IRC Logs

Log of #ghostscript at irc.freenode.net.

Search:
 <<<Back 1 day (to 2012/09/05)2012/09/06 
mvrhel finally. the chili label renders correctly Gigs02:56.20 
  that was a tough one02:56.27 
  need to see if anything breaks with the fix02:56.35 
alexcher mvrhel: I'm back.03:46.11 
henrys mvrhel:ever heard of a north michigan wine made from tart cherries, Sabrina brough back some, pretty good.04:05.57 
  s/brough/brought04:06.11 
mvrhel henrys: I know they grow a lot of cherries up there and I have had a cherry stout beer from michigan but no wine04:18.01 
  ugh. why can't Acrobat open the file tests_private.comparefiles.Tesfform.v1.0.2.pdf04:21.28 
  ugh. why can't Acrobat open the file tests_private.comparefiles.Testform.v1.0.2.pdf04:21.41 
  my fix for bug 693018 shows some odd diffs with this file and I can't even look at it to know what it should be04:22.42 
  since that is the only file I am pushing forward with the fix04:23.23 
  henrys: ok. I will now get back to the customers issue04:25.32 
alexcher mvrhel: what was your question?04:31.36 
mvrhel alexcher: never mind. I was able to figure out what I needed05:11.06 
alexcher crypt_sha2.c06:35.39 
Robin_Watts tor8: ping09:47.20 
tor8 Robin_Watts: hi09:47.27 
Robin_Watts Hi.09:47.32 
  I updated the commits on robin/master09:47.45 
  Ignore the latest one for now if you want (it's a work in progess, but any thoughts etc you have would be appreciated), but the 2 before that are ready to go.09:48.23 
  no hurry.09:48.49 
tor8 Robin_Watts: why keep pdf_image_params as a type?09:50.10 
  the drop commit LGTM09:50.19 
Guest78441 hi everyone, I need help with the compilation of mupdf, can anyone help out?09:50.41 
Robin_Watts pdf_image_params has a colorspace in too, IIRC.09:50.51 
  Guest78441: We can try. What platform ?09:51.01 
tor8 Robin_Watts: yeah, why is that not part of fz_image_params?09:51.14 
Guest78441 I have the same error under Windows / cygwin and ubuntu09:51.19 
Robin_Watts because we don't need it in the fz_ case.09:51.33 
  Guest78441: OK. and that error is? (If it's a small paste, here is fine, otherwise, drop the whole lot in a pastebin and paste the URL here)09:52.46 
Guest78441 it starts like this: 09:53.00 
  SharedLibrary : libmupdf.so09:53.01 
  /home/flip/proj/mupdf/android/obj/local/armeabi/libmupdfcore.a(pdf_xref.o): In function `pdf_init_document':09:53.01 
  /home/flip/proj/mupdf/android/jni/../../pdf/pdf_xref.c:784: undefined reference to `pdf_new_js'09:53.01 
Robin_Watts OK, what make command are you using?09:53.23 
Guest78441 I mention that I have run "make" successfully, and that I'm running ndk-build from android folder09:53.31 
tor8 oh. did we forget to "fix" the android makefile for the forms branch?09:53.33 
Guest78441 I had some problems at make with some X11 extensions, but I got that fixed09:54.05 
Robin_Watts Guest78441: OK. We have some alpha code in there for dealing with forms in pdf files.09:54.38 
tor8 Guest78441: if you are building for android, you shouldn't be seeing any X11 stuff at all09:54.46 
Guest78441 I've tried with both NDK r8b and r6b, as the readme states, but I haven't manage to downgrade my sdk09:55.00 
  yeah, I got that during the "make" step09:55.11 
Robin_Watts That got merged to master the other day. It is possible that we've broken the android makefile.09:55.26 
tor8 tries an android build on the latest master09:55.51 
Guest78441 even with the X11 errors, I still got the "generated" folder, so it's a side issue09:55.51 
Robin_Watts Right. As the instructions say you don't need the make step to complete - just to get far enough to make the generated directory.09:55.54 
  right.09:55.59 
tor8 Guest78441: oh, right. to get the "generated" folder. "make pregen" should make that without hitting the x11 stuff.09:56.16 
  and yes, we've broken the android makefile09:56.21 
  sorry, "make generate"09:56.37 
Guest78441 so, the code that deals with forms is breaking the compilation process?09:56.40 
tor8 Guest78441: yeah, android/jni/Core.mk is out of date09:58.33 
  it's missing a few new files09:58.46 
  pdf_form.c, pdf_js_none.c and pdf_write.c should be added09:59.39 
Guest78441 so, I should git pull the project again - when is it safe to do that?09:59.40 
tor8 Robin_Watts: check tor/master10:00.27 
  Guest78441: a few minutes, hopefully, once Robin_Watts has vetted my commit we can push it to master10:00.43 
Guest78441 thank you, I'll leave for lunch, and check it when I return, then I'll let you know the outcome :)10:01.38 
tor8 Robin_Watts: I need to go for lunch, so push if it looks good and I'll see you later10:02.08 
Robin_Watts tor8: looking now. will do.10:03.15 
  Guest78441: New version pushed. Does that solve it for you?10:07.58 
Guest78441 great, that solves it!10:56.57 
  thank you very much, I'll see if the next steps are good10:57.29 
  everything went good.11:05.19 
Mickey_ Hello everybody14:13.28 
  I'm using GhostScript to convert some pdfs to Jpegs, but when using a file of 87MB, I get an error14:15.45 
kens and the error is ? What version of GS ?14:16.21 
Mickey_ GPL Ghostscript 9.04 : Unrecoverable error, exit code 114:16.22 
kens 1st recommendation, try 9.0614:16.45 
Mickey_ Btw I'm using a commandline tool (pdf2toimage) that makes use of gs14:16.57 
kens OK 2nd recommendation, sdon't do that14:17.12 
Robin_Watts Mickey_: Drop the WHOLE error output into a pastebin - there must be more than that.14:17.17 
kens use gs directly, because we'll bneed the command line too14:17.23 
Robin_Watts but kens advice is good.14:17.31 
Mickey_ Ok, please give me a minute14:18.49 
  Thanks! :D14:18.58 
  I've never used gs before, so I'm still figuring it out14:34.39 
  Only used wrappers so far14:34.43 
  Since I only need to convert pdfs to images14:36.04 
sebras Mickey_: http://www.terminally-incoherent.com/blog/2007/04/23/convert-ps-and-eps-images-to-jpeg/ I think this might be a good start for you.14:38.54 
Robin_Watts Presumably pdf2Image is using a command line like: gs -sDEVICE=jpeg -r200 -o out%d.jpg in.pdf but the devil is in the details.14:39.05 
paulgardiner Robin_Watts: I may have misunderstood the cluster commands. I thought bmpcmp worked on the basis of the last mujstest, independently of the current state of my local repo. Is that right?14:40.16 
Robin_Watts no.14:40.48 
paulgardiner Ah!14:41.07 
Robin_Watts The clusterpush command always uploads your current code.14:41.08 
  and builds that. If you ask it to bmpcmp it runs the files that differed in the last test through both the new binary and the 'last HEAD binary'.14:42.00 
paulgardiner I see. ok.14:42.31 
Robin_Watts If you ask it to test, it runs all the files and compares their md5 sums.14:42.41 
  I might possibly have done it differently if I was building from scratch, but that's the way it fell out, and I've never felt strongly enough about it to want to open pandoras box to fix it.14:43.21 
paulgardiner So I should ensure my repo is in the same state as the last test when issuing bmpcmp?14:43.51 
Robin_Watts Yes.14:45.08 
  or at least be prepared for the consequences if it isn't.14:45.35 
Mickey_ Thanks sebras !14:45.44 
paulgardiner Presumably a bmpcmp doesn't change the last-test info. So I can still issue a correct bmpcmp now, without rerunning mujstest14:46.04 
Robin_Watts indeed.14:46.10 
paulgardiner Magic14:46.15 
chrisl Robin_Watts: I was going to bring your Byrds boxset to the staff meeting to give back, but it occurs to me that if you're getting your camera from Marcos, you might not want the extra bulk of the CDs. What do you think?14:47.09 
Robin_Watts chrisl: Urm... I'm in no hurry for it.14:47.31 
  so shall we wait til next time?14:47.38 
  It's only going to take space on a shelf :)14:48.02 
chrisl Yep, no problem - I wouldn't want to tip you over into "real luggage" territory14:48.10 
Robin_Watts I think when I fly back I will be in real luggage territory anyway.14:48.25 
chrisl As long as it's not my fault......14:49.19 
Robin_Watts No, it's my inability to control my spending on toys :)14:49.47 
sebras Robin_Watts: reading engadget too much..?14:50.33 
chrisl That's better than spending on cars :-(14:50.40 
Robin_Watts sebras: Nah, I've been waiting a long time for this camera. And curse electronics for being so much cheaper in the states...14:51.13 
  chrisl: I've just run the MOT gauntlet for this year. Failed on the handbrake.14:51.35 
  Only 250quid for brakes/pads, so I guess that's par for the course.14:52.04 
chrisl Robin_Watts: Ah, I finally got my 924 back last Friday, and realised earlier in the week that it's misfiring at higher revs, under power - feel like torching it......14:52.53 
Robin_Watts chrisl: Well, clearly the solution is drive it at low revs all the time :)14:53.33 
chrisl A solution that had already occurred to me - but rather defeats the point of owning a car like that, I reckon!14:54.07 
Robin_Watts I thought the purpose of classic porsches was to keep porsche dealers in ski holidays and shiny suits? :)14:55.02 
chrisl And that's why I avoid Porsche dealers......14:55.34 
  That and the fact there isn't one near here14:56.14 
Robin_Watts Yeah. Find a local garage that's busy enough that they don't need to 'find stuff that's wrong', but not so busy that they rush stuff.14:57.29 
  Or do like me, and luck into moving next door to a garage owner that needs his computers fixed on a regular basis :)14:58.04 
paulgardiner chrisl: I had that fault (in not quite such an impressive car). Was predicted to be the crank sensor.14:58.10 
chrisl paulgardiner: the crank sensors on the 924 generally work okay, or not at all - it is a possibility, but unlikely. Also, the revs where it misfires vary (it seems) on how long the car's been running. I suspect fuel pump.15:00.03 
paulgardiner Yes, makes sense... and I've just remembered in my case it was predicted to be the crank sensor because the ECU said "It's the crank sensor" :-)15:01.18 
  Could even be a slightly blocked filter, couldn't it?15:01.52 
Robin_Watts Yeah, I had a car once that had an inline fuel filter. It would run perfectly for a while, but when you were using lots of fuel (long runs, high speed) would start to sputter and die as the fuel ran short.15:02.56 
  Pull over, wait 20 minutes, ran fine again.15:03.08 
chrisl Yes, it could be a filter. The reason I'm leaning towards the pump is a worn out pump will overheat, and lose efficiency as you run it, so would explain the varying behaviour with the running time.15:03.16 
Robin_Watts This particular model of car had 2 fuel filters, one in the engine, one by the tank. No one other than a specialist dealer ever knew about the second one.15:04.08 
chrisl They often have a large, coarse filter near the tank, and then a fine filter near engine - the fine one you change often, the coarse one much less often15:05.25 
Robin_Watts Right. It was the coarse filter that was getting clogged.15:06.08 
  I'd get that checked as it sounds like it would explain your problem too (and it's likely to be a lot cheaper than a pump).15:06.35 
chrisl Yeh, well, the car is overdue a service (which will be both fuel filters), so I'll take it back into the garage after the staff meeting - I can't be bothered to work on it myself at this stage....15:07.21 
  The only thing that really bugs me (other than having been without the car since March!) is that I really wanted to use it to drive to Heathrow, but I don't want to risk it getting worse en-route15:09.47 
Robin_Watts Gah. I hate PDF text.15:13.08 
  chrisl: Yes. Unreliability is the curse of old cars.15:16.45 
chrisl Robin_Watts: No, unreliability is the curse on the owners of old cars ;-)15:17.25 
Robin_Watts But you're not unreliable...15:17.49 
chrisl I could be, if it would help.....15:18.18 
Robin_Watts http://ghostscript.com/~robin/out.pdf Woo Noo!15:48.07 
  Woo Hoo! even.15:48.11 
kens rewritten ?15:48.33 
  text looks 'poor'15:48.40 
Robin_Watts It's absolutely perfect.15:48.49 
  apart from all the fonts being Helvetica.15:48.55 
kens that would do it15:49.01 
Robin_Watts And the encoding being assumed to be ascii.15:49.02 
confusedDev hi, guys! I need some help with opening a PDF document in android, do you have some time?15:49.05 
Robin_Watts confusedDev: Ask away...15:49.17 
  kens: basically, it's the crappest possible implementation to get me *something* on screen that looks reasonable.15:50.05 
kens Robin_Watts : fair enough15:50.38 
  Hmm, looks like my page hint table is badly messed up. The Header look sOK though15:51.47 
confusedDev so, what I did is I took MuPDFCore, MuPDFPageView, PageView and all the classes referenced by them, and moved them, along with the .so, in another project. There, I am loading a pdf file by calling mCore = new MuPDFCore("/mnt/sdcard/Download/test.pdf");. The log tells me that the file was opened correctly, but, when I try to render it into a bitmap with drawPage(), when calling gotoPage, it has numPages = -1. What am I doing15:53.02 
kens Hmm page 1, first objetc number 262144, I don't think so somehow....15:53.18 
Robin_Watts confusedDev: OK, just so you know, the internals of the Android app are not a public interface.15:54.01 
  We reserve the right to change them at any time.15:54.11 
  The C level interface is fixed; the classes that wrap those are considered internal.15:54.42 
  That's not to say that you can't work that way, but don't be surprised if changes have to be made in future versions.15:55.03 
confusedDev yes, I've tried a quick fix, so I can see that it works at a more basic level than the sample. And, sadly, it doesn't.15:55.32 
Robin_Watts so MuPDFActivity is our bit of code that calls MuPDFCore.15:56.30 
confusedDev yes15:56.44 
Robin_Watts We call new MuPDFCore(filename) to get a core.15:57.10 
paulgardiner Robin_Watts, confusedDev: It might be just that you need to call countPages before gotoPage15:58.46 
Robin_Watts Right. Then we call countPages.15:59.07 
  I think paulgardiner may be right.15:59.17 
confusedDev thank you, that worked. I haven't figured that it might not be called anywere. You've saved me again! :)16:04.43 
kens OK off for the night, bye all16:08.11 
ray_laptop bye, kens16:09.26 
  Robin_Watts: got a moment for a git question ?16:20.00 
Robin_Watts ray_laptop: Sure.16:22.58 
ray_laptop Robin_Watts: sorry had a call... Any idea how I can get a bunch of 'M' files listed AFTER I do a git co my_branch16:28.50 
Robin_Watts Ok. I *think* that when you 'checkout' it only replaces the files in the working set that have changed between the old and the new branch.16:30.19 
ray_laptop Usually I see those when it tells me I CAN'T switch branches16:30.24 
  so to get my files back to my branch I do a reset --hard to that branch ?16:31.07 
Robin_Watts So if I have a repo that has files a,b,c,d,e and only a and b have change between branch1 and branch2, if I'm on branch1 and git checkout to branch2, with modifications to c,d, or e, the checkout will complete without a problem.16:31.18 
  ray_laptop: That's one way, yes.16:31.23 
  In general, you should be very careful when using git reset --hard, as it's one of the few git commands that can actually destroy data without you having a way of getting it back.16:32.03 
ray_laptop Robin_Watts: OK. I'll just try that. It un-did a bunch of my changes due to a conflict when I pulled in changes from master16:32.25 
Robin_Watts git stash/git stash drop is safer I reckon.16:32.27 
  cos it leaves a record of those files in gits history that you can get back if you ever need to.16:32.55 
mvrhel brb16:32.57 
marcosw chrisl: I'm in the process of making a change to the cluster code. I think it's all working, but if your clusterpush job fails strangely it's my fault.16:33.15 
chrisl marcosw: okay, thanks. Although I actually expect this one to behave strangly....16:33.57 
Robin_Watts Roll up, Roll up. Today for 1 day only... it'll be Marcos breaking the cluster rather than me!16:33.58 
henrys that's how I first read his sentence the first time "strangely it's my fault" - not Robin's ...16:35.46 
  dangling modifier or something like that.16:36.06 
Robin_Watts :)16:36.21 
chrisl I think you can get an ointment for dangling modifiers nowadays........16:37.08 
henrys wow the agenda is taking a long time because there really was a large volume of work this time, commits analysis, very productive few months here - (I know it is a longer than usual period)16:38.47 
  but it should be ready in the next few hours or so.16:39.32 
Robin_Watts It's clearly been too long between meetings, because I've had to lookup some of the agenda items to remember what on earth I supposedly did on them.16:39.56 
henrys don't let tor8 hear you say that.16:40.26 
chrisl Hmm, a long a agenda means a l..o..n..g meeting - I better make sure to pack the *strong* headache pills........16:42.39 
Robin_Watts It may just be that I need a memory expansion.16:43.38 
sebras henrys: I just signed a copyright assignment for tor8 to bring to the meeting. I'd better not be an exception or cause more anyone here more grief... ;)16:44.50 
henrys Hmm, Miles agreed to make a new assignment so I'm afraid there will be another one to sign.16:46.57 
Robin_Watts sebras: I think it's fair to say that the benefit we gain from you far exceeds any grief you may cause :)16:47.00 
henrys sebras:you seem to be griefless as far as I can tell.16:47.51 
  chrisl:not being married you haven't learned the skill of not listening. That can be very stressfull and headache inducing for a guy.16:50.38 
chrisl henrys: I do get to practice no listening to my mother - maybe I'll just bring my noise cancelling headphones to the meeting room :-)16:52.00 
Robin_Watts chrisl: Kens usually buys strong headache pills in the states to bring back - they have better drugs than us.16:52.55 
chrisl Good point!16:54.16 
ray_laptop Robin_Watts: BTW, thanks for clarifying it. I did the stash and my files are back where I want them. I have _NO_ idea how they got messed up, but it was probably VS which was open while I switched to 'master' and back to my branch. Safe to blame Microsoft :-)16:55.23 
Robin_Watts ray_laptop: fab.16:56.17 
  Git can be confusing in that it checks out through the index.16:56.37 
  and git diff shows differences between your working set and the index, rather than the working set and the current head.16:57.09 
  I don't think that's at play here, but it's something else to bear in mind.16:57.37 
henrys chrisl:actually I don't know that it will be a bad meeting. Most of the stuff I'm plowing through is fixed and won't need to be discussed.16:58.58 
chrisl henrys: well, that's good!16:59.51 
ray_laptop cust 532 has been quite since the last fix I sent them. Either they haven't tested, or it worked and they haven't bothered to let me know (hopefully the latter)17:00.37 
Robin_Watts ray_laptop: On tuesday they sent an email saying "7 regressions left"17:01.44 
ray_laptop Robin_Watts: really ? I missed that. I wonder where that went. Thanks.17:02.08 
Robin_Watts Dennis Fan: 21:50 to you CCd to support.17:02.28 
mvrhel_laptop oh did I miss ray17:13.32 
Robin_Watts mvrhel_laptop: He may still be here.17:14.11 
mvrhel_laptop henrys: you there?18:19.08 
  so I seem to have the threshold halftone stuff working with a device parameter as control. 18:19.30 
  with the value set to true, it will prerender if it results in a smaller clist18:19.47 
  I need to cluster push and make sure all is OK18:20.09 
  but this change is minor compared to the change to get all the cases working18:20.25 
  Robin_Watts: If I do another cluster push now, why my one is already running, will the current one complete18:20.54 
  s/why/while/18:21.00 
Robin_Watts Unsure, but if the current one fails for any reason, it will restart, and will use the new code instead.18:21.40 
mvrhel_laptop ok. best to wait then18:21.51 
  thanks18:21.53 
Robin_Watts I'm not sure you can actually queue 2 jobs.18:21.54 
  but if you try, you'll either have your job not show up on the dashboard at all, or you'll just run the risk of the restarting problem I just mentioned.18:22.33 
mvrhel_laptop it seemed that marcosw had 2 mujstest jobs queued18:22.35 
Robin_Watts marcosw is a special case :)18:22.48 
mvrhel_laptop I will just wait 18:22.49 
  yes he is "special"18:22.55 
Robin_Watts Be nice to him. He has a car full of toys for me.18:23.14 
mvrhel_laptop oh that reminds me18:23.27 
  I have a favor to ask you18:23.31 
Robin_Watts You've got this friend that needs a "package" carried on the flight?18:24.06 
mvrhel_laptop my wife is a fan of English Breakfast Tea. And the EBT here is rather lame/weak compared to what I brought her back after our London meeting18:24.06 
Robin_Watts Ah, ok. I'll try and get some tomorrow.18:24.29 
mvrhel_laptop I was hoping you could get me a box from the Grocery.18:24.30 
  I will hook you up with a milk shake or something in SF18:24.43 
Robin_Watts :)18:24.49 
  Any particular brand she liked?18:25.10 
mvrhel_laptop Here we usually have Twinnings, but it wasn't as good as what ever I bought her there. I don't remember the brand. I will trust you in the tea dept.18:26.01 
Robin_Watts Ah, cos Twinings is the big brand here.18:26.17 
mvrhel_laptop I wonder if it is the same18:26.26 
Robin_Watts for "quality" teas at least.18:26.29 
mvrhel_laptop ok. well what ever you think will best is fine18:27.19 
Robin_Watts It was definitely English Breakfast?18:27.54 
mvrhel_laptop yes def.18:28.12 
  she has it every morning18:28.20 
Robin_Watts OK. I'll look in the shop for what EB they have. If they only have Twinings, should I still bring it? (on the grounds that it may be different US to UK)18:28.44 
mvrhel_laptop Yes. That would be fine.18:28.56 
  Robin_Watts: Thanks!18:29.01 
Robin_Watts fab.18:29.01 
  no worries.18:29.06 
mvrhel_laptop good grief the mujstest takes awhile...18:48.09 
  some how the jobs don't seem to be well distributed18:48.24 
  amongst the cluster nodes18:48.32 
  or kilometers is just slow18:49.12 
  or must have restarted18:49.16 
  bbiaw18:49.32 
Robin_Watts mujstest should take < 30 mins.18:50.49 
  tor8: Updated stuff on robin/master. As usual the last one is a work in progress.18:55.23 
  So... clipping through an image, patterns, shadings, softmasks, transparency groups... I'm running out of easy stuff to do :)18:57.32 
henrys Robin_Watts:would you be interested in taking on the multiple instance gs project?19:03.00 
Robin_Watts I'd be interested in doing the first bit at least (finding out why it won't work with multiple instances)19:05.22 
  I suspect the answer may have been globals.19:05.36 
  just about to walk dogs. bbs.19:05.55 
sebras Robin_Watts: I hope so too. but better be correct about it. :)21:40.41 
Robin_Watts hunts for what that is in response to...21:41.15 
sebras henrys: yeah. let me know when it arrives.21:41.21 
  copyright assignment.21:41.32 
henrys Robin_Watts:we can talk about multiple instances at the meeting it's on the list22:03.48 
fommil hi all – I'm have a little script that creates a PDF file containing a small piece of text http://pastie.org/4676817 But I'd like to be able to specify the background colour to be solid white (instead of transparent). Can anyone help?23:03.19 
 Forward 1 day (to 2012/09/07)>>> 
ghostscript.com
Search: