patches." | 16:37.08 |
raul_ | my coomand line is c:\Program Files/gs/gs9.06/bin/gswin64c -q -dNODISPLAY -P- -dSAFER -dDELAYSAFER --pdfopt.ps c:\inputfile.pdf c:\outputfile.pdf | 16:37.13 |
| excuse me the / is \ in my command line | 16:37.43 |
Robin_Watts | The initial agreement was reached with Kryswotsit, who I think was the sole developer of Sumatra at the time. | 16:37.49 |
| Then zeniko came on board, and it never occurred to us that the situation had changed. | 16:38.44 |
| When zeniko mentioned it, tor looked back and saw that the agreement he had in place only covered K's code. | 16:39.33 |
raul_ | what is the patch for gs9.06 version? because i aplied http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=535d11e8 | 16:39.48 |
chrisl | raul_: 9.06 has that patch in the release | 16:40.14 |
raul_ | may you the link for download | 16:41.03 |
chrisl | You said your were using 9.06 already | 16:41.22 |
raul_ | yes | 16:41.34 |
chrisl | So, you must have downloaded it already, then...... | 16:41.56 |
Robin_Watts | So we discussed it with zeniko, and he said that he was pretty much happy to sign a copyright assignment (subject to the wording being OK), and that we could continue for 1.0 as if he'd signed such a thing. | 16:41.59 |
raul_ | i download the this link http://www.ghostscript.com/download/ | 16:42.15 |
chrisl | Yep, that's where you get the latest GS releases | 16:42.37 |
Robin_Watts | Now, I assumed that all this had been silently sorted out in the meantime, because I didn't hear anything back from either Miles or zeniko. | 16:42.57 |
| (last mail I have was on March 19th) | 16:43.11 |
raul_ | i download the this page http://www.ghostscript.com/download/gsdnld.html ghostscript GPL Release win64 | 16:43.30 |
henrys | Robin_Watts:my vote was to pull his work out and now I feel even stronger about it. But let's continue down this path. | 16:44.02 |
raul_ | but i dont understand what happen with pdfopt command because no work | 16:44.07 |
Robin_Watts | henrys: This is what I don't understand; what has changed between march and now that stopped us just getting a signed CLA from him? | 16:45.04 |
chrisl | raul_: well, the pdfopt.ps in the 9.06 release has that change in it. It may be you have a case that shows a bug in pdfopt.ps | 16:45.10 |
henrys | Robin_Watts:I don't know lost in Miles mail. | 16:45.36 |
raul_ | ok,, one moment please | 16:46.12 |
henrys | Robin_Watts:I think he got a bounty check for one of the jbig2 problems and the light bulb went on, but without the email difficult to say. | 16:46.32 |
Robin_Watts | Well, I'd hope that offering him bounties should smooth the problems away. | 16:46.56 |
henrys | I wonder what sort of hell windows 8 has in store for us? | 16:49.11 |
Robin_Watts | avoid avoid avoid avoid. | 16:49.29 |
raul_ | My command line is c:\Prorgram Files\gs\gs9.06\bin\gswin64c -q -dNODISPLAY -P- -dSAFER -dDELAYSAFER --pdfopt.ps c:\tt.pdf c:\outputfile.pdf, but the program no produce bug, the command finish whitout problems but the file (outpufile) is not generated, | 16:49.32 |
Robin_Watts | henrys: For desktop builds, no differences at all. | 16:49.42 |
| For TIFKAM builds, I had to disable the printer integration. | 16:50.12 |
| and the display device. | 16:50.17 |
| because all that is win32 specific. | 16:51.07 |
| But it means people can add that stuff back in themselves (or render direct to memory with their own devices etc). | 16:51.38 |
| So I suspect that actually it's no big deal for us. | 16:51.48 |
chrisl | raul_: An empty output file sure sounds like a bug to me | 16:51.57 |
Robin_Watts | The only people that it would affect are people that want to build their own TIFKAM apps that use gs - and they'll be doing the UI for that themselves, right? | 16:52.43 |
henrys | the big piece of work is mupdf on window RT | 16:56.18 |
raul_ | no, the output file no is produced in c:\, i dont understand, but when i put the command c:\Prorgram Files\gs\gs9.06\lib\pdfopt.ps c:\tt.pdf c:\outputfile the command produced a bug gswin64c no internal command ... but the output file is produced whit cero bytes | 16:56.26 |
Robin_Watts | henrys: That's 'just' another display app (we have a win32 one and a linux one). | 16:56.57 |
| but it would be silly to do a TIFKAM one now when the gtk based one is just around the corner. | 16:57.28 |
chrisl | raul_: I suggest opening a bug, and attaching a representative example of a non-working input PDF: http://bugs.ghostscript.com/ | 16:59.45 |
| it should then get investigated in due course..... | 17:00.08 |
raul_ | ok chrisl thank for you help, i will go to opening a bug | 17:01.00 |
henrys | Robin_Watts:just in time you have mail regarding the subject. I'm skeptical there won't be problems. | 17:01.16 |
raul_ | the last question, you do say that patch is inside gs9.06 release version? | 17:02.31 |
Robin_Watts | henrys: with WinRT? yeah, someone is bound to want us to do their work for them. | 17:02.58 |
| My feeling is that if someone is writing a winrt based app, then they are doing it specifically because they want to use the TIFKAM UI. | 17:03.52 |
| And the sole reason for doing that is so they can have their own 'style' of controls. | 17:04.17 |
chrisl | raul_: the patch was committed in May 2012, gs 9.06 was release in August 2012 - so the fix is in 9.06 | 17:04.37 |
Robin_Watts | (i.e. they want their own touch sensitive controls, or fade in/out, slide on/off effects etc) | 17:05.06 |
| and we can't possibly be expected to do that for everyone. | 17:05.20 |
chrisl | has to go..... | 17:05.43 |
Robin_Watts | The UI is what sets TIFKAM apps apart from one another - so it's not unreasonable to expect OEMs to do the UI coding. | 17:06.01 |
raul_ | may you help me whit the structure the command line for pdfopt in windows, for compaire with my command line | 17:07.58 |
henrys | raul_ it looks like you need a space after -- | 17:12.01 |
raul_ | ok i will go to the practice | 17:12.45 |
henrys | you want '-- pdfopt.ps' not '--pdfopt.ps' | 17:16.09 |
raul_ | ahh ok, in this moment i will go to make the change | 17:16.51 |
| please moment | 17:17.01 |
henrys | windows 8 also has some bad news for us, it finally has a native pdf viewer. | 17:22.41 |
ray_work | henrys: why is that bad news ? AR has been free from the get go and has been on Windoze | 17:23.21 |
raul_ | this is my command but after execution dont work look this C:\Program Files\gs\gs9.06\bin>gswin64c -dNODISPLAY - -pdfopt.ps c:\tt.pdf c:\t.pdf GPL Ghostscript 9.06 (2012-08-08) Copyright (C) 2012 Artifex Software, Inc. Allrights reserved. This software comes with NO WARRANTY: see the file PUBLIC for details. | 17:23.32 |
ray_work | raul_: as henry said, you need " -- pdfopt.ps " not "- -pdfopt.ps" | 17:24.35 |
Robin_Watts | raul_: Not "- -pdfopt.ps" "-- pdfopt.ps" | 17:24.38 |
raul_ | commadn C:\Program Files\gs\gs9.06\bin>gswin64c -dNODISPLAY - -pdfopt.ps result GPL Ghostscript 9.06 (2012-08-08) Copyright (C) 2012 Artifex Software, Inc. Allrights reserved. This software comes with NO WARRANTY: see the file PUBLIC for details. but the output file no is produced | 17:24.39 |
| ok, please moment, excuse me | 17:25.26 |
| try again | 17:25.35 |
ray_work | raul_: in doc/Use.htm it describes the options. '--' is an option | 17:25.53 |
henrys | ray_work:I think if folks have to go out in the world and make a choice they will consider us, if you have something on the box you may not consider shopping about. I think many folks use explorer even though chrome is better for example. | 17:25.57 |
ray_work | henrys: it's not like ANYBODY goes to mupdf (sumatrapdf) or ghostscript on windoze. If they use something other than Adobe, it would probably be Foxit | 17:27.17 |
raul_ | ok gays my command produce a bug check C:\Program Files\gs\gs9.06\bin>gswin64c -dNODISPLAY -- pdfopt.ps c:\tt.pdf c:\t.pdf | 17:27.50 |
| the results is a bug | 17:27.59 |
ray_work | cust 532 is upgrading to 9.06 :-) | 17:28.41 |
| Steve had asked me about it a couple of weeks ago, and Len is calling me with some questions as HE is doing the merge | 17:29.30 |
raul_ | Read objects, t = 11 Replaced references, t = 17 433 objects total Propagated attributes, t = 18 [[2]] Identified Catalog, t = 18 Error: /typecheck in --.knownget-- | 17:29.33 |
| Operand stack: --nostringval-- FirstPageNs 6 6 Contents | 17:30.03 |
henrys | ray_work:I don't know what you mean, lots of people use sumatrapdf. | 17:30.07 |
Robin_Watts | raul_: OK. At this point, update the bug with everything you know. | 17:30.07 |
raul_ | ok | 17:30.30 |
Robin_Watts | Then hopefully kens will be able to help you tomorrow. | 17:30.33 |
raul_ | the full error. line to line | 17:31.21 |
| GPL Ghostscript 9.06 (2012-08-08) | 17:31.23 |
| Copyright (C) 2012 Artifex Software, Inc. All rights reserved. | 17:31.31 |
| This software comes with NO WARRANTY: see the file PUBLIC for details. | 17:31.37 |
Robin_Watts | raul_: NOT HERE. On the bug. | 17:31.42 |
raul_ | a ok, excuse me | 17:31.54 |
| ok in this moment i posted the bug, thanks guys | 17:42.24 |
Robin_Watts | raul_: Can you attach your tt.pdf to the bug please? | 17:43.27 |
raul_ | of course | 17:43.49 |
Robin_Watts | The aim here is to give people enough information that they can reproduce your problem and then tell you how to solve it. | 17:43.52 |
raul_ | let me one moment | 17:44.01 |
| Robin the tt.pdf has 6.9 mb is possible upload this file? | 17:45.05 |
Robin_Watts | raul_: Yes. Use the "Add an attachment" button. | 17:45.49 |
raul_ | ok | 17:46.21 |
alexcher | Robin_Watts: I must be missing something. "-c foo" is a single argument "-c" "foo" are 2 arguments. | 17:46.23 |
raul_ | the file is upload now | 17:46.39 |
Robin_Watts | alexcher: With the winRT email? | 17:47.32 |
| That may be exactly the problem. | 17:47.44 |
| Without seeing his calling code it's hard to see exactly what he's doing. | 17:48.18 |
alexcher | Robin_Watts: yes, gs doesn't parse data that follow -c in the same argument on any platform. | 17:48.28 |
raul_ | Robin the file is already in my post bug, can you see it | 17:50.42 |
ray_laptop | chrisl_away: (and others) Does anyone know why we have BOTH lcms and lcms2 in the distro ??? | 17:51.07 |
Robin_Watts | I can. At this point you'll need to wait for kens tomorrow. | 17:51.14 |
| ray_laptop: Because you can configure which CMS to use. | 17:51.26 |
| WHICH_CMS="lcms" or WHICH_CMS="lcms2" or even WHICH_CMS="myowncms" | 17:51.50 |
ray_laptop | but why would anyone use lcms | 17:52.07 |
Robin_Watts | Because for a while lcms coped with stuff that lcms2 didn't. | 17:52.23 |
| now, lcms2 is better all around, and we could strip lcms out, but it does serve as an example to show that we are independent of the exact cms we use. | 17:53.01 |
raul_ | ok thank you very much | 17:54.27 |
Robin_Watts | tor8, henrys: http://ghostscript.com/~robin/zeniko2.txt | 17:56.56 |
| actually, ignore the XPS png fixes one. | 17:58.53 |
raul_ | ohh Robin, i has been the test with the same command with another file and results is perfect, the tt,pdf has a problem. | 17:59.53 |
Robin_Watts | Right. | 18:00.08 |
| It may be that tt.pdf is fine, it's just upsetting our pdfopt.ps script. If so, then that's a bug in pdfopt.ps, and kens would need to look at that. | 18:00.39 |
| ray_laptop: Something strange happened with your git commit earlier to fix my = {}; stuff. | 18:01.38 |
| it got pushed to the git server on the "HEAD" branch, rather than on "master". | 18:01.59 |
| What command did you use to push it please? | 18:02.20 |
henrys | Robin_Watts:so we'll accumulate "too small to mention" into 1 bounty and now we have 8. tor8 are you good? | 18:03.39 |
Robin_Watts | henrys: Oh, right. I was intending to ignore them :) | 18:03.56 |
alexcher | henrys: One more bug has been recently filed against pdfopt. Do we still need to maintain it? | 18:16.13 |
ray_laptop | alexcher: is your thought that we would make folks move to mupdf for that ? | 18:36.19 |
henrys | I was away sorry. | 18:39.07 |
ray_laptop | Robin_Watts: that's strange. I wonder why | 18:39.53 |
henrys | I think it should be maintained | 18:40.11 |
alexcher | ray_laptop: the script is (1) broken, (2) ineffective. The files are not recognized as linearized. | 18:40.11 |
ray_laptop | Robin_Watts: I didn't know HEAD was a branch, I thought it was a name for a version. | 18:40.40 |
henrys | alexcher let's talk to kens about it tomorrow | 18:40.50 |
Robin_Watts | indeed. | 18:40.51 |
ray_laptop | henrys: was that to answer alexcher ? | 18:40.52 |
henrys | ray_laptop:yes | 18:41.04 |
Robin_Watts | ray_laptop: That's why I was asking what command you used to push it. | 18:41.24 |
ray_laptop | I didn't think that kens had linearized working yet | 18:41.29 |
Robin_Watts | Did you perhaps checkout to HEAD rather than master? | 18:41.32 |
henrys | alexcher:have you verfied mupdf works for the users file? If so it might be worthwhile to suggest it. | 18:42.08 |
ray_laptop | Robin_Watts: I never use "HEAD" | 18:42.19 |
| I use 'master' and 'origin/master' | 18:42.39 |
alexcher | henrys: not yet. | 18:42.47 |
ray_laptop | or when I push I use push origin | 18:42.56 |
| Robin_Watts: my command history shows I used: git push origin/master | 18:44.13 |
Robin_Watts | and the msys git command prompt says (master) ? | 18:44.43 |
ray_laptop | Robin_Watts: my git logg shows: | 18:46.22 |
| * a1cf288 (origin/master, origin/HEAD, Fix-broken-Windows-build-fb55f251) Fix broken Windows build with patch fb55f251. | 18:46.24 |
| | * 7ed8765 (HEAD, master) Fix broken Windows build with patch fb55f251. | 18:46.25 |
henrys | bbiab | 18:49.08 |
Robin_Watts | right. There should be no such thing as origin/HEAD | 18:49.16 |
ray_laptop | Robin_Watts: how do I get rid of it ? (and why is it there ?) | 18:50.08 |
Robin_Watts | I'd imagine it's there because you used HEAD instead of master by accident at some point. | 18:50.31 |
| git branch -D origin/HEAD | 18:50.40 |
ray_laptop | Robin_Watts: I checked my 'history' and there is no "HEAD" there | 18:51.21 |
Robin_Watts | actually, ignore me. | 18:52.42 |
| origin/HEAD may not be wrong. | 18:52.49 |
ray_laptop | error: branch 'origin/HEAD' not found | 18:52.59 |
Robin_Watts | Professor google says it can be used by remote bare repos to indicate what should be checked out by default on clones. | 18:53.15 |
| so that's fine. | 18:53.17 |
| Anyway, I have no idea how it happened. Let's just ignore it and pretend it didn't happen. | 18:53.38 |
ray_laptop | Robin_Watts: so is my commit in or not ? | 18:53.44 |
Robin_Watts | ray_laptop: I committed a version of it. | 18:54.00 |
| so your exact commit is not in. | 18:54.07 |
ray_laptop | strange. Wonder how it got misplaced | 18:54.25 |
Robin_Watts | Thanks for fixing my mistake though :) | 18:54.28 |
ray_laptop | Robin_Watts: np. I just happened to need to build on Windows ;-) | 18:54.45 |
| I often rely on my clusterpush to spot any linux compile issues -- but we don't build windows with a clusterpush (like we should) | 18:55.52 |
| Robin_Watts: somehow my git is messed up. I did a git reset --hard origin/master and all are now on a chain inset from the main tree that branches in at "1e53dc6 More PACIFY_VALGRIND magic" | 19:19.49 |
Robin_Watts | What is your current SHA ? | 19:20.33 |
| The latest should be 0279e3b | 19:20.49 |
ray_laptop | Robin_Watts: the top of the logg looks like: | 19:20.59 |
| * 735fc4d (Fix-clist-copy-planes) Fix clist probelms with copy_planes, writer and reader out of sync. | 19:21.01 |
| * a1cf288 (Fix-broken-Windows-build-fb55f251) Fix broken Windows build with patch fb55f251. | 19:21.02 |
| | * 0279e3b (HEAD, origin/master, origin/HEAD, master) Bug 693298; Memento build tweak for memset. | 19:21.04 |
| | * 2dba4f4 Fix broken Windows build with patch fb55f251. | 19:21.05 |
Robin_Watts | Right, so that's fine. | 19:21.18 |
ray_laptop | Robin_Watts: the leading '| ' is OK ? (I don't recall seeing that) | 19:21.54 |
Robin_Watts | The lines are listed in order of time. | 19:22.09 |
| 735fc4d was obviously committed more recently than the others. | 19:22.26 |
| (I think). | 19:22.32 |
| The leading | is purely for the chart. | 19:22.38 |
ray_laptop | Robin_Watts: those other two were NOT more recent | 19:22.56 |
Robin_Watts | actually, yes, you're right. it's not ordered by time. | 19:23.07 |
| but it's just laid out that way by the algorithm that determines the ascii art to use. | 19:23.27 |
ray_laptop | Robin_Watts: I guess so. I just did a git am with a patch I wanted and it changed to: | 19:28.25 |
| * bb15658 (HEAD, master) Fix clist probelms with copy_planes, writer and reader out of sync. | 19:28.27 |
| * 0279e3b (origin/master, origin/HEAD) Bug 693298; Memento build tweak for memset. | 19:28.29 |
| * 2dba4f4 Fix broken Windows build with patch fb55f251. | 19:28.30 |
| I misread the clusterpush results and thought I still had 2407 differences, but those were from the previous clusterpush that included mvrhel's patch "Enabling of fast threshold code for color source images as well as monochrome images to cmyk devices" | 19:31.35 |
| so he'll have those differences to check out (if he hasn't already). At least it now runs his problem file without the FPE or a SEGV | 19:32.17 |
| when mvrhel gets back remind me to ask if he's OK with removing lcms | 19:33.50 |
tor8 | Robin_Watts: henrys: robin's list looks good to me | 20:28.56 |
| Forward 1 day (to 2012/08/24)>>> | |