| | <<<Back 1 day (to 2014/09/26) | 2014/09/27 |
| chrisl_away | kens: just briefly - that crash (at least, the one I see) is definitely DCT decode stream related...... | 08:15.20 |
| kens | chrisl_away : OK I cna look at it if you are out | 08:15.33 |
| | I see you also grabbed the %stdin one | 08:15.40 |
| chrisl_away | That's trivial..... | 08:15.51 |
| kens | While I was styill rading the mail :-) | 08:15.59 |
| | Going to have to send Marcos maill, he still hasn't attached the customer file | 08:16.28 |
| chrisl_away | The cause of the DCT one is that when we close the stream, we free a load of the "state", but the file object remains reference in the Postscript world. When the file object is later garbage collected, we then try to free the same parts of the state...... | 08:16.52 |
| kens | Oh that sounds painful | 08:17.09 |
| | Maybe have to figure out which parts not to free when the stream is closed | 08:17.30 |
| | I'll take a poke at it in as bit, got to finish SO first | 08:17.46 |
| chrisl_away | Yes, fixing the specific crash is trivial, but it would benefit from another hour or so just checking if the same applies to other parts of the "state". | 08:18.33 |
| kens | I cna look at it, probably we're about equal on that | 08:18.52 |
| chrisl_away | It's up to you, I don't mind finish up with it on Monday - depends what else you have going on | 08:19.23 |
| kens | In the absence of the customer file, just a stupid file from paps, reported on SO | 08:19.42 |
| chrisl_away | There, that's the %stdin fix pushed..... | 08:22.01 |
| kens | THat was quick :-) | 08:22.11 |
| chrisl_away | Like I said, pretty trivial..... there's actually a comment in the code which pointed me in the right direction | 08:22.38 |
| kens | Well that's remarkable, must be a first to have a useful comment..... | 08:23.11 |
| chrisl_away | Pretty much - I will say that it was a hacky work-around for a broken API, and it just hadn't been applied everywhere it was needed. Of course, I'd argue for fixing the API, but...... | 08:24.24 |
| kens | :-( | 08:24.40 |
| chrisl_away | I'll grant that in this case, it's a relatively innocuous one | 08:25.13 |
| | Okay, off to see how my ankle holds up on a squash court! | 08:25.38 |
| kens | Have unf | 08:25.54 |
| | fun* | 08:25.57 |
| chrisl_away | No, I think you were right the first time!! | 08:26.08 |
| kens | LOL | 08:26.13 |
| chrisl_away | let me know about the DCT issue - don't want to pick it up on Monday if you're working on it | 08:26.37 |
| kens | No problem | 08:26.51 |
| chrisl_away | Bye | 08:26.55 |
| kens | Just finishing up my SO reply now | 08:26.57 |
| | bb | 08:27.00 |
| | Forward 1 day (to 2014/09/28)>>> | |
IRC Logs