| <<<Back 1 day (to 2017/09/20) | 20170921 |
chw | good morning | 07:29.03 |
kens | Morning | 07:29.08 |
chw | i've just recompiled and retested the pdfs | 07:29.20 |
kens | Hopefully improved ? | 07:29.29 |
chw | neither acrobat nor verapdf complain :) | 07:29.38 |
kens | Well that sounds better... | 07:29.45 |
chw | yes, it's fixed as far as i'm concerned | 07:30.06 |
kens | OK thanks for letting me know, useful to get the update | 07:30.22 |
chw | the only thing left would be pdf_info.ps, but i don't think that is a priority | 07:30.29 |
kens | Ah, I had rather forgotten about that | 07:30.43 |
chw | it still failes on the file gs generates... | 07:30.46 |
kens | I should probably look at it | 07:30.53 |
| I bet its a different problem though | 07:31.07 |
chw | semms like it | 07:31.26 |
| seems* | 07:31.31 |
kens | I have to write a (lengthy) n email, I'll look at iit after that | 07:31.57 |
chw | not a problem, it's just an option after all | 07:33.36 |
kens | Yeah, but its bad that it doesn''t work with our own output | 07:33.51 |
chw | yes, kinda weird too. i would have thought it was the encoding mismatch, but apparently it's something else entirely | 07:35.35 |
kens | The XMP boilerplate is differently constructed, its probably overflowing a string or something | 07:36.01 |
chw | that could be it. it dies with an error in readline | 07:39.27 |
kens | Its just a theory, I'll have to debug the PostScript | 07:39.44 |
chw | sure. the error appears to occur right after the documnetid tag at the start of the description tag. | 07:42.10 |
deekej | hello chrisl :) do you have any ETA on when the 9.22 will be released? | 11:13.23 |
kens | We'll need to do another release candidate | 11:13.38 |
| So at least another week | 11:13.43 |
deekej | ah, okay :-/ | 11:14.17 |
| because fontconfig upstream is pushing on me to have the release of URW fonts ready | 11:14.37 |
kens | Some problems turned up, but that's what the release candidate is for :-) | 11:14.38 |
| Sorry, you'll just have to tell them its going to be some time yet | 11:14.59 |
deekej | okay | 11:15.18 |
kens | There's still at least 2 problems to be reoslved, one of which hasn't even been looked at yet | 11:15.24 |
| And the ones which have been fixed caused differences in approximately 1/3 of our test suite, so we'll need to look at all those outputs again | 11:15.56 |
| anbd that takes days, even if there are no actual problems | 11:16.06 |
chrisl | deekej: If they get really nasty, I *could* do the font release early - but I'd very much rather have the releases happen together | 11:16.54 |
deekej | chrisl: I understand. I will ask them if they could wait (we need to sync the releases for now) | 11:17.24 |
| chrisl: okay, so fontconfig upstream can't wait with the release in Fedora, since it contains other bug fix changes. | 12:03.34 |
| chrisl: would it be possible to realease the fonts sooner, please? Otherwise the fontconfig might revert the changes that were accepted before. | 12:05.12 |
chrisl | deekej: I'll do it tomorrow | 12:06.01 |
deekej | chrisl: okay, thank you! | 12:06.16 |
norbertj | hello, just a quick question. Can pdfwrite device also generate multiple pdfs for 1 job, i.e. split on e.g. 5-10 page boundaries? | 14:12.05 |
kens | You can produce one file per page | 14:12.23 |
| You can use FirstPage and LastPage to produce a single set | 14:12.34 |
| You can use the Pages command to produce a single set of arbitrary pages | 14:12.54 |
| You can't choose where to split the PDF file otherwise | 14:13.05 |
| You would have to modify the pdfwrite device | 14:13.12 |
| Its probably technically possible by starting with a device other than pdfwrite, execute a save, load teh pdfwrite device. Set a custom EndPage. If you hit the page count, restore to the old device, then set pdfwrite again as the device. | 14:14.23 |
| Though thinking about it, I'm not sure that would work with the EndPage procedure | 14:14.43 |
norbertj | so when x pages are produced, then I want to tell the pdfwrite device to close the pdf and start a new pdf (with same resources like collected subsets of fonts etc) | 14:15.25 |
kens | You *definitely* can't tell it to keep the resoruces stored in the first PDF file produced, if that's what you mean. It will maintain the existing state in the interpreter, so any fonts loaded will still be present, and will be emitted (if required) in the second PDF file | 14:16.33 |
| But you don;t do it by telling pdfwrite to close the PDF file. | 14:16.45 |
| You do it byb unloading the device entirely, and starting a new one | 14:16.59 |
| It would probably be easier to modify pdfwrite. | 14:17.17 |
norbertj | That's what I was thinking of. Mod the pdfwrite, but was just wondering whether the state would be preserved. You just confirmed this. So thanks again. | 14:18.30 |
kens | NP | 14:18.36 |
norbertj | bye\ | 14:19.47 |
| Forward 1 day (to 2017/09/22)>>> | |