IRC Logs

Log of #ghostscript at irc.freenode.net.

Search:
 <<<Back 1 day (to 2015/05/27)20150528 
rayjj so, why does the cluster build ghostxps (gxps) if I specify filter=pbmraw (guessing, marcos hasn't taken it into account in the increasingly complex cluster scripts)?04:17.07 
  it's not really a big deal, but once marcos's machines go down (which slows the testing down a lot), it just is a wasted step04:18.16 
kens chrisl well done on finding that memory problem!07:40.40 
chrisl kens: it was a bit of a mission, that one. Although, once again, I wonder how it hasn't caused more problems......07:41.42 
kens Yeah, it does look like it kind of ought to07:41.56 
chrisl Well, we've had quite a few marcosw has mentioned that tiny differences on the command line make a problem appear and disappear, so maybe it has, and we don't know it07:42.55 
kens Hmm I guess that's possible07:43.10 
  Its a good one to get fixed anyway07:43.31 
chrisl Definitely. I'm quite relieved to put that one away07:44.29 
kens chrisl the readme you pointed me at is not really relevant now we have a signed ghostpdf.inf. You can just add the printer on Windows 8 now.07:47.15 
chrisl I was a bit on the fence about whether to update or to delete it. I sort of feel it worth changing the readme to say you need to "Add Printer", though07:48.31 
kens I can do that, but it means that practically all the text will be removed though, since its mostly not true with a signed file (I did try this yesterday).07:49.19 
  I'll change the text a bit and send it to you to see what you think07:49.52 
chrisl I just figured on copying what you have in the .inf comment into the readme, nothing more07:50.09 
kens Hmm, OK I cna do that easily enough07:50.23 
chrisl The reason being (and the reason I added it originally) is that I doubt many users of the .inf file would open it and read the comments07:50.29 
kens True07:50.36 
chrisl kens: that looks fine to me08:05.11 
  I mean, practically no one will read the readme, either, but at least we can just say "RTFM"....08:06.20 
  Out for a bit - back in a couple of hours.....08:14.31 
kens Robin_Watts : ping09:00.56 
Robin_Watts pong09:44.38 
kens I was thinking about adding info on signing the .inf files to the twiki, not sure where to do it. Do you think adding something under Ghostscript makes sense ?09:45.20 
Robin_Watts kens: Absolutely.09:48.42 
kens OK then I'll have a bash at that :-)09:48.56 
Robin_Watts http://twiki.ghostscript.com/do/view/Ghostscript/WebHome09:49.08 
  Start there.09:49.11 
kens Yeah that's where I am09:49.17 
Robin_Watts Choose 'Raw Edit' (I think Raw edit is easier for programmers than Edit is)09:49.31 
kens Hmm, OK I'll see where I get to in a few minutes09:49.47 
Robin_Watts Add a wiki word in the list under "RefactorForLanguages"09:50.02 
kens Yeah I was thinking of adding 'signing ghostpdf.inf for WIndows'09:50.45 
Robin_Watts kens: You need to pick a Wiki topic name for it. A WikiWord09:51.35 
kens Umm OK09:51.42 
Robin_Watts Like SigningDriver or something.09:51.52 
kens I got involved with something else, I'll get back to it shortly09:51.55 
  Robin_Watts : can you have a look at that please and tell me what style policies I may be violating :-)10:06.42 
Robin_Watts Looks good to me.10:08.50 
kens Ah, thanks10:08.58 
  coffee time....10:10.32 
  chrisl the website documentation looks good to me now, thanks!13:26.16 
chrisl kens: the script used to work - no idea why it stopped....13:26.41 
  I am surprised 8.71 didn't work, though13:27.14 
kens Well the installer defintely doesn't work13:27.23 
  Gives an error running setupgs.exe13:27.33 
  TO be frank, I really wan't that interested :-)13:27.44 
chrisl Of course, that was before I took over, so.....13:27.47 
kens I did check 8.60 on my WIndows 95 and it did work, so I assumed the reporter was correct.13:28.11 
  I wasn't going to ask why he was still running Windows 95 in case he told me13:29.05 
chrisl I've got to think it takes more effort, these days, to keep running Win95 than to use something a bit more modern13:29.25 
kens I presume he's running it in a VM of some kind, I can't believe he's running it on real hardware13:29.50 
chrisl Yeh, but still.... for what purpose?13:30.20 
kens Some ancient crufty piece of software which he finds vitally important for his workflow I guess......13:30.45 
chrisl In which case, keep using the ancient crufty version of Ghostscript13:31.15 
kens I was assumign he was freshly adding GS to the mix, though again, why rather escapes me....13:32.25 
chrisl 'Course, you can't get VS2003 these days13:34.12 
kens I can't say I'msurprised13:34.47 
chrisl I think you *can* still get it with an MSDN subscription13:35.35 
kens I really, really don't want it :-)13:35.49 
  Mind you, I don't wnat VS 2013 either. Ugly looking thing13:36.10 
chrisl I think anything after VS2008 looks pretty naff13:36.47 
kens Oh, and my community edition only has a 30 day trial licence, I thought ths was the equivalent of Express13:37.04 
chrisl Is it pre-release?13:37.29 
kens No, its the real deal13:37.39 
  Maybe bits stop working in 29 days13:38.05 
chrisl So, maybe you just have to register? I thought it was free for open source projects13:38.20 
kens Possibly that's why it keeps wanting me to 'sign in'. I thought that was just for collaboration, but perhaps not.13:38.53 
  TBH I don;t thnk I care if it stops working in 30 days, I really don't thnk I want to be forced to use it13:39.14 
chrisl Past experience says the command line utils keep working, so you'll have the bit(s) we need13:39.49 
kens Won't be a problem, I still have VS 200813:40.27 
chrisl I thought you needed 2013 for the signing13:41.10 
Robin_Watts chrisl: You need VS2013 to install the signing tools.13:42.11 
  but I believe the signing tools are independent once installed.13:42.21 
kens http://twiki.ghostscript.com/do/view/Ghostscript/SigningGhostpdf13:42.40 
henrys kens: without windows 8 or future support shouldn't we be pulling that stuff out of the tree (deprecating)?14:09.09 
kens It works on Windows 8, I tested it14:09.27 
  I have no idea if RedMon works, but the .inf file works fine14:09.43 
henrys kens: ha I was reading your commit diff backwards14:10.52 
  sorry14:10.56 
kens No problem, I can't recall what I said, maybe I wasn't clear14:11.11 
chrisl henrys: since when does obsolescence result in removal from Ghostscript??14:12.00 
kens Oh I see, its the comment change in the diff14:12.06 
  The old comment said 'does not support Windows 8 and beyond' thich is what I removed :-)14:12.35 
henrys kens: yes I read that as being added not removed14:12.55 
kens I cna see that would be confusing14:13.10 
henrys reading a diff, not really ;-) just got out of bed14:13.40 
chrisl I see I was being particularly careful with the "at your own risk" repetitions.....14:14.24 
kens Doing the .inf signing was something we discussed at the last staff meeting as a nice side benefit of getting a certificate.14:14.42 
  chrisl it was worth repeating....14:15.23 
chrisl I was acutely aware of the implications of saying "now, turn off all the security....."14:16.04 
kens Yeah, not really the best plan14:16.25 
henrys chrisl: I'm a bit baffled about the metrics, I looked at a few samples and don't see anything that would result in visible differences between PCL and PDF/PS. The design units are different but after accounting for they look very close. I am asking URW to look at it because they claim the metrics are incompatible.14:16.35 
chrisl henrys: was that a recent claim, or a historic one?14:17.23 
kens They may be thinking in terms of a .afm file, which is not the same as the metrics stored in a TT file. But we don't use the .afm metrics anyway14:17.50 
henrys kens: oh good point14:18.14 
kens I have to go out for a bit, back later (I hope)14:18.14 
henrys chrisl: recent.14:18.33 
chrisl I was thinking, with everyone converging on the Windows TTF fonts in recent years as the benchmark, I'd surprised if PCL and PS/PDF could get away with differing metrics14:18.55 
henrys chrisl: the response to that has been embed the fonts, but we have to pass quality logic14:20.25 
  chrisl: not really an issue with PXL or a lot of PDF where character positoning xyshow is common14:21.04 
  PXL is only xyshow actually14:21.16 
chrisl henrys: I guess that is true.... still, we see enough cases without the fonts embedded14:25.08 
henrys chrisl: one of the few things XPS got right ... get rid of the damn resident fonts14:26.11 
chrisl henrys: can't argue with that!14:28.00 
tor8 Robin_Watts: one last commit on tor/master before I leave14:28.11 
Robin_Watts tor8: looking...14:28.22 
  I give you a fortnight before you'll be hacking epub again :)14:28.48 
tor8 it's ... unfortunately too credible for me to bet against ;)14:29.34 
Robin_Watts looks plausible to me.14:29.52 
chrisl tor8: I'd think epub would be one of the things you'd not miss!14:30.10 
henrys tor8: have a great time, we'll miss you!14:30.29 
Robin_Watts yeah, have fun.14:32.29 
mvrhel bye tor8!14:36.19 
Robin_Watts Right, let's convert all those tabs to spaces...14:37.51 
henrys yes apply http://www.ghostscript.com/doc/9.16/C-style.htm to mupdf ;-)14:40.44 
tor8 thanks everyone!14:41.07 
  I'll be back in a couple of weeks...14:41.45 
  ...maybe sooner if my tabs have evaporated ;)14:41.54 
chrisl tor8: the staff meeting just won't be the same without your happy, smiling face!14:43.14 
sebras ha! I'll pester tor8 with lots of mupdf-questions during his sabbatical... ;)14:45.45 
tor8 chrisl: careful, if you yank too hard my leg might come off... ;)14:45.59 
chrisl tor8: okay. So, it won't be the same without you there to swap complaints with......14:47.11 
tor8 chrisl: that sounds more like it :)14:47.21 
  http://arstechnica.com/information-technology/2015/05/sourceforge-grabs-gimp-for-windows-account-wraps-installer-in-bundle-pushing-adware/ I hope we deleted our account there...14:47.29 
chrisl We haven't deleted it, but it is no longer used - for exactly that reason14:48.25 
rayjj wow, what do these customers expect -- they want us to test to make sure changes work ? ;-)18:10.56 
  cheeky18:11.05 
Robin_Watts rayjj: The mupdf question? Yes :)18:30.36 
rayjj now the HT tiling code patterns match, and pixel transition points match, but the transfer functions differ (at least if I add -c "{} settransfer" to the test cases, it matches18:40.01 
henrys Robin_Watts: hold on - no response for them right? did you see miles' last email?18:58.31 
Robin_Watts henrys: oh, crap.18:58.45 
  I forgot.18:58.48 
henrys Robin_Watts: don't worry about it.18:59.01 
Robin_Watts I sent them a binary for them to try.18:59.38 
  If they want more than that I'll check with Miles.18:59.57 
henrys Robin_Watts: good enough.19:00.09 
 Forward 1 day (to 2015/05/29)>>> 
ghostscript.com
Search: