| <<<Back 1 day (to 2015/02/12) | 20150213 |
kens | Robin_Watts : Seq The PDF interpreter converts annotations into pdfmarks which it passes to pdfwrite to reconstruct the annotation, it doesn't render them. For annotations that the PDF interpreter understands. | 08:11.28 |
Robin_Watts | kens: So annotations like stamps will stay as annotations? | 08:40.38 |
kens | Robin_Watts : yes, provided the PDF interpreter knows how to convert them into pdfmarks | 08:41.44 |
| It knows about quite a lot of stuff these days | 08:41.56 |
| IIRC< if the original annotation lacks an appearance stream, running it through pdfwrite will produce an appearance stream in the pdfmark, so the resulting file will work on all PDF consumers (except Acrobat, which *always* ignores apperance streams and generates new ones) | 08:43.11 |
Robin_Watts | nice | 08:43.59 |
kens | This Cambodian text 'problem' turns out to be an example,the text is an annotation with no appearance stream. SO Their 'pdf form' application, Ghostscript and Acrobat all render the result differently. And the erporter thinks its a Ghostscritp bug. | 08:45.07 |
chrisl | kens: Presumably if you have the Cambodian language pack, Acrobat gets it as intended? | 08:49.14 |
kens | chrisl, no | 08:49.23 |
| The BBox is insufficient for the width of the text , which is what the reporter is seeing I believe. | 08:49.52 |
| But because we generate a slightly different appearance to Acrobat, its not quite the same (a different glyph gets truncated). However, neither Acrobat nor Ghostscript matches what he posted as the 'correct' result | 08:50.27 |
| I'm busy cutting the file down to a minimum to try and demonstrate | 08:50.48 |
| Of course, and 8 page acrobat form is not a great place to start | 08:51.07 |
chrisl | Oh, I hadn't realised he'd uploaded another test..... | 08:51.12 |
kens | Yeah :-( | 08:51.19 |
| If you open the file in Acrobat and look at the section '14' you may see what he means | 08:51.41 |
| There are (I think) 4 glyphs there, but the BBox is only wide enough for 3 when Acrobat draws it, and 2 when we draw it | 08:52.03 |
chrisl | Frankly, I'm finding it hard to care: if you can't upload the right test file, you go to the back of the queue..... and if the eventual test file is pointlessly big, you go way behind the back of the queue! | 08:53.20 |
kens | Thinking about it, I beleive Acrobat rewrites its generated appearance stream, tha'ts why it offers to 'save the changes'. So I'll upload that too. That will render the same in GS as Acrobat I expect | 08:53.28 |
| chrisl, I got this far figuring out what was wrong, so I figured I'd finish it up | 08:53.55 |
| Hmm, actually the annotations *do* have an apperance stream, but of course Acrobat ignores those :-( | 08:59.24 |
chrisl | kens: I guessing it *might* be okay if the font were embedded? | 09:03.51 |
kens | The font is embedded. | 09:03.59 |
| The problem is that the declared clip/bbox/width is insufficient for the full width of the text (I htink) | 09:04.16 |
| I'm having trouble getting the remaining text out atm | 09:05.07 |
chrisl | Oh, I thought we were substituting for KhmerOSSiemreap but it must be just the messages being jumbled together | 09:05.35 |
kens | Hmm, I didn't see a substitution message, just a minute | 09:05.50 |
| No, not getting a substitute with the cut down file | 09:06.18 |
chrisl | Using 9.10: | 09:07.55 |
| Can't find (or can't open) font file /usr/share/ghostscript/9.10/Resource/Font/KhmerOSSiemreap. | 09:08.00 |
| Can't find (or can't open) font file KhmerOSSiemreap. | 09:08.00 |
| Querying operating system for font files... | 09:08.01 |
| Can't find (or can't open) font file /usr/share/ghostscript/9.10/Resource/Font/KhmerOSSiemreap. | 09:08.01 |
| Can't find (or can't open) font file KhmerOSSiemreap. | 09:08.01 |
| Didn't find this font on the system! | 09:08.01 |
| Substituting font Courier for KhmerOSSiemreap. | 09:08.02 |
| Loading NimbusMonL-Regu font from /usr/share/fonts/type1/gsfonts/n022003l.pfb... 4085220 2563290 11106568 6761291 3 done. | 09:08.02 |
kens | Oh well I'm using current code.... | 09:08.16 |
| Why are you using 9.10 ? | 09:08.37 |
chrisl | I didn't have a build of master, and 9.10 is what's installed on this system | 09:09.09 |
kens | Fair enough | 09:09.16 |
| Mine's the current master and the font is OK for me | 09:09.31 |
chrisl | The executable I had is from my build consolidation project, and I'm not sure I'd trust that right now..... | 09:09.58 |
kens | Well he claims to be using 9.15 (though the report says master), and its OK for me, so.... | 09:10.53 |
chrisl | Hmm, yes, current code gives no warnings.... I don't understand that.... | 09:11.00 |
kens | I don't understand why changing the clip, bbox and rect makes no difference to what's rendered, even using Ghostscript :-( | 09:11.38 |
| Oh wait, I think the rect may be wrong, let me check the spec | 09:12.01 |
| Hmm, no, still no difference. | 09:13.47 |
kens | smacks head | 09:14.55 |
| Of course, this is a file I cut down using Acrobat, and Acrobat has added the apperance stream. Idiot! | 09:15.20 |
jo-tud | hi | 09:24.29 |
ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 09:24.29 |
jo-tud | I have a weird problem for which i was not able to find a solution by googling (a lot) | 09:25.11 |
chrisl | jo-tud: that's probably not enough information for us to help - if, indeed, we can help | 09:29.18 |
jo-tud | sorry | 09:33.09 |
| i was still thinking how to explain | 09:33.17 |
| here is an example file: https://tatul.owncube.com/public.php?service=files&t=e52fde8c4641e8128da695d29ffdb26f | 09:33.49 |
kens | Can you at least tell us which product this is about ? | 09:34.02 |
jo-tud | I used gs -dBATCH -dNOPAUSE -sDEVICE=pdfwrite -sOutputFile="min-pdfa.pdf" "min.pdf" | 09:34.21 |
kens | OK so Ghostscript, then. WHich version ? | 09:34.38 |
chrisl | I gotta go out for a while..... | 09:35.08 |
jo-tud | GPL Ghostscript 9.14 (2014-03-26) | 09:35.12 |
kens | And the problem is ? | 09:35.27 |
jo-tud | when pasting some text from the page letters from a-j are replaced by 0-9 | 09:35.41 |
kens | Well that's not entirely surprising | 09:36.02 |
jo-tud | there are some copy+paste problems that i found while searching - but never this | 09:36.06 |
kens | Did you try copy and paste from the original file ? | 09:36.58 |
jo-tud | it works | 09:37.03 |
kens | No it doesn't, I just tried it | 09:37.10 |
jo-tud | it works from the original file | 09:37.23 |
kens | If this isn't the original file, then what is it ? | 09:37.32 |
jo-tud | should I upload it, too? | 09:37.33 |
| ok, wait a second | 09:37.40 |
kens | This file is produced by pdftk, so where does Ghostscript come in ? | 09:37.50 |
| Actually produced by itext, modified by pdftk | 09:38.09 |
jo-tud | original: https://tatul.owncube.com/public.php?service=files&t=3424aac2dd32e7ad6f0527a00927b207 | 09:38.45 |
| the examples are page 50 from a large file to give an example. The behaviour is the same | 09:39.31 |
| p50 was extracted with pdftk | 09:39.45 |
| the original is produced by latex and works fine (copy-paste) | 09:40.07 |
| after processing with gs the number replacing occurs | 09:40.18 |
kens | Hmm, well there's a TopUnicode CMap, so that should work, and clearly doesn't. Your best bet is to open a bug report | 09:41.09 |
jo-tud | the gs prosess is done to make the pdf compatible with publication rules (-dpdfa) | 09:41.36 |
kens | Odd, the output file also contains a ToUnicode CMap | 09:42.08 |
| advice is the same though, open a bug report | 09:43.07 |
jo-tud | ok - I don't really know how to explain the problem though... I'll try | 09:43.33 |
kens | Your explanation her was perfectly fine | 09:43.47 |
jo-tud | ok | 09:43.57 |
kens | Attach the original file, and give the GS command line, then simply say that the output file doesn't copy/paste correctly | 09:44.14 |
jo-tud | ok, thanks | 09:45.05 |
kens | You're welcome | 09:45.17 |
jo-tud | http://bugs.ghostscript.com/show_bug.cgi?id=695834 | 10:10.52 |
kens | ok | 10:11.14 |
tor8 | Robin_Watts: why does fz_page_presentation take both a float *duration and return the duration in the struct? | 11:24.28 |
Robin_Watts | tor8: Can't remember offhand. Let me look. | 12:31.07 |
| duration is page duration. | 12:32.57 |
| the duration in the fz_transition struct is the transition duration. | 12:33.13 |
| So... show this page for 10 seconds, then fade over 1 second to the next page. | 12:33.29 |
tor8 | right | 13:17.00 |
henrys | with yesterday's new discoveries we are going after NVA again with the lawyer, trying to get a take down. | 14:23.45 |
kens | take down would be good | 14:24.12 |
henrys | I have mixed feelings about implementing NVA it could give the impression we support it, OTOH it undermines the business model ... | 14:24.32 |
kens | I don't think we should, it contaminates us in any argument, and I wouldn't really want to be complicit in the copyright theft. | 14:25.05 |
Digit_ | Anybody here can help me with question about a commercial license in a private chat? | 14:25.13 |
kens | If anyone asks we can point them to the available script to convert .nva back to .pdf | 14:25.25 |
henrys | kens: I lean that way too. | 14:25.33 |
Robin_Watts | Digit_: No one here can help with a commercial license. | 14:25.47 |
| But we can tell you who can :) | 14:25.55 |
Digit_ | only the email provided on the website? | 14:26.12 |
Robin_Watts | You need to mail scott.sackett at artifex.com (or sales at artifex.com) | 14:26.22 |
| Tell him as much about what you are doing as possible. | 14:26.42 |
kens | If you sent an email and didn't get a reply we can chase it up, or find out why (could be Scott is on the road) | 14:26.48 |
Robin_Watts | He'll get back to you with a long list of questions (many of which may not apply to you). Just answer as best you can. | 14:27.11 |
| He'll then get you a customised quote. | 14:27.19 |
| In general Scott is very good at replying to people. | 14:27.36 |
barq | malcolm | 16:38.17 |
Robin_Watts | barq: I fear I got it wrong. APV is by maciej, not malc. | 16:40.06 |
| I haven't seen maciej on here in a while. | 16:40.21 |
barq | hmm :'( | 16:41.18 |
Robin_Watts | But he should be emailable. | 16:44.11 |
| https://code.google.com/u/mpietrzak/ | 16:45.16 |
| Forward 1 day (to 2015/02/14)>>> | |