| <<<Back 1 day (to 2012/06/29) | 2012/06/30 |
mvrhel | henrys: huge win on 693042 | 00:25.01 |
| file goes from taking 12 minutes+ to a few seconds with this fix in the xps interpreter | 00:25.19 |
| I will do the cluster push testing this weekend as there may be a few minor fixes to deal with but over all this will be a nice fix | 00:25.53 |
| Robin_Watts: thanks for you help with the clip path stuff earlier today | 00:26.15 |
| off to get dinner going, and clean the 16.5 pounds of strawberries that got picked today | 00:26.48 |
norbertj | mvrhel: nicely done. I suspected that this could be the problem. IIRC there was a similar problem a couple of years back (when we started with xps). Have to look (at work) what the bugreport was for this. | 14:30.09 |
sebras | Robin_Watts: you guys really ought to upgrade to a later version of bugzilla... it complains if my password is longer than 16 chars. :-P | 15:26.07 |
| Robin_Watts: I reviewed your patch at robin/master. I think fz_write_buffer_bits() is still buggy. | 15:31.54 |
Robin_Watts | sebras: really? You can see cases in which it will go wrong? It's quite possible - it's not had much use. | 15:32.30 |
sebras | so I cherry-picked your patch to sebras/master, added a function to hexdump fz_buffers, implemented a test program http://pastebin.com/m46UPcT7 which illustrates the bug I;m seeing. | 15:32.33 |
| and fixed it over at sebras/master. | 15:32.40 |
| lastly there is a whitespace fixing patch. | 15:32.54 |
| if you feel like it, have a look. | 15:33.04 |
| it's hard for me to find the details of how the current implementation is misbehaving, but I think my patched one does what it advertises to do at least. :) | 15:33.56 |
| Robin_Watts: so basically I always keep val shifted to the left so that the most significant bit is the next one to be written to buf->data. | 15:34.54 |
Robin_Watts | sebras: Thanks, I'll look, but probably not today. | 15:34.55 |
sebras | Robin_Watts: I realized, so I set the pastebin timeout to a month... ;) | 15:35.07 |
Robin_Watts | I'd like to understand why my code is wrong before fixing it (I'm vain like that :) ) | 15:35.43 |
sebras | Robin_Watts: I fully sympathize with you. it's just that I can't really quantify _why_ :) | 15:36.18 |
| Robin_Watts: though the pastebin snippet illustrates it quite well. | 15:36.31 |
| that one does require my fz_buffer_hexdump though. | 15:36.55 |
marcosw | Robin_Watts: I've installed g++ on the all the cluster nodes that I control. | 23:00.30 |
Robin_Watts | marcosw: Thanks. | 23:26.26 |
| I can't quite get bmpcmp working; for some reason it doesn't like the format of the images. I'll look into it more on monday. | 23:26.55 |
| Then I need to hook it up so that commits on the forms branch get tested with it automatically. | 23:27.16 |
| Forward 1 day (to 2012/07/01)>>> | |