| <<<Back 1 day (to 2015/07/12) | 20150713 |
Guest49075 | Hi All good morning | 05:57.40 |
mvrhel_laptop | Robin_Watts: for the logs. I pulled your gsproof changes. did not get a chance to look them over in detail. I am going to be out Tuesday-Thursday so it may not be until the end of the week before I can jump in. | 06:23.08 |
Guest49075 | Hi robin | 06:41.56 |
kens | Hmm, have we changed Bugzilla version ? | 07:20.58 |
| AH no, my browser jujst displayed it oddly | 07:21.30 |
| If anyone's read Niven/Pournelle's "Fallen Angels": | 08:02.28 |
| http://www.theregister.co.uk/2015/07/13/mathematician_predicts_sun_to_go_quiet_for_ten_years_from_2030/ | 08:02.28 |
sebras | kens: do you know the idea behind this gsproof thing that robin implemented for mupdf? | 12:03.12 |
kens | yes | 12:03.21 |
sebras | kens: as far as I can tell the idea is that ghostscript renders the pdf and that mupdf visualizes it..? | 12:03.29 |
kens | Yep | 12:03.34 |
sebras | kens: and the reason for doing this is that you need color accuracy for proofing..? | 12:03.49 |
kens | For GSView | 12:03.55 |
sebras | kens: what do you me by "for gsview"? | 12:04.12 |
| kens: how is gsview related to mupdf? | 12:04.22 |
kens | That's what iots for | 12:04.24 |
Robin_Watts | sebras: Hi. | 12:04.33 |
sebras | Robin_Watts: good morning! | 12:04.42 |
Robin_Watts | Yes, it's for color correct proofing. | 12:04.51 |
| Also, for correctness w.r.t. the 'any part of a pixel' rule that we don't strictly implement in mupdf. | 12:05.24 |
sebras | ok, and the idea is to use mupdf as a viewer just because it is already a handy app..? | 12:05.33 |
| Robin_Watts: but those things are bugs in mupdf really, and ought to be fixed at some point..? | 12:05.56 |
Robin_Watts | sebras: MuPDF is more suited to display use, yes. | 12:05.56 |
| Color correctness should be sorted in mupdf at some point. | 12:06.10 |
sebras | mmm, probably using lcms or something I guess. | 12:06.23 |
Robin_Watts | We also don't get transfer functions or transparency right. | 12:06.27 |
| yes, we've been planning lcms2 integration for a while, but Michael has never been free enough. | 12:06.43 |
sebras | transparency I thought was correctly implemented. | 12:06.50 |
Robin_Watts | No, it's wrong in various cases. | 12:07.05 |
sebras | ah! I didn't know that. :) | 12:07.17 |
Robin_Watts | It takes me about a week to remind myself what's wrong with it each time. | 12:07.29 |
| and then I can kinda see a route to fixing it. | 12:07.42 |
| but it involves rewriting lots of the rendering code. | 12:07.52 |
sebras | right, so it gets pushed into the future each time. | 12:08.12 |
Robin_Watts | I may possibly have a go at it on holiday. It all depends. | 12:08.26 |
sebras | or we wait until tor[58] wakes up one day and can't resist the urge to fix it. ;) | 12:09.00 |
| thanks, now I get the idea behind those formats at least. | 12:10.34 |
Robin_Watts | reckons tor should be back soon... | 12:13.03 |
| mvrhel_laptop: Hey | 14:39.56 |
mvrhel_laptop | morning Robin_Watts | 14:40.08 |
Robin_Watts | I've fixed the leaks in the gproof stuff. | 14:40.21 |
| but I noticed a problem. | 14:40.30 |
| where the image tiles abut, the gridfitting that mupdf does means we lose a pixel between neighbours most of the time, and that looks horrible. | 14:41.03 |
mvrhel_laptop | oh no | 14:41.10 |
Robin_Watts | I've got a fix here for it - I change the gridfitting algorithm so that we round edges to the nearest pixel, rather than always rounding outwards. | 14:41.44 |
| Quality looks much improved. | 14:41.48 |
mvrhel_laptop | oh great | 14:41.52 |
Robin_Watts | Now I need to figure out a nice way to turn it on just for these files. | 14:42.00 |
mvrhel_laptop | I see | 14:42.07 |
Robin_Watts | I could add a new device method for setting render flags, the first of which could be 'gridfit for tiled images'. | 14:52.02 |
| or I could try and get away without the method. | 14:52.19 |
| by abusing the dev->flags. I think I prefer the former though. | 14:52.34 |
mvrhel_laptop | beating gsview to death here with docs that have wacky sized pages, (e.g. very small next to very large). Found several issues.... | 17:56.57 |
Robin_Watts | ugh. | 17:59.03 |
mvrhel_laptop | it is a big pain | 18:04.18 |
| Forward 1 day (to 2015/07/14)>>> | |