| | <<<Back 1 day (to 2018/04/22) | 20180423 |
| ray_laptop | StevePhillips: in preparation for guiding you through the clist bug HenryStiles bestowed upon you, I had a look at what the clist code is doing. The writer and the reader "state" gets out of sync in that the delta height (dheight) calculated by the writer when it writes the fill_rect is -2, based on a height of 1 and a previous height of 3 | 22:46.12 |
| | StevePhillips: but then in the clist reader, in gxclrast.c line 987, it sets the state.rect.height from the tile height, which is 4 (in the copy: section for the copy_mono_planes cmd). This leads to the delta height calculation for the subsequent fill_rect giving 4 - 2 = 2 (rather than 3 - 2 = 1) | 22:49.10 |
| | I'll leave it to you, but I think the problem is that the actual copy clips the height to dev->height - y (as the writer did -- see line 356 of gdevm24.c) but the tile actually goes on past that. If I set the state.rect.height to 3, then I get a height of 1 for the subsequent fill_rect | 22:59.45 |
| | Forward 1 day (to 2018/04/24)>>> | |
Log of #ghostscript at irc.freenode.net.