| <<<Back 1 day (to 2014/07/13) | 2014/07/14 |
mattchz | oops. forgot to sign in here. | 11:14.38 |
| would it be possible to setup some kind of testflight@artifex.com mail alias? | 11:14.55 |
pedro_pc | hi folks | 12:50.11 |
mattchz | hi pedlo | 12:53.11 |
Robin_Watts | pedro_pc: You back from hols now ? | 13:47.11 |
marcosw | mattchz: I can setup a testflight@ alias. To whom shall it forward? | 14:14.30 |
mattchz | marcosw> thanks | 14:14.50 |
| if you could set it to forward to matt@emobix.co.uk for now, thatâd be great, thanks | 14:15.02 |
marcosw | I'll set it up and send a test message. | 14:15.37 |
mattchz | marcosw> works thanks, but you used the alias testdrive rather than testflight - not taht it really matters | 14:34.04 |
Robin_Watts | tor8: (or anyone else who knows)... jbig2globals... | 14:38.57 |
| is there any easy way to put globals back into the image? | 14:39.22 |
| i.e. to combine globals and image data together into a single stream? | 14:39.39 |
| Can they simply be concatenated? | 14:39.47 |
mattchz | marcosw: actually, is it possible to change to testflight please, just to avoid confusion? | 14:53.16 |
Robin_Watts | marcosw: Can we have a 'PDF-Export' component for GhostDocs please? | 17:05.51 |
| (in bugzilla) | 17:05.55 |
| and probably, a "Renderer" component too ? | 17:06.20 |
henrys | mvrhel_laptop: I did look at xpswrite performance a bit, and going with paths for text results in 1 meg per page ⦠thatâs much better than a 600 dpi contone bitmap out of the old gsview. But if we are really entertaining printing the plrm I should just print one page at a time. There is no way we want to buffer that much no matter how much it is compressed | 17:16.41 |
henrys | will never be convinced printing with a non streamable format is anything short of nonsense. | 17:17.36 |
Robin_Watts | It does seem to be saddling yourself with a whole heap of problems, yes. | 17:23.54 |
mvrhel_laptop | henrys: 1 meg per page sounds like it might work | 17:27.16 |
| henrys: the way I have gsview set up to work right now is that the xps output from gs is stored in a temp file and then I do all the print stuff required to take the xps to the windows print pipeline | 17:28.35 |
| I need to double check the code though | 17:29.14 |
| to see how I would handle 1 page conversion at a time | 17:31.07 |
henrys | mvrhel_laptop: I do wonder if you shouldnât run it in the background a page at a time using FirstPage and LastPage to go through, avoiding a long startup delay | 17:31.21 |
mvrhel_laptop | I am going to try to wrap up this SOT stuff first though | 17:31.57 |
henrys | mvrhel_laptop: okay, chrisl_away also noted it was slow and Iâll look at that also I have some questions about this âTrioâ printf stuff we added. That seems very pokey | 17:33.18 |
Robin_Watts | henrys: The problem is that printf is (stupidly) locale specific. | 17:33.40 |
| Hence either we need to change locale while we run (which plays badly when we are a lib or a DLL), or we need to provide our own non-locale respecting functions. | 17:35.29 |
henrys | Robin_Watts: right I know why it was added I wanted to see if we knew about the performance problem. | 17:35.31 |
Robin_Watts | oh, right, I wasn't aware of such problems, sorry. | 17:36.19 |
mvrhel_laptop | henrys: I am running it async. I did everything I could to get it on a back ground thread but there were issues. I will revisit this. | 17:46.41 |
| henrys: I may ping my contact at MS about this | 17:48.34 |
rayjj | henrys: mvrhel_laptop: Note that -dFirstPage -dLastPage is NOT good for printing individual pages from a large PDF. It close and re-opens the PDF each time. Better is the runpdfbegin .. First# Last# dopdfpages method | 19:47.28 |
| henrys: mvrhel_laptop: that leaves the scanned (possibly repaired) PDF in VM and puts out the selected page (range if First# and Last# are different) | 19:48.21 |
henrys | rayjj, mvrhel_laptop: however you want to do it, but the dramatic slowness chrisl discovered is to do with the management of many pages in the device. Using Firstpage and LastPage I see .5 sec and 1 Meg per page for the PLRM. If you want to generate many page jobs and send that to a printer xpswrite will need some changes. | 21:52.12 |
| Forward 1 day (to 2014/07/15)>>> | |