| <<<Back 1 day (to 2013/09/13) | 2013/09/14 |
mvrhel_laptop | henrys: for the logs, I know it is late. I stepped through the glyph filling stuff when we are filling the mono bitmap characters. | 05:32.31 |
| so during the fill operation there is logic in there to avoid doing any color conversions in these cases. | 05:32.49 |
| essentially, the current color space is set to gray and the fill color is set and it matches the device and there is not a remap, so hence no ICC stuff is invoked. | 05:33.25 |
| so the short of it all is that it is safe to simply use the forward procedure like I did with gs_make_mem_mono_device | 05:34.48 |
| alternatively, it may make sense to have it return null though since it should not be used anyway and then address those bugs accordingly | 05:39.44 |
| we can talk a bit more about it next week if you would like. | 05:40.12 |
| when I say "have it return null" I mean have the profile for the device be null and if someone is trying to get and use it, things will trip up. | 05:41.15 |
| done for night | 05:41.43 |
vtorri | hey | 07:31.29 |
| just in case it interest you, the latest libgs code seems to work on win 7 64 bits | 07:32.02 |
| but when i compile libgs, there is a huge number of warnings | 07:32.46 |
| especially when using %x and a ulong | 07:33.04 |
| or conversion from a pointer to an int | 07:33.15 |
kens | Its not possible to remove all warnings from all compiler on all operating systems,so we ignore benign ones | 07:33.43 |
| new warnings are flagged up when copmmits are made, and investigated at the time for potantial consequences | 07:34.11 |
vtorri | ok | 07:34.22 |
| kens: but better using %lx with a ulong, no ? :) | 07:35.17 |
kens | Without looking at the code I can't be certain. | 07:35.36 |
| It may well be that the value cannot exceed a long | 07:35.54 |
vtorri | and ? | 07:36.10 |
kens | And it's not a problem | 07:36.18 |
vtorri | and using %lx is also not a problem | 07:36.32 |
kens | We're not about to go rummaging through the code looking for all the places and changing them | 07:36.52 |
| That's much more likely to have conseuquences when people make mistakes | 07:37.07 |
| And we'd rather spend the time working on real problems | 07:37.31 |
vtorri | indeed | 07:38.12 |
| btw, you (the devs) have fixed a problem recently which make libgs not working at least on Windows | 07:38.46 |
kens | Not that I'm aware of | 07:38.57 |
vtorri | bug 694534 | 07:40.10 |
kens | THat's nothign to do with libgs, it was a general problem | 07:40.44 |
vtorri | and as i'm not a dev, it's not surprising | 07:41.05 |
| and i'm also using only libgs, so... | 07:41.22 |
| anyway, it seems i bother you | 07:43.01 |
| sorry for that and bye | 07:43.11 |
Noldorin | hmm. is it possible to pass in arrays to ghostscript via -d⦠? | 13:09.50 |
ray_laptop | Noldorin: seems nobody else was around. -d cannot be used to define things other than numbers, true, false, null, or PS names (start with / ) | 18:38.18 |
| Noldorin: you can define things using -c "..." command line argument to define arrays. For example: -c "/myarray [ 1 2 3 ] def" | 18:41.21 |
| Noldorin: note that the objects cannot be put in systemdict as -d and -s options are. They will be put in userdict by default | 18:42.38 |
Noldorin | rayjj, aha | 21:43.40 |
| thanks | 21:43.42 |
| Forward 1 day (to 2013/09/15)>>> | |