| <<<Back 1 day (to 2016/04/12) | 20160413 |
twohot | Greetings everyone | 00:13.44 |
| I'm running Fedora rawhide and just realized pdftk isn't there. How do I get the tool? | 00:14.13 |
rayjj | marcosw: RU around ? | 02:19.45 |
| or anyone else know of the best way to port a group of users and their .ssh directories from one machine (e.g. casper) to another ? | 02:20.44 |
| I could script it, but surely someone has this laying around | 02:21.20 |
| interesting that google doesn't really turn up any scripts, but marcosw surely must have done this | 02:25.37 |
mvrhel_laptop | got my laptop fixed. had to pull it almost completely apart | 03:01.57 |
| the whole usb connector had come off the board | 03:02.10 |
| lost a usb port now, but at least its not rattling around inside the laptop | 03:02.44 |
rayjj | mvrhel_laptop: that sounds BAD | 03:22.20 |
| mvrhel_laptop: usually the power connector is the first thing to fall off the m/b. Unfortunately, there's only one of them. | 03:23.15 |
| My dad has re-soldered at the power adapter on at least 15 laptops that people gave him. They usually tell him "It probably isn't worth fixing, so just keep it". He finds kids that can't afford a laptop that love them. | 03:24.44 |
| a couple of times he couldn't get the power connector to attach, so he just soldered a pigtail to an cable and connector that hung outside, but the kid that got it didn't mind a bit | 03:25.39 |
| mvrhel_laptop: let me guess, it was the USB slot you usually had your mouse receiver dongle plugged into. Bumping into that a few times can really strain things | 03:28.55 |
Robin_Watts | Morning tor8 | 09:35.42 |
| Various commits for you to look at. | 09:36.15 |
tor8 | Robin_Watts: morning. I got distracted by the bug report about FB2 | 09:37.40 |
Robin_Watts | ha. I wondered if you might. | 09:37.52 |
tor8 | turns out, we can already do FB2 if we just cook up a different default style sheet | 09:37.54 |
Robin_Watts | So... should I look at the fz_document as fz_image thing? | 09:41.54 |
| Or the fz_image as a displaylist thing at least. | 09:42.24 |
tor8 | fz_image as a display list would probably be better | 09:44.40 |
Robin_Watts | I think the latter is the key component of the former. | 09:46.19 |
tor8 | Robin_Watts: warning fix LGTM | 09:46.34 |
Robin_Watts | tor8: So... rc2 ? | 09:47.07 |
| pushed, thanks. | 09:47.16 |
tor8 | Robin_Watts: for text aa, that's not affected by compile time AA_BITS defines; that's just us calling freetype with different flags. | 09:47.25 |
| yes, rc2 time. | 09:47.50 |
Robin_Watts | tor8: Not quite. When text is too large for freetype, then we use outlines. | 09:48.09 |
tor8 | Ah. That's a crack the setting can fall through then. | 09:48.34 |
Robin_Watts | I believe my code does it right. | 09:49.07 |
tor8 | AA levels LGTM then. | 09:49.19 |
| though, text bits as done by freetype is either 8 or 0 | 09:49.58 |
Robin_Watts | Yes. | 10:00.05 |
| if > 0 ? 8 : 0 or something is what the code does. | 10:00.25 |
| or some freetype defines like that. "SMOOTH" maybe. | 10:00.48 |
tor8 | yes | 10:00.55 |
| the mupdf 1.9 RC2 is uploaded | 10:04.00 |
Robin_Watts | tor8: sanity check please... | 10:17.58 |
| pdf_dict_put(ctx, outlines, PDF_NAME_Count, pdf_new_int(ctx, doc, old_count > 0 ? nc : -nc)); | 10:18.06 |
| That's leaking the pdf_obj int. | 10:18.24 |
tor8 | yes. should be pdf_dict_put_drop | 10:18.42 |
Robin_Watts | Ah, that's why we have pdf...yes. | 10:18.46 |
| Will fix, tha. | 10:18.51 |
| ta | 10:18.53 |
| tor8: Should a mutool clean -dif renumber the file objects? | 10:42.41 |
| tor8: Sanity check please. http://ghostscript.com/~robin/wtf.pdf | 10:48.47 |
| Object 18 is a Flate decoded stream. | 10:49.30 |
| OK, ignore me. | 10:50.19 |
tor8 | Robin_Watts: okay :) | 10:52.30 |
bvx89 | Hi. Does anyone here had problems with CreationDate not being equal to xmp CreateDate (and likewise with ModDate/ModifyDate)? | 11:32.11 |
| I'm using gs compiled inside a docker container | 11:32.54 |
Robin_Watts | problems in what sense? | 11:33.10 |
bvx89 | When i perform a validate on the output PDF using pdfbox I get an error on those fields | 11:33.32 |
| I'm trying to create a valid PDFA/1-b | 11:33.44 |
Robin_Watts | bvx89: Go to bugs.ghostscript.com, open a bug, attach a source file, and give the exact gs command that you have to run to generate the file that's incorrect. Give full details of WHY it's incorrect. | 11:34.39 |
| But first... what version of gs are you running ? | 11:34.51 |
bvx89 | 9.06 | 11:34.56 |
| I assumed that this was the cause at first: http://bugs.ghostscript.com/show_bug.cgi?id=692887 | 11:35.11 |
| But I checked the source code, and found that it was fixed in the 9.06 source code | 11:35.32 |
Robin_Watts | 9.19 is out now. You should really try that. | 11:36.21 |
| Cos it's the first thing you'll be asked to do when you open a bug report, I imagine. | 11:36.46 |
bvx89 | Well, 9.06 is the last one that uses the GPL license | 11:37.54 |
Robin_Watts | bvx89: We don't care about that :) | 11:38.27 |
bvx89 | haha, okay | 11:38.47 |
| Well, I can give a shot at compiling 9.19 and see if that fixes the problem in validation | 11:39.05 |
Robin_Watts | If you're using gs in a way that the GPL allows, but the AGPL doesn't, then we *really* don't want to fix your problem in 9.06 :) | 11:39.26 |
bvx89 | Yeah, I understand | 11:42.52 |
| I'm being enforced by my employer to use GPL version | 11:43.15 |
Robin_Watts | Who is your employer? | 11:44.19 |
bvx89 | It's a Norwegian firm, that makes money from scanning and archiving documents for the state/municipalities | 11:46.10 |
| Don't wanna be more specific | 11:46.28 |
| Why is that relevant? | 11:49.35 |
Robin_Watts | Just interested. | 11:50.24 |
| If they are genuinely using gs in a way that the AGPL would not allow them to do, then they are using it against the spirit in which we release it as freeware. | 11:51.26 |
| They might be within their legal rights, but we're probably really uninterested in helping them. | 11:51.52 |
bvx89 | I'm talking to my team leader now about this | 11:54.41 |
| Do you have any estimates about pricing? | 11:57.40 |
kens | bvx89 : We're just engineers, you would have to speak to the commercial people in the US office. mail to sales@artifex dot com | 12:09.08 |
bvx89 | Okay, I'll pass on the information | 12:09.56 |
| Robin_Watts : We decided to abide by the AGPL | 12:43.18 |
| I'll try to compile 9.19 and see what the result is | 12:44.01 |
| Nope, same result. | 12:58.26 |
| Should I post a bugreport then? The same file and the same commands works on a different machine, so I'm thinking it might be dependencies or something that's prohibiting the date being set correctly | 12:59.21 |
Robin_Watts | Yes, that's probably best. | 12:59.43 |
kens | We can't comment without seeing the file, and you would need to be quite specific about the problem. Openeing a bug report is definely the way to go | 12:59.45 |
| Err, are you saying thast the same file processed through the same version of GS on a different machine yields different results ? | 13:00.20 |
bvx89 | Yes | 13:00.24 |
kens | I can't see how that can happen..... | 13:00.37 |
| How are you comiling Ghostscript ? | 13:00.56 |
bvx89 | ./configure && make && make install | 13:01.10 |
| I didn't compile the 9.19 version, only 9.06 | 13:01.25 |
kens | Well in that case there should be no dependencies (other than the compiler and C runtime of course), it uses the third party code we supply | 13:01.57 |
Robin_Watts | The same command, binary and input file on 2 different machines is giving different results ? | 13:02.40 |
kens | The XMP metadata is set from the document informatoin (which is where the document informaton is stored) in pdf_wrte_document_metadata in gdevpdfe.c I can't see how the two can be anythign other than the saem | 13:04.29 |
bvx89 | Well, not exactly 2 different machines | 13:04.32 |
| on is on my current machine, and the other is run from within docker | 13:04.48 |
| one* | 13:05.10 |
Robin_Watts | But it's the same command, binary and input file ? | 13:05.14 |
bvx89 | Correct | 13:05.30 |
kens | But, as I said, the two pieces of information are set from *one* value, the date and time stored in the document Info dict. So I can't see how they can be different | 13:05.44 |
bvx89 | I get the same error on creation/mod date for any file | 13:06.00 |
| I'm thinking this might have something to do with how date is parsed inside the stripped down docker image i use as a base | 13:06.31 |
Robin_Watts | bvx89: Open the bug report, attach the source file, give the command line, and the 2 different failed files. | 13:06.43 |
kens | It seems unlikely. I could understand that making the dat eincorrect, but the document Info and the XMP data are created from the same fate | 13:07.02 |
Robin_Watts | kens: Might it be something dim like the C runtime printing differently in different locales? | 13:07.51 |
bvx89 | Yes, I've been thinking about that too | 13:08.03 |
Robin_Watts | But... all of this is just speculation until you *open the bug. | 13:08.30 |
bvx89 | The only locale available (from $ locale -a) is C, C.UTF-8 and POSIX | 13:08.33 |
kens | Well, its possible I suppose, but I didn't think hte code used the C runt-time. In any event open a report and I'll look at it when I have a chance | 13:08.34 |
| We parse the PDF date/time information ourselves to build the XMP string, so I don't thnk it can be anything like the locale, but lets see the file anyway | 13:11.58 |
bvx89 | I'll try setting the locale first, will post bug report if that doesn't work | 13:13.01 |
Robin_Watts | tor8: Just found a bug in mutool clean. | 15:00.49 |
| Fixed on robin/master. | 15:00.55 |
| I've still got a case here when sanitize is corrupting stuff, but it's independent of the fix there. | 15:01.23 |
| I'm looking into that now. | 15:01.32 |
| tor8: And another fix. | 16:07.54 |
| Fixed version of the SVG commit on my master too (wouldn't compile on VS) | 16:39.48 |
mvrhel_laptop | lunch time | 19:30.48 |
| and page merging seems to be working nicely with mupdf now | 21:38.50 |
| tor8: please take a look at my branch (graft) when you get a chance | 21:39.15 |
| thanks | 21:39.17 |
| taking a break for a bit. bb shortly | 21:39.42 |
| Forward 1 day (to 2016/04/14)>>> | |