| <<<Back 1 day (to 2014/02/07) | 2014/02/08 |
Robin_Watts | So if I remove the vdmx table in arial, it matches swissA vertically. That's progress at least. Perhaps I can add a vdmx table to swissA. | 12:31.16 |
| and hinting explains most of the rest. | 12:36.04 |
madmoose | mupdf is delightfully well-designed... In practically no time I (a bit hackily) added multi-page tiff reading, and now mupdf can view them and mudraw (with a small fix to pdf output) can convert a multi-page tiff to a multi-page pdf. | 13:50.00 |
Robin_Watts | madmoose: Oh, you mean you updated the tiff format reader? | 13:50.47 |
| so that each image comes out as a new page? | 13:50.56 |
| cool. | 13:51.19 |
| I'd be interested in seeing both those changes. Would you be prepared to let us include the changes in mainline MuPDF? | 13:51.44 |
madmoose | Robin_Watts: Sure, and I can redo them properly if you have design comments. | 13:54.02 |
Robin_Watts | cool. | 13:55.41 |
| In order for us to take them on, we have to get you to sign a copyright assignment to Artifex. | 13:56.29 |
| That way we can continue to license stuff commercially too. | 13:56.48 |
madmoose | It's yours if you want it | 13:56.55 |
Robin_Watts | If you go to bugs.ghostscript.com and open an bug, add details of your enhancement, and attach your patch. | 13:57.31 |
| On the screen where you attach your patch there is a link to the CLA. | 13:57.55 |
| I'm sorry that we have to ask people to go through this pain, but it's a necessary evil. | 13:58.15 |
| We will be preparing the next version of MuPDF very soon, so if we can get this in before that, it would be great. | 13:58.46 |
| Gotta go. back soon, sorry! | 13:58.52 |
madmoose | Robin_Watts: I submitted it in three bug reports (695021 | 16:08.54 |
| 695022, and 695023) and mailed a CLA to Miles. | 16:09.25 |
Robin_Watts | madmoose: Brilliant! Thanks. | 17:40.46 |
toothrot | Robin_Watts, you were right about the spans, they are split across the whitespace. i was a little surprised at some of the blocks were arranged/overlapped at the top though: http://imgur.com/a/S0mhM | 18:06.13 |
Robin_Watts | such as? | 18:06.50 |
toothrot | like up in the top right corner where HHNC and 13/20/53 aren't in the same block | 18:08.16 |
| but like you said, i can always group the spans however i want | 18:08.40 |
Robin_Watts | toothrot: Well, the problem is that they may not be output in the PDF in the order you would expect. | 18:08.57 |
| possibly we are missing some tricks when assembling stuff in strange orders. | 18:09.22 |
toothrot | so the blocks aren't really arranged strictly by coordinate position, but influenced by position in the file? | 18:09.41 |
| would you be willing to entertain changes to the way mupdf handles this? | 18:10.06 |
Robin_Watts | Basically, all the text output stuff sees is a stream of "Put this char here". | 18:10.12 |
| And we gather those up into spans in a really dumb way by saying "does this one fit on the end of our previous span?" | 18:10.40 |
| If it does, we add it. If not, we start a new span. | 18:11.05 |
| And then at the end we run through and try and add spans together. | 18:11.28 |
| Yes, we would absolutely be interested in changes to this. | 18:11.46 |
toothrot | Alright, I'll approach it with that in mind then instead of trying to wrangle the spans externally | 18:13.56 |
| thanks again | 18:14.33 |
Robin_Watts | toothrot: You should look over the existing cide (in stext.c probably) to understand the stuff we do. | 18:15.05 |
| We are at pains to work with text at any rotation, and to copy with superscripts/subscripts etc. | 18:15.27 |
| and to try to cope with inserting spaces etc as required. | 18:15.42 |
| It's all horrible :( | 18:15.47 |
cypher707 | heya - i need the ghostscipt v9.07 sources on debian for 64bit arch. - does there exist a 64 bit version anywhere? its listed for win only as 64 bit :/ | 18:42.49 |
toothrot | cypher707, i could be wrong but i'd be surprised if there were a separate source dist for 64 bit. compile it as 64 bit. | 18:45.33 |
cypher707 | i see , its generic, hm :) ? | 18:46.15 |
| great great and thanks toothrot | 18:46.48 |
ray_laptop | cypher707: as toothrot commented, the ghostscript sources are both 32 and 64. If you 'configure' or 'autogen.sh' on your system, if it's 64-bit it will be a 64-bit build. There are configure options to force 32-bit. | 20:33.22 |
| On windows, where folks aren't used to building from source, we have binaries, which is why there arre 32 and 64 bit | 20:34.03 |
cypher707 | ray_laptop: oh yes , good to know ^ | 20:44.07 |
ray_laptop | psdcmykog is _really_ confused with SeparationOrder. i want to make it act more 'sanely'. It's really not set up to be a DeviceN device. I'm thinking that it would be better to just ignore SeparationOrder | 21:11.23 |
| to see real misbehaviour, try: gs -sDEVICE=psdcmykog -o x.psd -c "<< /SeparationOrder [ /Black /Cyan /Magenta /Yellow ] >> setpagedevice" -f examples/tiger.eps | 21:12.53 |
| Forward 1 day (to 2014/02/09)>>> | |