| | <<<Back 1 day (to 2016/03/31) | 20160401 |
| chrisl | neverpanic: the checksums should be correct now.... there were a couple of Mac specific (hidden) files which caused Mac users to have problems unpacking them, so I re-uploaded. I generated new checksums, but must have failed to select them for upload. Sorry. | 08:35.47 |
| jnoetzel_ | Hi | 14:23.18 |
| ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 14:23.18 |
| rayjj | RobinWattsLenovo: I saw (on the logs) that you were up late. The file with the somewhat excessive RAM usage with mupdf, even with setting the "STORE" to 1, _was_ the j12 file | 16:39.53 |
| RobinWattsLenovo | No, I was up early :) | 16:40.14 |
| | Does J12 use images? | 16:40.28 |
| rayjj | RobinWattsLenovo: I was surprised that the "page by page" mode did not exactly correlate with what I saw with all pages | 16:40.44 |
| | RobinWattsLenovo: yes, some pages have images | 16:41.00 |
| RobinWattsLenovo | That may explain the lumps. | 16:41.09 |
| | MuPDF decodes images in their entirety. | 16:41.19 |
| rayjj | RobinWattsLenovo: but not why none of the pages in page by page mode are as big as the peaks when all pages are run | 16:42.07 |
| RobinWattsLenovo | rayjj: Suppose page 1 uses objects 1-100, page 2 uses 101-200, page 3 uses 201-300 | 16:42.44 |
| rayjj | The pages with images _are_ larger than the ones without | 16:42.49 |
| RobinWattsLenovo | decoding page 1, then page 2 will result in objects 1-200 being held in memory. | 16:43.13 |
| rayjj | does it keep the image data laying around for pages as it does pdf objects ? | 16:43.43 |
| RobinWattsLenovo | wheras decoding page 2 alone will only result in 101-200 | 16:43.54 |
| | Image data lives in the store, so it will be kept around if the store is large enough. | 16:44.08 |
| | With the store set to 1, it will not be kept around. | 16:44.18 |
| rayjj | RobinWattsLenovo: looking at what's laying around, it looks like the fonts might be the culprits (many are 200k) | 16:44.48 |
| RobinWattsLenovo | I have to go out now. Will double check stuff and get back to you. | 16:45.52 |
| rayjj | RobinWattsLenovo: no problem. This isn't a top priority | 16:49.05 |
| RobinWattsLenovo | rayjj: Yes, that's almost certainly it. | 18:04.04 |
| | I will ponder on a 'low memory' mode that ditches everything between pages. | 18:04.20 |
| sebras | https://www.humblebundle.com/indiefonts any takers on halfvetica? | 20:32.22 |
| malc_ | have bought helvetica lt com back in the day. worst... purchase... ever... | 20:47.03 |
| | Forward 1 day (to 2016/04/02)>>> | |
IRC Logs