| | <<<Back 1 day (to 2018/03/17) | 20180318 |
| vtorri | velix : you can use MSYS2 | 06:01.30 |
| Hufokus | Hello! It looks like fz_run_page is very slow operation. Am I right, or may be just using it in a wrong way? | 10:35.31 |
| malc_ | Hufokus: depends on the page | 11:48.41 |
| Hufokus | malc_: so generally it is a wrong idea to use fz_run_page or fz_run_display_list in real-time? The best way is to preload, store them in memory and only copy fz_pixmap_sample at render stage? | 11:52.31 |
| malc_ | Hufokus: sorry, i don't have an answer to that. you probably should ask sebras and tor8 and possibly robinwatts here (when they are online that is) | 11:57.42 |
| Hufokus | malc_: thanks anyway) | 12:36.19 |
| sebras | Hufokus: with fz_run_display_list() you first need to prepare a display list which means parsing the file into the display list datastructure. | 16:05.51 |
| | Hufokus: with fz_run_page() you would reinterpret the file every time. | 16:06.07 |
| | Hufokus: so if you want to render the same page multiple times using a display list is the way to go. | 16:06.27 |
| | Hufokus: this is a good read to get an overview of mupdf https://www.ghostscript.com/~robin/mupdf_explored.pdf | 16:07.08 |
| Hufokus | sebras: fz_run_display_list takes about and fz_new_pixmap_from_display_list take about 30 % of rendering time each (on some pdfs, not on all of them) Is it a way to reduce time? Previously I was using preload of pixmap at loading stage and then just copy pixmap->samples at render stage. But now I need zooming, so it looks like I need scaling at real time( | 19:13.18 |
| | sebras: may be I should store display list and also pixmap obtained from this display list at loading application stage? And then, at render stage just use fz_run_display_list with desired scaling matrix? | 20:01.37 |
| | Forward 1 day (to 2018/03/19)>>> | |
Log of #mupdf at irc.freenode.net.