| <<<Back 1 day (to 2018/03/28) | 20180329 |
mvrhel_laptop | HenryStiles: if you can keep chains disabled for me, that would be appreciated | 04:28.14 |
| This testing is taking a bit long | 04:28.23 |
chrisl | I have all the third part libs updated except freetype (2.9 breaks our output, I need to debug it and see if the problem is there's or ours) and jpegxr (ballache, little used - will get to it at some point) - but in half an hour, I'm going out for a couple of hours, so I'll not push the commits until I get back, just in case. | 09:02.48 |
kens | Righto | 09:03.03 |
tor8 | chrisl: some commits on ghostpdl tor/master that could do with a look over if you've got a spare moment | 11:23.47 |
| some fixes to jbig2dec mainly | 11:23.55 |
chrisl | tor8: the xps one bothers me: Robin_Watts specifically extended that buffer to fix a problem: http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=b69d68995d065d61b2defe919c038cf9cf60b598 | 11:41.53 |
tor8 | chrisl: he extends it so the pointless null-terminator has a place to go. the null terminator serves no purpose so removing both that and the extra space for allocation should be safe. | 11:43.43 |
| but best to double check | 11:44.06 |
kens | lunches | 11:44.37 |
chrisl | tor8: Okay, I get that... the stdint one makes me twitch, too, but only because that's a bit a of a house of cards..... I don't actually see anything wrong with it | 11:50.04 |
tor8 | chrisl: though I can't reproduce a buffer overrun that triggers valgrind, but that could just be gs allocator? | 11:50.36 |
| chrisl: yeah, I anticipated some twitching because it's a real house of cards | 11:50.51 |
| but it's causing me unneccessary headaches with my cmake experiments in mupdf, needing to set HAVE_STDINT_H when it really shouldn't be necessary | 11:51.40 |
chrisl | tor8: Possibly the allocator, yes - I can take a look after lunch if you like | 11:51.43 |
tor8 | chrisl: I tried reverting robin's commit to see if I could trigger, but valgrind is not being helpful and I'm too rusty on debugging ghostscript | 11:52.24 |
chrisl | Possibly you need to build with CFLAGS=-DSINGLE_OBJECT_MEMORY_BLOCKS_ONLY | 11:53.08 |
| tor8: But, tbh, I'm sure you're right about dropping the space, so I'm fine with it | 11:55.06 |
tor8 | chrisl: okay, I have verified that my fix does not reintroduce the bug fixed by robin's patch | 12:07.11 |
chrisl | tor8: I didn't really expect it to be a problem, I just remembered Robin's change, couldn't remember the circumstance... so your commits LGTM | 12:08.48 |
tor8 | chrisl: thanks. I'll give them a final cluster before pushing. | 12:09.37 |
chrisl | reboots | 13:55.51 |
kens | chrisl HenryStiles in Bugzilla I see the only component we have for GhostXPS appears to be 'General' which is OK, but rather limited. Could we have the same components as for Ghostscript ? Or at least, some of them ? It would at least be nice to get the PDF Writer component in there. | 14:42.24 |
chrisl | I can't think why not..... | 14:42.58 |
kens | And thanks for adding the 9.23 version whoever did that :-) | 14:43.32 |
chrisl | I did - someone prompted me on bugzilla | 14:44.02 |
| Uh, this could take a while :-( | 14:44.53 |
kens | Don't let me take you away from anything more important | 14:49.53 |
| I only noticed it because of the XPS bug report when I was trying to change the component | 14:50.12 |
chrisl | I've added a bunch of the more important (it seems to me) ones. I can others as and when people notice a lack | 14:51.03 |
| I also fixed the contents of a couple of the gs ones..... | 14:51.28 |
kens | Thanks chris | 14:51.48 |
| Yeha that's great I can change ti to PDF Writer now | 14:52.21 |
| Forward 1 day (to 2018/03/30)>>> | |