| <<<Back 1 day (to 2015/03/20) | 20150321 |
Robin_Watts | tor8: For the logs, I have my stuff rebased onto the top of your stuff in a form I'm happy with. | 17:33.46 |
| It's on robin/merge_csi | 17:33.57 |
| The cluster is still showing 750ish differences though. | 17:34.20 |
| I wonder if the path problems in your stuff are as simple as you not closing paths... | 17:47.30 |
| aha. | 18:18.15 |
| op_s and op_f were being called incorrectly. | 18:18.39 |
| down to 237 diffs now. | 18:33.54 |
| op_w was not flushing text. | 18:57.17 |
| crap. The remaining problems (or at least some of them) are down to us not having the right pdf resources when processing a softmask. | 23:55.08 |
| I suspect the correct solution to this will require us to pass a pdf_obj *resources to every op function. | 23:55.41 |
| This means that every op function is now taking (as a minimum) fz_context, pdf_processor *proc, and pdf_obj *resources. | 23:56.12 |
| At this point I wonder whether it's not worth putting those into a structure and passing a single pointer to that. | 23:56.34 |
| Forward 1 day (to 2015/03/22)>>> | |