| | <<<Back 1 day (to 2012/12/27) | 2012/12/28 |
| jzmer | what is the recommended windows color printer driver to be used to produce postscript files so that it can be processed with ghostscript? | 02:20.53 |
| ray_work | jzmer: The "generic PostScript color printer" but there is a .inf file in the gs distribution that has what you need. | 02:33.48 |
| | jzmer: search for ghostpdf.inf | 02:34.43 |
| jzmer | ray_work: is it signed with a cat? | 02:36.14 |
| ray_work | jzmer: I'll check further when I get back to my laprop | 02:36.16 |
| | laptop | 02:36.25 |
| jzmer | thanks. i will look it up | 02:36.29 |
| ray_work | jzmer: we don't do MS signed stuff | 02:36.43 |
| | we distribute the source with a MD5 sum published so you can verify your down.oad | 02:37.12 |
| | bbiaw | 02:37.27 |
| jzmer | but windoze would complain if a non-signed drive is installed, and in this case, the inf file | 02:37.57 |
| | does it mean i will have so sign it myself then ... | 02:38.11 |
| | s/so/to | 02:38.17 |
| | anyway to extract files from gs906w64.exe without running it as administrator? | 02:42.33 |
| | wait, generic postscript color printer? | 02:52.16 |
| | under which vendor? | 02:52.34 |
| | can ghostscript provide zip files for all windows flavors? | 03:30.36 |
| | exe need admin privelege | 03:30.46 |
| JakeSays | does gs have a debug mode that will trace the execution of the interpreter? | 04:03.41 |
| bencc | how can I convert series of images to a single pdf, each image as one page? | 12:02.44 |
| | I see viewjpeg.ps is there something similar for png? | 12:21.59 |
| jzmer | graphicsmagick should be able to convert image to pdf | 12:58.16 |
| | and then you concat them pdf into one | 13:04.29 |
| | gm/im should be capable as well | 13:04.47 |
| robin_watts_mac | JakeSays: Yes, various debug flags. | 13:25.27 |
| | In a debug build of gs, call: gs --debug and it will give you a list of switches. | 13:25.54 |
| | There are docs in gs/doc/Use.htm I believe. | 13:26.08 |
| bencc | jzmer: graphicsmagick is like imagemagick? | 15:08.30 |
| jzmer | aye | 15:10.31 |
| JakeSays | Robin_Watts: yeah found 'em - thanks | 15:26.49 |
| | robin_watts_mac: i believe i've tracked down the issue i've been having with txtwrite and invalid characters. i think the pdfs i'm dealing with have embedded fonts and txtwrite can't resolve them correctly. however other drivers work fine - for instance png*. can you give me an idea where i can find the code png* uses to render text? | 15:33.40 |
| robin_watts_mac | JakeSays: The problem is that lots of PDFs are created with enough information to render correctly, but with not enough information to actually give a valid unicode char for each glyph. | 17:22.03 |
| | I can make a PDF with an embedded font, and tell it to draw glyphs 1,2,3,4,5. Now, those glyphs might happen to spell out 'fred!', but unless I somehow note the unicode value that each glyph number corresponds to, the file doesn't have that info in it. | 17:25.23 |
| JakeSays | robin_watts_mac: yeah thats exactly what these pdfs are doing :( | 17:57.03 |
| | robin_watts_mac: the thing is, though, the text writer driver can't even resolve the glyph. | 17:58.10 |
| | i'm half suspecting it might be easier to just OCR the whole thing from a png | 17:59.04 |
| | robin_watts_mac: actually its the function gs_font_map_glyph_to_unicode that is failing | 18:10.59 |
| | which makes sense | 18:11.33 |
| sebras | JakeSays: are you able to copy the text you want using another PDF viewer? | 18:11.46 |
| JakeSays | because i imagine the actual unicode value is useless for rendering | 18:11.52 |
| sebras | if that works then it might be the simple option, and if that doesn't work either then maybe the problem is with the PDF instead of with the PDF viewers...? | 18:12.23 |
| JakeSays | sebras: i can select and copy it in acrobat reader, but the characters it copies are not what is being displayed | 18:13.12 |
| | sebras: and when i try to edit it in foxit's editor, it gives me an error saying the font doesnt support the character i'm typing | 18:14.01 |
| | which fits robin_watts_mac's theory | 18:14.08 |
| sebras | exactly. | 18:14.13 |
| JakeSays | which, if the glyphs are exactly the same for every pdf, then i can come up with a glyph to char mapping easy enough | 18:15.08 |
| | but my guess is they're not | 18:15.19 |
| | actually, i bet the glyph for 'a' isn't even the same for every 'a' in one pdf. there are several embedded fonts | 18:16.27 |
| sebras | that sounds plausible. | 18:16.41 |
| JakeSays | i'm going to tweak txtwrite to dump the glyph # so i can see whats up | 18:17.25 |
| sebras | actually the same 'a' in the same font might have different indicies on different pages unless the same font object in the PDF is re-used on each page. which it may or may not be. | 18:17.37 |
| JakeSays | these are single page pdfs | 18:18.05 |
| | they're actually generated from a pdf print driver | 18:18.19 |
| sebras | JakeSays: this might help you. in mupdf there is a tool called mudraw, if you use it like this: mudraw -x texty.pdf 3 then you will get a xml-representation of page 3. | 18:19.03 |
| | that would contain e.g. the ucs-charcter (if any) and a gid. | 18:19.21 |
| JakeSays | sebras: did that. mudraw shows "?" for the chars | 18:19.39 |
| | however, dont remember seeing a gid. hmm. | 18:19.56 |
| | sebras: i dont see a gid | 18:22.15 |
| | sebras: <char bbox="-30.8162 346.557 -25.9543 355.486" c="?"/> | 18:22.36 |
| sebras | JakeSays: um.. are you sure you are using mudraw -x and not mudraw -ttt ? | 18:25.27 |
| JakeSays | sebras: ah, yeah thats much better | 18:25.57 |
| sebras | JakeSays: great. :) | 18:26.08 |
| JakeSays | dang. thats very helpful! | 18:27.10 |
| sebras | JakeSays: you still need to make the mapping table, and it might not be 100%, but you have something to start with at least. | 18:28.16 |
| JakeSays | right | 18:28.28 |
| | lol now i know what a display list is | 18:29.26 |
| | sebras: are the coordinates in pixels? | 18:36.46 |
| sebras | JakeSays: no, they are probably expressed in user space. in addition I'm unsure if all translation matrices are taken into account. I think so, but I might be wrong. | 18:40.21 |
| JakeSays | sebras: i was thinking the matrices weren't applied since they're included in the output, but i'm not sure. i'll consult the source to be sure tho | 18:41.33 |
| sebras | that's probably a good idea. I'm having too much of a cold to be fully reliable. ;) | 18:42.11 |
| JakeSays | sebras: would you happen to know where the code is that generates the display list xml? | 22:24.48 |
| | sebras: lol found it. | 22:25.57 |
| | Forward 1 day (to 2012/12/29)>>> | |
IRC Logs