| <<<Back 1 day (to 2012/11/09) | 2012/11/10 |
henrys | Robin_Watts_:I think I'm seeing a problem with gapto and closepath. it seems that when closepath goes back to the last moveto - it seems that is exactly right when filling but stroking shouldn't work like that I think I can work around it in the pcl code. The problem is the graphics library doesn't know if it will be a stroke or a fill when the closepath is issued. | 00:14.31 |
| more briefly I think: closepath should go back to the last gapto in stroke mode and back to moveto in fill mode. | 00:16.02 |
Robin_Watts_ | oh. | 00:17.33 |
| That would be problematic for me to handle in the gfx library I think, due to the stupid subpath representation we have in gs. | 00:18.31 |
henrys | I can probably fix this in pcl if gx_path_new_subpath is exposed to PCL. I assume that won't be a problem. | 00:19.14 |
Robin_Watts_ | That doesn't sound good to me. | 00:19.36 |
henrys | I'd feel the same if gx_path_new weren't available but the cat's out of the bag so to speak | 00:21.03 |
Robin_Watts_ | No, having looked at the code, I'm even less keen. | 00:21.13 |
| No, gx_path_new makes perfect sense. | 00:21.38 |
| gx_path_new_subpath is exposing part of the internals of gs' data structure. And not even a nice part. | 00:22.12 |
henrys | oh I thought you were concerned about the languages having the gx_path data type. | 00:22.39 |
Robin_Watts_ | no. | 00:22.50 |
henrys | I haven't read subpath_new carefully | 00:22.56 |
Robin_Watts_ | "closepath should go back to the last gapto in stroke mode." | 00:23.25 |
| I'll need to stare at the code for a bit to think if that can be done. | 00:24.22 |
| strokes are a bit of a git, cos if memory serves the first section is rendered differently if they are 'closed' or not. | 00:25.10 |
| If the subpath doesn't 'close' back to the start of the subpath, then I need a different 'is_closed' flag. | 00:25.38 |
| Let me ponder on it for a bit. | 00:26.51 |
henrys | sure | 00:27.00 |
Robin_Watts_ | There may be a way to do it using notes. | 00:27.10 |
tor8 | Robin_Watts_: ping | 08:53.01 |
| Robin_Watts_ I read the irc log from last night, about gapto, moveto and closepath. If gapto is used to create a gap when dashing an outline, and moveto is used to create a new subpath, it may be similar to what happens in XPS. In XPS a closepath on a stroked, dashed path will join the last segment to the first with a linejoin rather than end caps. It requires a fair bit of voodoo to get right. | 08:56.01 |
| Robin_Watts_: and if not, disregard what I just said :) | 08:56.46 |
sebras | tor8: Robin_Watts_: a slightly controversial patch over at sebras/master for review. I simply delete variables that a unused that cause warnings under gcc 4.7.1. | 10:52.38 |
| xps attributes might be useful, but reintroduce them when they are used in that case. | 10:53.05 |
| I also silenced a fz_warn in pdf/pdf_cmap.c that causes cmap generation to print warnings. | 10:53.35 |
| paulgardiner: http://www.uscis.gov/files/form/n-400.pdf http://bugzilla.gnome.org/show_bug.cgi?id=687951 I can't see the form either -- should I be able to even if I don't compile with v8? | 11:10.39 |
Robin_Watts_ | sebras: Did you see the tech preview release binaries are up? | 18:11.03 |
| Forward 1 day (to 2012/11/11)>>> | |