| <<<Back 1 day (to 2011/09/16) | 2011/09/17 |
Lafriks | thanks for information | 00:02.14 |
Robin_Watts | np. | 00:03.18 |
Lafriks | I'll test your code on monday to see how it performs in my use case | 00:03.39 |
Robin_Watts | My code is not complete enough for you yet, I fear. | 00:05.24 |
Lafriks | for what I have found that first static variables to move to context are error/warning message and freetype library pointer | 00:08.17 |
| with these move to context it does perform very good and I have not found any problems with it | 00:09.07 |
| btw why are there reasons behind changing error/warning handling in mupdf? | 00:11.35 |
| *there -> the :) | 00:11.48 |
| why -> what | 00:12.59 |
Robin_Watts | Currently, when mupdf runs out of memory, it just exits. | 00:25.55 |
| We want to make it cope more gracefully. | 00:26.11 |
| So we either need to make it 'bubble' errors at every possible allocation site, or we need to move to an exception based scheme. | 00:26.56 |
| I've used the exception based scheme in other projects and found it to be much clearer; the flow of the program is maintained with much less intrusive "If that failed, then cleanup" code. | 00:28.20 |
| exceptions are used for what really are exceptional cases. | 00:28.34 |
| I'm off to bed now, but will check logs in the morning in case you have any more questions. | 00:29.04 |
Lafriks | that sound very nice | 00:29.18 |
| I'm good, thanks again | 00:29.30 |
| will clean up my code on monday and prepare patch | 00:29.55 |
| hope it will be any use :) | 00:30.10 |
hemin | Hi, can I use ghostscript to convert a PDF into a PCL? | 02:40.18 |
sebras | wow. the pdf in the zip-attachment over at http://bugs.launchpad.net/inkscape/+bug/846389 has a 100Mbyte page content stream where they are simulating shadings using paths... | 14:00.32 |
| tor8: and pdfdraw -dx on it segfaults in fz_trace_clip_path() due to rect being NULL. | 14:07.10 |
frafra | hi all | 15:40.19 |
| I've a problem with some pdf | 15:57.33 |
| I have four slides on a pdf, and I would like to cut them, in order to get 1 slide for 1 page | 15:57.57 |
| is it possible? | 15:58.00 |
sebras | tor8: mupdf is giving an error on bugzilla 692320 due to Pages having Count as real instead of int. it's generated by some Java pdf-library, so it may be a question of what we desire to support or not. should we truncate reals in this case? | 16:20.41 |
| tor8: 692419 and 692158 seem to affect mupdf performance-wise. 692419 likely to unsorted dict keys. | 16:34.22 |
cryptopsy | how are djvu translated to pdf via the intermediary ps form? The pdf looks like it was transformed into images and the scrolling in mupdf is very slow --- slower than any other pdf i've scrolled through in that program | 22:40.47 |
sebras | cryptopsy: to the best of my knowledge djvu are most likely converted to some rather high-resolution images that (may be compressed and) are added on each page in a pdf. running pdfinfo mynicepdf.pdf may give you some information on the size of the images. | 23:17.21 |
tor8 | cryptopsy: djvu is an image format, so you'll end up with a big full page image per page | 23:31.27 |
| Forward 1 day (to 2011/09/18)>>> | |