| | <<<Back 1 day (to 2020/06/03) | Fwd 1 day (to 2020/06/05)>>> | 20200604 |
| paulgardiner | ator: your fix for 700236 looks to do the job, thanks. | 13:19.35 |
| ator | paulgardiner: nice. is it needed for other annotation types too, or are only the Text annotations affected? | 13:30.03 |
| paulgardiner | Oh, I don't know. I'll get back to you on that. | 13:33.56 |
| ator | malc_: (as usual, again) when I get around to answering your questions you're offline! | 13:35.40 |
| sebras | ator: malc_: you are both here! talk before the meeting! :) | 14:06.55 |
| malc_ | sebras: thank you :) | 14:08.14 |
| | ator: https://github.com/moosotc/llpp/issues/135 (summary a Nix guy wants me to tag something) i'd rather wait for exec stack and msfntp outline solution, question is - is it going to happen soonish? (say in 2-3 days) | 14:09.47 |
| ator | what's the msfntp outline thing? | 14:15.20 |
| | oh, msfents. | 14:15.31 |
| malc_ | ator: microsoft pdf that i put to boblycat for you (i deleted it form the server so have forgotten the proper name) | 14:17.36 |
| ator | malc_: yeah. I've got a patch for that one. it needs review, but should go on master in a couple of days. | 14:19.17 |
| | the execstack one is trickier | 14:19.22 |
| | the default *should* be noexecstack, so I need to figure out why it's forcing execstack on | 14:19.41 |
| malc_ | ator: cool | 14:20.09 |
| | as for exec stack... yeah it can be tricky to figure out | 14:20.37 |
| ator | malc_: ah... it's the objcopy generated data files that trigger it. | 14:34.48 |
| | pax-utils and 'scanelf' to the rescue! | 14:34.59 |
| malc_ | cool! | 14:35.51 |
| malc_ | feels ashamed having pushed for objcopy for resource generation for years | 14:36.37 |
| | thats almost four fors in one bloody sentence... | 14:37.08 |
| ator | should've sad "for four years" :P | 14:38.50 |
| | malc_: there's a fix on tor/master if you want to try it out | 14:39.13 |
| malc_ | ator: it's been more than four :) and give a moment to try your branch | 14:39.53 |
| | ator: seems to work, thanks! (GNU make based build system mupdf utilizes forced me to nuke first, only then did it properly rebuilt the object files... sigh) | 14:45.44 |
| kens | quit | 15:48.52 |
| Kantor | hi, I have a pdf file which can not be modified and saved because some errors in the file | 16:43.52 |
| | I run mutool clean -d and I get the following errors | 16:44.15 |
| | error: expected 'R' keyword (2629 0 R) | 16:44.26 |
| | warning: PDF stream Length incorrect | 16:44.49 |
| | warning: cannot load object (2629 0 R) into cache | 16:45.23 |
| | how can I fix this? | 16:45.31 |
| ator | Kantor: make sure the file is complete. is it a truncated download perhaps? | 16:48.26 |
| Kantor | I have no idea how to check if it is complete or not. I ust downloaded from the Internet. It is a pdf book. | 16:49.26 |
| | I'm new to mupdf and also to pdf in general | 16:49.38 |
| | the file can be opened and read as other pdf's | 16:49.56 |
| | the problem comes only when you want to save or analyze the file | 16:50.14 |
| ator | Kantor: the error you describe looks like either a truncated or corrupted file. | 16:50.47 |
| sebras | Kantor: can you view the file without problems? every page of the file? | 16:51.09 |
| ator | Kantor: link the file or upload it somewhere, and I can take a look. | 16:51.43 |
| Kantor | yes, I'm using sumatrapdf , I scrolled quickly the pages and I get no errors. | 16:51.59 |
| | Also Adobe Acrobat opens and reads the file. | 16:52.14 |
| | but can't save ot | 16:52.20 |
| | *it | 16:52.22 |
| | I will try to re-download the pdf | 16:52.47 |
| | Unfortunately the file is bad even on the server | 16:55.09 |
| | (I checked the checksum) | 16:55.24 |
| | ator: it is possible to manually in a hex editor fix it? | 16:56.17 |
| ator | Kantor: I doubt it. You can try. | 16:57.54 |
| Kantor | how it is that you can view every page, but can't save the file? | 16:58.51 |
| ator | but if it can be fixed, mutool clean should do it. does mutool clean abort without saving the file, or does it just complain? | 16:58.52 |
| | there can be data in the file that is not immediately visible just by viewing the pages, like file attachments. | 16:59.35 |
| Kantor | ator: it saves the file but the resulting file can't be opened with any pdf software | 16:59.43 |
| ator | I can't really tell you any more without access to the file. | 17:00.07 |
| | everything so far is just me guessing based on a vague description of the problem :) | 17:00.23 |
| Kantor | could I share a link with the file? | 17:00.35 |
| | *may I | 17:01.02 |
| ator | as long as I don't have to register an account, sure. | 17:01.04 |
| Kantor | of course you don't need that, I have a google drive space | 17:01.41 |
| malc_ | Kantor: at least in the past sumatrapdf was using mupdf to handle pdf's, so the situation you are describing is pretty bizarre (not inconceivable but... weird...) | 17:04.07 |
| Kantor | https://drive.google.com/file/d/1dxxgEWgxghKC_bfK4OKS_u_UooFW6-p5/view?usp=sharing | 17:07.46 |
| sebras | malc_: I'm not sure how up to date the mupdf version in sumatrapdf is. | 17:08.47 |
| malc_ | sebras: that was precisely my point | 17:09.11 |
| ator | Kantor: I get warning: expected 'endobj' or 'stream' keyword (2674 0 R) just by opening that file | 17:09.13 |
| malc_ | a progression where life spelt 'pro' as 're', so to speak | 17:09.58 |
| Kantor | I did not got such warnings . . . not with sumatrapdf nor with mupdf GUI interface | 17:10.06 |
| ator | and looking at the file, there are some serious syntax errors in there | 17:10.21 |
| Kantor | in the pdf structure? | 17:10.40 |
| ator | object 1609 doesn't end with "endobj" it ends with "endobj1610" | 17:10.43 |
| | which means our repair code can't fix it | 17:11.02 |
| | and the object it "crashes" on has the same error (missing a space between "endobj" and the start of the next object) | 17:12.08 |
| | 2629 0 obj 648 endobj2630 | 17:12.28 |
| Kantor | A funny thing. The latest version of Adobe Acrobat pdf reader and editor even freezes and crashes if I want to open the structure view | 17:12.41 |
| | there are a lot of such errors? | 17:13.25 |
| ator | several, yes | 17:14.03 |
| | at least 3 | 17:14.31 |
| Kantor | can I myself fix that with a hex editor? just searching for those strings and renaming them? | 17:15.41 |
| ator | if you fix those three errors (search for endobj[0-9] and add a space after the endobj) then mutool clean can repair it | 17:15.46 |
| | you can do it in a regular text editor (as long as it doesn't mangle binary data) | 17:16.03 |
| | vim is fine | 17:16.07 |
| Kantor | I have notepad++ | 17:16.20 |
| | so should I search for endobj? | 17:20.15 |
| malc_ | wow... 9899:202x working draft is typeset in URW PalladioL... | 17:25.01 |
| malc_ | sbrains melts and explodes | 17:25.16 |
| Kantor | A generic pdf related question. I inspected more pdf file hex data. The last string is %%EOF | 17:38.35 |
| | But after that some pdfs have nothing, others have only a 0D, others a 0D 0A | 17:39.12 |
| | does a pdf need a windows/linux newline after EOF? | 17:39.41 |
| | <<<Back 1 day (to 2020/06/03) | Forward 1 day (to 2020/06/05)>>> | |
Log of #mupdf at irc.freenode.net.