| <<<Back 1 day (to 2018/11/06) | 20181107 |
paulgardiner | tor8: I'm looking again at in-situ form filling. A while ago, you added something on a branch, I believe, to generate fz_text from an annotation. Can you point me at it? | 09:45.15 |
tor8 | paulgardiner: fz_new_stext_page_from_annot | 10:36.30 |
paulgardiner | Thanks. | 10:43.23 |
| I had imagined it was still on a side branch somewhere. Hadn't realized you'd committed it to master | 10:44.04 |
tor8 | paulgardiner: is your patch on paul/master good to push? | 12:49.02 |
paulgardiner | I believe so | 13:34.18 |
tor8 | Robin_Watts: http://git.ghostscript.com/?p=user/tor/mupdf.git;a=commitdiff;h=2997ee339c1b260e95dc9e449f63e9b41600be03 | 14:41.11 |
| I don't understand the code, it's too twisty for me. My patch is probably wrong. | 14:42.28 |
Robin_Watts | tor8: Well, if you remove that line, you might as well remove the whole section. | 14:55.48 |
tor8 | Robin_Watts: yeah, I just realized. question is ... is that the right thing to do? | 14:58.41 |
| the comment sounds like it's supposed to prevent double headers | 14:58.58 |
Robin_Watts | yeah. | 14:59.11 |
| Let me prod at it for a mo. | 14:59.16 |
tor8 | Robin_Watts: bah. more color management bugs. fz_pixmap_convert cannot convert CMYKA to RGBA -- the output pixmap is just blank full of zeroes | 15:19.53 |
| with ICC profiles | 15:20.07 |
| cmsFLAGS_COPY_ALPHA is not being set properly in fz_lcms_init_link | 15:25.42 |
| yeah, this design is broken. | 15:32.47 |
| fz_new_icc_link hardwires into the link whether it should copy alpha or not | 15:33.06 |
| but that function takes colorspaces as input, and we don't know there whether or not we'll be using the link to convert a pixmap with or without alphas | 15:33.50 |
Robin_Watts | tor8: http://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=86330c6fd626a861daba86994dff8596d939fddd | 15:47.31 |
| tor8: urgh. | 15:47.59 |
tor8 | if I force copy_spots=1 when asking for a link in icc_conv_pixmap I have gotten it to work | 15:49.34 |
| clustering now | 15:49.36 |
| maybe rename copy_spots to copy_extras everywhere? | 15:49.45 |
| cluster shows no diffs, so at least it hasn't broken anything obvious | 15:50.22 |
Robin_Watts | cluster doesn't test spot rendering. | 15:54.06 |
tor8 | Robin_Watts: double header check LGTM, I think | 15:59.36 |
Robin_Watts | tor8:Ta. | 15:59.49 |
tor8 | Robin_Watts: could you nod at the top commit on tor/master? | 16:00.58 |
Robin_Watts | tor8:looking. | 16:01.05 |
| ok. | 16:02.15 |
| Not entirely sure I agree that it's a broken API, but it'll do. | 16:02.34 |
tor8 | I'll reword the comment to something less strong then :) | 16:03.52 |
moolc | sebras: 9e604ba1b8f05e94c177266cb7bdf3d2a01b529c - shame on you | 16:24.12 |
| sebras: i take that back... shame on my eyes | 16:25.08 |
| Forward 1 day (to 2018/11/08)>>> | |