| | <<<Back 1 day (to 2017/09/07) | 20170908 |
| sebras | tor8: there are two easy patches on sebras/master | 12:35.02 |
| tor8 | sebras: those two LGTM | 12:57.44 |
| | as does the third | 12:58.26 |
| sebras | tor8: ok. | 13:00.12 |
| | tor8: and it clustered ok, I'll refetch origin and push in a minute. | 13:00.46 |
| | I especially like the fact that one of the patches removed 200 lines of code but changes nothing. :) | 13:01.45 |
| | Robin_Watts: there's a commit on sebras/wip that I'd like to discuss with you. | 13:02.56 |
| | in do_mark_line_app() I believe that } else if ((y_steps = ey - sy) > 0) { will handle any positive y_steps | 13:04.22 |
| | in the else you handle y_steps <= 0, and you start by y_steps = -y_steps | 13:06.45 |
| | next you check for y_steps < 0, and I feel that this can never be true..? | 13:07.08 |
| Robin_Watts | y_steps = 0x80000000 | 13:13.15 |
| sebras | yes, I see it now, it is because MIN_INT is one less than MAX_INT and so when inverting it is still negative. | 13:13.48 |
| Robin_Watts | exactly. | 13:14.03 |
| sebras | Robin_Watts: I concur with "awkward". :) | 13:14.19 |
| Robin_Watts | And I had a fuzzing file that hit exactly that case. | 13:14.20 |
| sebras | Robin_Watts: I saw something in clang and/or coverity pointing to this bit of code while you were in .mg. | 13:15.17 |
| | Robin_Watts: of course I have lost what prompted me to look at it by now. oh well. | 13:15.34 |
| Robin_Watts | mvrhel_laptop: Morning. | 13:56.19 |
| mvrhel_laptop | Morning Robin_Watts | 13:56.28 |
| Robin_Watts | I found a fix for the compatible overprint thing. | 13:56.30 |
| mvrhel_laptop | oh great | 13:56.36 |
| Robin_Watts | brace yourself... | 13:56.56 |
| | http://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=0f15c6e2a9f2c0377ac2c66a555fdfc476eb9c02 | 13:56.58 |
| mvrhel_laptop | nice! | 13:57.32 |
| Robin_Watts | looking at the page 13 image things now. | 13:57.52 |
| mvrhel_laptop | ok. you probably want to unwind my one fix when working on that | 13:58.14 |
| Robin_Watts | have done. | 14:01.08 |
| | sebras: You here? | 14:13.13 |
| sebras | Robin_Watts: I am. | 14:13.41 |
| Robin_Watts | Got time to talk about this jpx thing? | 14:13.53 |
| | In commit fc2e086 you change the code slightly. | 14:14.30 |
| | Rather than being passed in an 'indexed' bool, you check the colorspace to see if it's indexed. | 14:14.50 |
| | What do we gain by that? | 14:14.56 |
| | Presumably there is some reason why that's better? | 14:15.04 |
| sebras | let me have look. of course I forgot. | 14:17.47 |
| | Robin_Watts: naïvely I would say that we pass fewer arguments and only call fz_colorspace_is_indexed() at a later stage when it is necessary..? | 14:19.09 |
| | I can't see that this change would cause a behavioural change..? | 14:19.31 |
| Robin_Watts | but we call fz_colorspace_is_indexed twice now. | 14:19.49 |
| | If it's a "done for tidiness, not for effect" commit, I can live with that explanation. | 14:20.48 |
| | I was worried there was some cunning behavioural benefit that I was missing. | 14:21.23 |
| sebras | none that I remember. | 14:22.07 |
| Robin_Watts | ok, ta. | 14:22.34 |
| sebras | Robin_Watts: I presume that is related to the indexed jpx commit you have on robin/spots? | 14:22.52 |
| | Robin_Watts: how close are you to getting robin/spots merged by the way? | 14:23.06 |
| Robin_Watts | sebras: This is me trying to understand what's going on w.r.t that commit. | 14:26.24 |
| | sebras: hopefully we can merge it soon. | 14:26.33 |
| sebras | tor8: old patch resurrected onto sebras/master | 15:36.48 |
| Robin_Watts | Ahem. I found the jpx colorspace issue. | 15:55.46 |
| | Probably best *not* to scale an indexed image before expanding it ;) | 15:56.06 |
| sebras | Robin_Watts: ah, that does indeed sound bad. | 15:59.56 |
| Robin_Watts | mvrhel_laptop: So, robin/spots contains something that fixes most of the problems. | 16:12.49 |
| | There is still 1 image that's a bit lighter than I'd hope for on page 13. | 16:13.05 |
| | I'm just trying to massage it into something we can start to merge back now. | 16:13.27 |
| mvrhel_laptop | Robin_Watts: great | 16:15.51 |
| Robin_Watts | So mvrhel_laptop: I think what I have on robin/spots is probably as 'coalesced' as I can reasonably get it. | 17:15.51 |
| mvrhel_laptop | Robin_Watts: ok | 17:16.03 |
| Robin_Watts | Do you want to give it a once over to see if you can spot anything before we merge it back? | 17:16.17 |
| mvrhel_laptop | Robin_Watts: How did you end up fixing the jpx issue | 17:16.40 |
| | ? | 17:16.43 |
| | I don't see that commit | 17:16.49 |
| Robin_Watts | mvrhel_laptop: Ahem. | 17:16.50 |
| | I stopped the drawing code scaling an indexed pixmap. | 17:17.10 |
| mvrhel_laptop | ah | 17:17.26 |
| | that would be a problem | 17:17.33 |
| Robin_Watts | Indeed. Stupid mistake on my part. | 17:17.48 |
| | Forward 1 day (to 2017/09/09)>>> | |
Log of #mupdf at irc.freenode.net.