IRC Logs

Log of #ghostscript at irc.freenode.net.

Search:
 <<<Back 1 day (to 2013/03/13)2013/03/14 
ray_laptop rather disappointing that cust 393 is now TERM. I sure hope that the improvement to the product justifies the time that Michael spent. The time that Marcos spent duplicating and entering their 21 bugs, and Chris spent making binaries for them is clearly "sunk costs".00:23.35 
mvrhel_laptop ray_laptop: I was disappointed to see that too. I do think we have something better now though01:18.50 
  It is going to be interesting to see if they run across any other issues. I think the cost for them should go up01:19.24 
jakllsch you folks are a bunch of horrible people.02:57.43 
  I can't even print without draconian licenesing restrictions now.02:58.03 
  FUCK YOU!02:58.11 
robin_watts_mac chrisl, mvrhel_laptop: I spoke to Marti about my outstanding lcms patches before the meeting thinking that they'd get mentioned at the meeting, and then forgot.03:24.08 
  He wasn't going to take them as was, and he proposed a way that they could be done better and 'independently' of core lcms. I liked his idea, but said that it would still be a shame not to see them as part of lcms. It seems from his latest reply that he may be reworking them into a form he likes, which is nice.03:25.30 
  WTF is jakllsch on about?03:28.26 
henrys I have no idea he left before anyone could say anything.03:41.01 
  robin_watts_mac: coeds in thongs?03:41.29 
robin_watts_mac henrys: Very short dresses, certainly.03:47.56 
  bedtime!03:48.17 
henrys good night03:48.25 
kens Well, upgraded anti-virus software from vewrsion 7 to 8 and so far my mail client won't work (SSL problems0 and it flase positives on mkromfs and won't let me run it. Excellent08:37.05 
  Just going to update Miranda, brb08:38.43 
docampo Hello, I would like to chat with a MUPDF developer if there is someone here. 13:16.12 
kens tor8 ping13:16.20 
tor8 docampo: I'm around13:16.43 
kens chrisl just found this in an EPS file from Adobe:13:38.16 
  :-O13:38.16 
  "/HqnSpots /ProcSet resourcestatus"13:38.24 
chrisl kens: oops, I wonder how that happened......13:43.58 
kens Its also a Illustrator file with broken logic I think. Fortunately from my POV Distiller does the same as we do :-)13:44.30 
chrisl :-)13:45.06 
henrys kens:I thought marcos said distiller worked, or am I missing something.14:08.38 
kens henrys it doesn't for me14:08.48 
  But it depends how you have it sert up14:08.54 
henrys I see14:09.02 
kens If its not preserving colour (or overprint) thenit will 'work'14:09.20 
  Obviously they could set the GS distillerparams also (I mentioed that in part in my response)14:10.10 
  FWIW the 'Press Quality' setting son Distiller produce the same output as GS14:11.17 
  THe 'Stamdard' settings produce a PDF file that looks like our output14:11.53 
  Err, our raster output...14:12.05 
  Ah, standard settings convert all colours to sRGB, so the colour conversion strategy is not 'LeaveColorsUnchanged' and so the spots are converted to process, and the section of marking operations is not skipped14:13.26 
henrys mvrhel_laptop:does word for windows create "real" XPS - my mac version seems to create XML variants but not the regular xps?14:50.52 
mvrhel_laptop henrys: yes it creates real xps files15:05.51 
henrys okay thanks.15:06.19 
kens Is that the MS XPS files or the ISO ones ?15:06.26 
mvrhel_laptop windows 8 can create either15:06.39 
kens OK15:06.56 
henrys from what I can gather MS is pushing folks to XAML - then XAML -> XPS.15:10.24 
mvrhel_laptop bbiab15:23.48 
henrys marcosw:I should have looked at ram before ordering the SSD - 8 G I think 32 G would have been a better upgrade cluster-wise15:33.02 
  the apple cheapskates gave me 4 2 G simms15:33.37 
marcosw henrys: I upgraded my MacPro from 8G to 16G and didn't see any increase in performance cluster wise. We have cluster nodes that only have 4G of memory and they do okay, so I don't think memory is an issue.15:34.44 
henrys well we'll see what the SSD does first it should be here soon.15:36.13 
marcosw I may have to change the number of jobs that are run simultaneously before we see an improvement. Currently it's 16, which make sense for all the other nodes, since they have at most a 8 cores available. When you initially got the MaPro I tested and there was no improvement over 16 jobs, but presumably that was due to disk bandwidth limitations.15:39.46 
henrys mvrhel_laptop: when you get back … I want to offer the customer 6 months of free support to make up for the bugs they encountered, thoughts? others?15:50.36 
  marcosw do you think that reasonable?15:53.08 
marcosw henrys: sorry, I haven't been paying attention. Which customer are we talking about?15:56.18 
henrys pretty long thread last night about it.15:56.55 
  probably best to read that first.15:58.04 
marcosw I read that, but I didn't think they were a customer anymore. I'm assuming they've moved on and are no longer using our software. I think offering free support at this point is sort of like suggesting marriage counseling after the divorce.15:59.58 
  Not that people don't remarry their ex...16:01.52 
henrys hey I remarried my X16:02.10 
  but that didn't go well16:02.17 
  marcosw:but yes you are probably right.16:03.30 
  OTOH there is no effort in giving it a shot.16:03.55 
  chrisl:692722, I know gp_fopen is not necessarily needed for unicode in many places - internal postscript, clist whatever else, but wouldn't it be better to simply #undef fopen and force everything through gp_fopen16:07.33 
  ?16:08.08 
chrisl henrys: eh? How does that help error messages?16:08.39 
henrys it doesn't but the file won't open16:09.36 
  see comment 716:10.02 
chrisl henrys: well, why is anywhere using fopen() directly anyway?16:11.12 
henrys my point exactly, why haven't you fixed that? ;-)16:11.47 
  I don't know if there is ever a case where we don't want the unicode behavior but I can't think of a reason.16:12.29 
chrisl Personally, I'm not a fan of silently replacing calls like that - I'd rather we just fixed the places that are wrong16:13.14 
henrys okay definitely lots of fopen's in the code.16:13.49 
  chrisl:I didn't mean silence it, I meant undefine fopen so the code does not compile. But the number of fopen's looks daunting.16:16.00 
chrisl henrys: I think undefining fopen() will only work on platforms where fopen is a macro, if it's defined directly as a function, that won't work16:17.26 
  henrys: I would suggest opening a bug about fopen() use in the code, assign it to me, and I can pass it on when I get sick of it16:19.25 
henrys probably right but #define fopen DONTUSE is okay16:19.49 
  will do16:20.03 
malc_ there's something fishy in ghostpdl.pdf, page 3 last word in penultimate line (probably has something to do with escaped periods in ghostpdl.tex)16:25.29 
chrisl henrys: yes, #defining fopen like that will work. If you could open the bug - I'm happy to work on it, but I have head out in a few minutes16:26.46 
henrys chrisl:done16:31.33 
chrisl henrys: thanks, I'll make a start tomorrow16:31.54 
  bye!16:32.31 
henrys malc_:thanks I'll have a look in a few minutes16:33.19 
malc_ henrys: welcome16:35.19 
kens2 goodnight folks17:14.15 
ray_laptop Oh, great. my virus program (avast) thinks that "mkromfs.exe" is infected with a virus. Note that I reported a "false positive", but had to figure out how to whitelist it :-(17:30.50 
  I doubt it is infected since I just did a clean build and am using VS 2008 (unless VS is infected), but it seems like EVERY executable built by VS should be problematic now17:31.59 
  found an infinite loop in the GC. Somehow I end up pointing into an imem object instead of to a valid header :-(18:15.10 
  this is with my changes, of course :-/18:15.45 
mvrhel_laptop henrys: I have mixed emotions about giving them free support. If they are still using our software the only thing that would entice them to get a support contract would be if they needed support. If they are not using our software then they don't need support. 18:42.58 
  18:42.59 
henrys mvhrel_laptop:I can see that but I do feel they were sort of "guinea pigged" and we owe them some sort of explanation. it's a lot of bugs they found. But we'll go with what you think, let's decide Tuesday18:47.41 
  as marcosw said it won't help get them back.18:48.03 
  probably18:48.24 
  ray_laptop:math was asking about his 690969 contribution.18:49.11 
  lunch18:49.52 
mvrhel_laptop bbiaw21:39.01 
 Forward 1 day (to 2013/03/15)>>> 
ghostscript.com
Search: