| | <<<Back 1 day (to 2012/08/14) | 2012/08/15 |
| prince3176 | hi ala | 06:01.13 |
| | i need help in mupdf | 06:02.09 |
| | any body help me | 06:02.49 |
| sebras | tor8: an untested PNG patch over at sebras/master | 11:27.25 |
| aksr | some of my pdf files are displayed in some strange symbols, i mean the entire document. If I try them in acroread (acrobat), they are displayed correctly, what could be the fault? | 11:38.07 |
| | i could upload pdf for testing purposes | 11:38.57 |
| chrisl | aksr: My guess would be that fonts (or more probably CIDFonts) are not embedded in the file, so you're relying on whatever (CID)fonts happen to be available - which differs between interpreters | 11:39.56 |
| aksr | chrisl: you think it's font's fault? | 11:41.03 |
| | how then acroread displays them correctly? | 11:41.26 |
| | (i don't understand :)) | 11:41.42 |
| robin_watts_mac | aksr: A 'well formed' PDF would have all the font details within it. | 11:42.00 |
| aksr | yes..? | 11:42.11 |
| robin_watts_mac | It's probable that your PDF is not well-formed, and just mentions some fonts by name. | 11:42.25 |
| chrisl | aksr: If acroread happens to have a matching CIDFont installed, it will work as expected - this is the reason the spec *stringly* advises CIDFonts should be embedded | 11:42.32 |
| aksr | so, there is no way to fix this, right? | 11:42.51 |
| robin_watts_mac | Unless those EXACT fonts, with those exact names are found, the pdf viewer has to guess. | 11:43.02 |
| chrisl | aksr: what interpreter are you using? | 11:43.11 |
| robin_watts_mac | Acrobat and gs both use different guessing rules (they aren't well documented) | 11:43.19 |
| | For CID fonts it's even worse. | 11:43.35 |
| aksr | i tried to open it via acrobat and gs(view) | 11:43.46 |
| | via acrobat is fine | 11:43.52 |
| | second-not | 11:43.58 |
| chrisl | And what platform? | 11:44.01 |
| robin_watts_mac | The fix is to ensure that whoever makes your PDFs makes them properly. | 11:44.03 |
| aksr | linux | 11:44.05 |
| | it's a book i downloaded from the *net | 11:44.20 |
| chrisl | A legit download, or something less so? | 11:44.46 |
| aksr | If you mean, did I pay for it-no. | 11:45.11 |
| | :) | 11:45.13 |
| | Can I, somehow, fix this? | 11:45.43 |
| | I mean to modify the pdf somehow .. | 11:45.53 |
| chrisl | Okay, can you post it somewhere, as you suggested? | 11:46.08 |
| aksr | Of course. | 11:46.16 |
| | Just a second. | 11:47.34 |
| aksr | now waits ... | 11:53.22 |
| | chrisl: https://anonfiles.com/file/3791ca2d0beb1dff06e5b5ad4054fe0f | 11:53.54 |
| | good? | 11:53.58 |
| chrisl | getting it now....... | 11:54.16 |
| aksr | ;) | 11:54.18 |
| chrisl | aksr: hmm, not what I was expecting :-( | 11:56.00 |
| aksr | try it with acroread | 11:56.34 |
| kens | Lots of embedded type 1 subsets | 11:57.29 |
| | But Symbol si not embedded | 11:57.35 |
| chrisl | Yeh, but seems like lots of non-Symbol stuff is messed up | 11:58.35 |
| kens | So I see, that's fairly weird | 11:58.51 |
| | Page 2 is using BOPORM2 as teh font | 11:59.24 |
| chrisl | FWIW, poppler has similar output to ours..... | 12:00.57 |
| aksr | chrisl: yes. | 12:01.06 |
| kens | So does MuPDF | 12:01.23 |
| aksr | yes i tried gsview, mupdf, poppler etc. | 12:01.43 |
| | acro and foxit get it right | 12:02.05 |
| kens | THe text looks sane enough: | 12:02.27 |
| | 29.888 0 0 29.888 257.835 563.587 Tm | 12:02.28 |
| | (T) Tj | 12:02.28 |
| | 17.933 0 0 17.933 277.203 563.587 Tm | 12:02.28 |
| | (rigonometric) Tj | 12:02.28 |
| | 29.888 0 0 29.888 413.185 563.587 Tm | 12:02.28 |
| | (D) Tj | 12:02.28 |
| | 17.933 0 0 17.933 437.613 563.587 Tm | 12:02.29 |
| | (elights) Tj | 12:02.29 |
| | Hmm, the Encoding only supplied a Diffrences array, I wonder if we are using a different default encoding | 12:03.50 |
| chrisl | If we weren't applying the differences, we'd have a *lot* of files not working. | 12:04.37 |
| kens | No what I mean is the basic encoding to which the DIffrences are applied | 12:04.51 |
| | The ENcoding dictioanry only seems to specify a Differences, not an Encoding | 12:05.09 |
| chrisl | looks up PDFRM..... | 12:06.16 |
| kens | We seem to be 'out by one' for the text on page 2 | 12:06.18 |
| | T->U, r->s, i->h etc | 12:06.36 |
| chrisl | Oh, lord more symbolic vs non-symbolic nonsense :-( | 12:08.22 |
| kens | Yes, I was just reading that | 12:08.33 |
| | BUt we don't need to worry about it because hte font is built in | 12:08.55 |
| | and rtherefore we should be using the built-in fonts base encoding | 12:09.05 |
| | Want to bet Acrobat does 'something else ' ? | 12:09.23 |
| | Going to have to decipher the type 1 font stream | 12:09.41 |
| | Type 1C font.... | 12:10.38 |
| chrisl | Hmm, font "Uses the Adobe Standard Roman Character Set: True"..... | 12:11.20 |
| kens | Which is presumably StandardEncoding, which should be fine | 12:11.37 |
| chrisl | Unless, the font's encoding *isn't* StandardEncoding..... | 12:11.59 |
| kens | Have you used cffdisasm on it ? | 12:12.11 |
| chrisl | No | 12:12.18 |
| kens | OK I'm still pulling the binary stream out | 12:12.26 |
| | Umm I have it deciphered, I just can't remember how to use teh Encoding in a CFF gtont :-( | 12:14.38 |
| kens | heads for the type 2 spec | 12:15.06 |
| | Both Charset and Encoding are format 0 | 12:16.29 |
| | This does not look like a standard encoding to me | 12:17.28 |
| aksr | Are you still sure it's font's fault? | 12:17.37 |
| kens | Still looking, but it seems likely | 12:17.50 |
| | Or more accuratley Acrobat is not following the specification | 12:18.02 |
| | "/A is code 63" | 12:18.14 |
| | And hsould be 65, so that is out byb 1 | 12:18.46 |
| | 2* | 12:18.48 |
| | B = 65, C = 66, D = 67 etc | 12:19.11 |
| | All out by one, just as expected | 12:19.26 |
| | So the font is *not* using a standard encoding, but Acrobat is applying one. I very strongly suspect teh PDF creation tool has created a mangled subset font, tried it on Acrobat @oh look it works' and that's the problem. | 12:20.09 |
| | I'm going to try fixing some of the font and see what happens | 12:20.59 |
| | cffasm doesn't liek the font, so I'm giving up | 12:25.15 |
| | In my opinion Ghostscritp et al are following the spec, Acrobat isn't (its ignoring teh Encoding in the built-in font, when it is not specified by the PDF file, but a Differences array is present). Whether ww choose to emulate it is up to chrisl and alexcher | 12:27.43 |
| aksr | hm. | 12:28.24 |
| chrisl | It's not obvious to me what Acrobat is using - hacking our PDF interpreter to apply the Differences to StandardEncoding doesn't *seem* to work :-( | 12:29.03 |
| kens | It looks like STandardEncoding ought to work | 12:29.41 |
| | THe few pieces of text I've looked at seem to be 'ASCII' | 12:29.56 |
| robin_watts_mac | how come mupdf gets it right ? | 12:31.24 |
| kens | It doesn't | 12:31.33 |
| robin_watts_mac | Oh, ok. That's much easier to believe :) | 12:31.47 |
| kens | Of course, it depends what you mean by 'right' :-) | 12:32.11 |
| | It matches GS, not Acrobat | 12:32.22 |
| robin_watts_mac | matching gs is fine. Had we been matching acrobat, then that's always harder to explain. | 12:33.07 |
| kens | Hmm, no wtha's odd.... | 12:33.11 |
| | page 3 comes out 'correct' | 12:33.18 |
| robin_watts_mac | OK. Lunch over. Heading to the pool. ttyl. | 12:33.30 |
| kens | enjoy robin_watts_mac | 12:33.42 |
| robin_watts_mac | ta | 12:33.46 |
| kens | chrisl here's an interesting thing. | 12:36.58 |
| | If I add '/StandardEncoding' to the Encoding dict in object 586 0, Acrobat shows *exactly* the same problems we do | 12:37.32 |
| | On page 3 | 12:37.38 |
| | and on other pages too | 12:38.04 |
| chrisl | Yeh, I'm actually struggling to find where (if!) we get the encoding from a Type1C font in order to apply the differences to it | 12:38.37 |
| kens | THat might be the problem then, we aren't applying it and Acrobat is. Or something | 12:39.02 |
| | By adding /BaseEncoding /StandarEncoding to the PDF file ,I believe I am preventing Acrobat looking up the Encoding in the CFF font. | 12:39.37 |
| | THe modified file displays exactly the same using GS and MuPDF which suggests to me they are not applying the actual encoding | 12:41.04 |
| | (in the font) | 12:41.10 |
| chrisl | Yeh, let me debug our PDF interpreter a bit more...... | 12:41.34 |
| kens | Going to try modifying the Differences array | 12:41.35 |
| | Huh, made no difference to GS or Acrobat :-( | 12:45.26 |
| | I'll leave it with you chris | 12:45.34 |
| chrisl | Er, hmm, I just "properly hacked" our PDF interp to use StandardEncoding for these fonts, and got the "correct" output - color me confused..... | 12:45.57 |
| kens | I guess I'm puzzled then. | 12:46.14 |
| | Maybe I misread the Differences array | 12:46.28 |
| | Certainly, what you say is what I thought was going on | 12:46.46 |
| chrisl | Ho, *torrential* rain here..... | 12:47.18 |
| kens | Oh, bet that's heading my way :-( | 12:47.32 |
| chrisl | aksr: do you have other files like this, or is this the only problem? | 12:48.05 |
| aksr | I think I encountered one (or more) but I don't now their name chrisl | 12:48.43 |
| | I'll look for them. ;) | 12:48.51 |
| | As soon as I find you them, you'll be informed! | 12:49.10 |
| | s/now/know now | 12:49.53 |
| | chrisl: so, you succeeded or..? | 12:50.56 |
| chrisl | aksr: I can get that file to come out legibly - I can say with confidence what Acrobat is *not* following the spec, but I cannot say what Acrobat *is* doing, for sure | 12:52.06 |
| aksr | hm. | 12:52.32 |
| chrisl | aksr: what I was thinking was that if this was the only file with the problem, I can give you a fixed version of the file. | 12:53.42 |
| aksr | i would love that and if can show me what to fix for other ones? | 12:54.23 |
| chrisl | aksr: they way I "fixed" it was to use Ghostscript's pdfwrite device with a hack in the PDF interpreter...... | 12:55.23 |
| aksr | ok... | 12:56.08 |
| | that isn't a fast way, is it chrisl? | 12:57.52 |
| chrisl | It didn't take long - but you do lose some metadata from the file | 12:58.35 |
| aksr | chrisl: what then i need to do? | 13:03.08 |
| chrisl | aksr: do you know where the Postscript resources are stored for GS on your distribution? | 13:03.44 |
| aksr | yes? | 13:06.12 |
| chrisl | OKay, so I'll put together a patch for you to apply, give me a minute | 13:06.43 |
| aksr | chrisl: /usr/share/ghostscript/9.06/Resource/? | 13:10.20 |
| | here â? | 13:10.29 |
| chrisl | aksr: yes. What I *strongly* suggest is that you create a directory somewhere in your user area, and copy the contents of /usr/share/ghostscript/9.06/Resource/Init into that - it should *not* have the string "resource" or similar in the path | 13:11.45 |
| aksr | i planned that ;) | 13:12.12 |
| chrisl | You can then tell Ghostscript to search *your* directory (as well as the default) for its init files by giving it teh "-I<path>" option | 13:13.07 |
| aksr | ;) | 13:13.59 |
| chrisl | aksr: once you have your own set of GS init files, you can apply this patch: http://www.ghostscript.com/~chrisl/cff-encoding.patch | 13:14.19 |
| aksr | chrisl: with what should I apply this patch? | 13:15.54 |
| | diff patch | 13:16.08 |
| chrisl | patch | 13:16.18 |
| aksr | or manually :P | 13:16.19 |
| chrisl | has a better idea..... | 13:19.00 |
| | aksr: here's the whole file: http://www.ghostscript.com/~chrisl/pdf_font.ps | 13:20.04 |
| | No patching required | 13:20.12 |
| aksr | chrisl: ok, thank you. | 13:24.14 |
| | when/if i find other files, you'll be informed. ;) | 13:24.34 |
| chrisl | aksr: could you do me a favour and create a bug (attaching that PDF) at: http://bugs.ghostscript.com/ | 13:26.03 |
| aksr | chrisl: of course | 13:27.22 |
| | i'm not sure about pdf, though. | 13:27.34 |
| | since it's copyright | 13:27.42 |
| | you have it, right? | 13:27.49 |
| | keep it. | 13:27.55 |
| | do i must create an account? | 13:28.57 |
| chrisl | Yes, you need an account. | 13:29.08 |
| aksr | ok. | 13:29.12 |
| | chrisl: http://bugs.ghostscript.com/show_bug.cgi?id=693279 | 13:38.07 |
| | good? | 13:38.08 |
| chrisl | aksr: that's fine, sure. Now you'll get notification when I finalise a proper fix (I've reassigned it to me). | 13:39.23 |
| aksr | chrisl: good. | 13:39.35 |
| | ;) | 13:39.37 |
| | chrisl: just one more question .. | 13:40.35 |
| | something like this: | 13:40.39 |
| | gs -I~/Resource/ Eli_Maor-Trigonometric_Delights.pdf | 13:40.41 |
| chrisl | Well, I did specifically say *not* to have the string "resource" (or similar) in the path..... I used "gs_init/" | 13:41.40 |
| aksr | gs -I~/gs_init/ Eli_Maor-Trigonometric_Delights.pdf | 13:44.30 |
| chrisl | Hmm, should work - maybe leave off the trailing "/" | 13:45.12 |
| aksr | chrisl: well, it isn't working.. | 13:47.00 |
| | i copied pdf_font.ps to gs_init/Init | 13:47.24 |
| chrisl | Oh, and Init is a directory? | 13:47.39 |
| aksr | yes | 13:47.53 |
| | i use arch linux | 13:48.01 |
| | it's in off. repo | 13:48.12 |
| chrisl | Okay, so the parameter is -I~/gs_init/Init | 13:48.19 |
| aksr | chrisl: i pointed/copied the whole Resource directory | 13:49.48 |
| | it should point to the Init inside Resoruse ;) | 13:50.02 |
| | *Resource | 13:50.07 |
| chrisl | You need to specify the path to the Postscript files | 13:50.35 |
| aksr | is there a way to recompile it, to see it normally in other viewers? | 13:50.41 |
| chrisl | gs -I~/gs_init/Init -sDEVICE=pdfwrite -o out.pdf Eli_Maor-Trigonometric_Delights.pdf | 13:51.17 |
| | Should give you a file called out.pdf which views correctly in other viewers | 13:51.39 |
| | But that patch causes all kind of mayhem with out regression tests :-( | 13:53.07 |
| aksr | chrisl: ok, it works | 13:53.59 |
| | but reports many errors/warnings | 13:54.07 |
| chrisl | Hmm, I don't get any errors or warnings | 13:54.38 |
| aksr | is there a efficient way to edit metadata? | 13:54.40 |
| chrisl | No | 13:54.50 |
| aksr | ok | 13:56.18 |
| | chrisl: thank you very much! | 13:56.28 |
| chrisl | aksr: thanks for find yet another case of Adobe failing to follow its own spec! | 13:56.54 |
| aksr | :D | 13:57.19 |
| | i'll report everything worth reporting ;) | 13:57.36 |
| | bye | 13:57.41 |
| | chrisl: notice on page 22 | 14:04.14 |
| | triangle in acro | 14:04.19 |
| | there isn't any after conversion | 14:04.27 |
| | you can compare versions | 14:04.38 |
| chrisl | Well, assuming he means a pyramid - it is there when I convert it...... | 14:08.22 |
| sebras | tor8 Robin_Watts: I got word from zeniko that he desires to be patch attributed as zeniko and not Zeniko. just so you know. :) | 14:08.53 |
| kens | chrisl the only triangle I can see on page 22 is the diagram | 14:09.40 |
| chrisl | kens: me too - of course, does he mean "Page 22" or the 22nd page in the file? | 14:10.10 |
| kens | THey are the same :-) | 14:10.22 |
| chrisl | No, Page 22 of teh book is the 37th page in the PDF | 14:10.42 |
| kens | Or possibly he means the 22nd page in the pages array | 14:10.47 |
| | Which has a number of triangles | 14:11.29 |
| | TBH the acrobat display on page 7 (22nd page) does not look correct to me | 14:11.59 |
| | "means to find a number such taht 250 times equals 180." | 14:12.25 |
| | I'm sure those blanks shouldn't be blank | 14:12.38 |
| chrisl | Well, all the triangles I can find seem to be there..... | 14:13.52 |
| kens | I cna't really comment abou thte triangles since I don't have a 'converted' file. I was just noting a problem with the so called 'correct' Acrobat display | 14:14.26 |
| chrisl | Those seem to glyphs missing from the font (or encoding) which is also interesting..... | 14:15.32 |
| kens | Yeah, that's what I was thinking | 14:15.42 |
| chrisl | It's a p*ss poor excuse for a PDF, whatever way you look at it.... | 14:16.44 |
| kens | The font is I'm sure not what they intended | 14:17.42 |
| chrisl | I think the entire file is not what they intended - but it sorta, just about works in Acrobat if you don't look too closely, so good enough :-( | 14:22.29 |
| kens | I think they key point there is 'if you don't look too closely'.... | 14:22.52 |
| | I wonder if Acrobat only ifgnores the font Encoding if its a CFF font.... | 14:23.36 |
| chrisl | The patch I threw together only does it for Type 2 fonts | 14:24.30 |
| kens | Was just thinking aloud really | 14:25.11 |
| chrisl | kens: it would make (some?) sense if it treated CFF fonts as "opaque" entities.... | 14:26.00 |
| kens | Yeah, they don't have an easy ENcoding, you have to assemble it yourself from the Encoding adn Charset dicts | 14:26.25 |
| | I bet Acrbat thinks that's too much like hard work | 14:26.52 |
| chrisl | Well, I certainly don't intend to faff about making a PDF with a Type1 font to try it | 14:30.33 |
| kens | I'm inclined to agree | 14:30.59 |
| | Ditalking of specs, did you see that Office 13 will finally support the Open XML document standard ? | 14:31.20 |
| chrisl | Oh, I thought they'd done that a while back | 14:32.25 |
| kens | No, they supported the 'transitional' format, which includes all the old Word formats... | 14:33.10 |
| | Not only for writing, but also for reading. SO MS Office couldn't open files in teh format that MS forced through the standards process | 14:33.35 |
| chrisl | So is this XML format the default file format for Office 13, or a "save as"/export thing? | 14:35.09 |
| kens | I think its the default, but you can still save in old formats | 14:35.32 |
| chrisl | I was just wondering if it's likely to gain traction outside of MS.... | 14:36.53 |
| kens | I think OPenOffice/LibreOffice already support it :-) | 14:37.49 |
| chrisl | But they use the other XML document standard by default, don't they? | 14:38.57 |
| kens | chrisl I'm not sure about that, not my field | 14:39.16 |
| chrisl | It's not something I pay a lot of attention to.... | 14:40.41 |
| kens | Me neither, but I have more confidence in teh OpenOffice people supporting an ISO standard than MS :-) | 14:41.13 |
| | I do remember MS forcing the spec through the ratification process. | 14:41.39 |
| | Its only taken them 7 years to get to the point of actually adopting it | 14:41.52 |
| chrisl | Well, maybe they realised that XML doesn't fix every problem! | 14:42.21 |
| kens | sounds too sensible to me | 14:44.37 |
| henrys | chrisl, tor8:can we put up mupdf 1.1 yet? | 15:24.54 |
| | on the gs downloads page? | 15:25.06 |
| chrisl | henrys: to the best of my knowledge, it's not released yet | 15:26.05 |
| henrys | right the candidate is out but it's not released. I guess the question was more for tor8. | 15:28.08 |
| tor8 | henrys: candidate is out, a few important patches in the pipe. | 15:30.11 |
| | going to build and upload final release tomorrow if all goes well | 15:30.40 |
| henrys | okay great | 15:33.42 |
| paulgardiner | tor8: do you have my android commits for the 1.1 release, or is that not an issue? | 16:23.05 |
| tor8 | paulgardiner: I am planning on taking on the commits on paul/master for 1.1, yes | 16:37.51 |
| sebras | tor8: I can feel the pressure... | 17:28.52 |
| mvrhel_laptop | switched IRC clients. Chatzilla was giving me issues | 17:45.51 |
| | bbiaw | 18:07.33 |
| lesshaste | hi | 18:34.22 |
| ghostbot | hello | 18:34.22 |
| lesshaste | is it possible to invert the colours of a pdf using ghostscript? | 18:34.33 |
| | I really just mean swap black and white | 18:34.38 |
| alexcher | lesshaste: yes, there are many ways to do this. Please give more details. | 18:45.57 |
| | lesshaste: What's the type of your output file. | 18:47.23 |
| lesshaste | pdf | 18:49.58 |
| | alexcher: ^^ | 18:56.13 |
| alexcher | lesshaste: I'd recommend to redefine setgray operator to invert the color. | 18:56.27 |
| lesshaste | how does one do that? | 18:56.40 |
| alexcher | lesshaste: If you have images in the file, there's more work to do. | 18:57.11 |
| lesshaste | I don't | 18:57.32 |
| alexcher | lesshaste: let me do a few experiments. | 18:57.46 |
| lesshaste | thanks | 18:57.51 |
| alexcher | lesshaste: Is the source a PDF file ? | 18:58.42 |
| lesshaste | yes | 18:58.51 |
| alexcher | lesshaste: Do you want black background ? | 18:59.20 |
| lesshaste | yes | 18:59.32 |
| | alexcher: too hard? | 19:22.24 |
| alexcher | lesshaste: wait a little | 19:22.35 |
| lesshaste | k | 19:22.56 |
| alexcher | lesshaste: First, you need -dDELAYBIND option to enable modifications, e.g | 19:26.52 |
| | ~/ghostpdl/gs/debugbin/gs -dDELAYBIND -sDEVICE=pdfwrite -o OUT.pdf inv.ps INPUT.pdf | 19:27.20 |
| | inv.ps is a PS file that does redefinitions. | 19:27.48 |
| | /setgray {1 exch sub setgray} .bind def | 19:28.42 |
| | /old currentpagedevice /BeginPage get def | 19:28.43 |
| | /new {gsave 0 systemdict /setgray get exec clippath fill grestore} .bind def | 19:28.45 |
| | <</BeginPage { //old exec //new exec } .bind >> setpagedevice | 19:28.46 |
| | .bindnow | 19:28.48 |
| | lesshaste: that's all for now | 19:29.17 |
| lesshaste | oh | 19:29.50 |
| | thanks.. this needs a blog post as it's non-trivial! | 19:29.58 |
| | or to be put on a wiki somewhere | 19:30.29 |
| pipitas | alexcher: I tested your recent suggestion in this #channel to invert black/white for a PDF. I tried it like this: | 20:13.03 |
| | gs -dDELAYBIND -o inverted.pdf -sDEVICE=pdfwrite -c "/setgray {1 exch sub setgray} .bind def /old {currentpagedevice /BeginPage get} .bind def /new {gsave 0 systemdict /setgray get exec clippath fill grestore} .bind def <</BeginPage {//old exec //new exec} .bind>> setpagedevice" -f input.pdf | 20:13.17 |
| | It works⦠but it doesn't invert :-) | 20:13.41 |
| alexcher | pipitas: .bindnow is nissing | 20:14.47 |
| | pipitas: .bindnow is missing | 20:14.52 |
| pipitas | Thanks for the hint, alexcher! | 20:15.24 |
| alexcher | pipitas: besides, this hack only inverts colors set bt setgray . | 20:16.36 |
| lesshaste | pipitas: did it work for you? | 20:49.03 |
| pipitas | lesshaste: yes⦠provided the "colors" are defined by 'setgray'. I tested with a PDF created with this command: | 21:12.07 |
| | gs -o noninverted.pdf -sDEVICE=pdfwrite -g5950x700 -c "/Helvetica findfont 48 scalefont setfont 10 10 moveto .7 setgray (Helvetica) show showpage" | 21:12.11 |
| | lesshaste: I also tested with http://www.adobe.com/content/dam/Adobe/en/devnet/acrobat/pdfs/PDF32000_2008.pdf | 21:13.27 |
| | lesshaste: My command to invert was: | 21:14.07 |
| | gs -dDELAYBIND -o inverted.pdf -sDEVICE=pdfwrite -c "/setgray {1 exch sub setgray} .bind def /old {currentpagedevice /BeginPage get} .bind def /new {gsave 0 systemdict /setgray get exec clippath fill grestore} .bind def <</BeginPage {//old exec //new exec} .bind>> setpagedevice .bindnow" -f input.pdf | 21:14.22 |
| lesshaste | ok | 21:37.44 |
| sebras | paulgardiner: http://www.mass.gov/hed/undersecretary-barbara-anthony-statement-on-uber-technologies.pdf | 23:51.17 |
| | paulgardiner: this pdf consistently crashes the version of mupdf/android you sent me. | 23:51.41 |
| | E/AndroidRuntime(19902): Caused by: java.lang.ArithmeticException: divide by zero | 23:54.28 |
| | E/AndroidRuntime(19902): at com.artifex.mupdf.MuPDFActivity.createUI(MuPDFActivity.java:309) | 23:54.31 |
| | and there we have it. if core.countPages() returns 1 then smax will be 0 which causes a division by zero. | 23:54.59 |
| | tor8: just to be overly clear -- there is a slight problem with one of the patches over at paul/master. | 23:55.23 |
| | Forward 1 day (to 2012/08/16)>>> | |
IRC Logs