IRC Logs

Log of #ghostscript at irc.freenode.net.

Search:
 <<<Back 1 day (to 2014/01/19)2014/01/20 
psycho_oreos Robin_Watts && sags && ray_laptop, hi thanks again for your help assisting a newb. :D If you don't mind I have blogged a mini-HOWTO on my blog and I have provided credits where they are due. It has been one hell of a roller coaster ride for me to get the ghostscript to work when I could've asked and thus got answers more quicker than itching my own head.01:09.55 
ray_laptop psycho_oreos: thanks. We know lots of 'tricks' with Ghostscript (and mupdf), but it's often the "first time" that's most difficult. In your blog, please mention this IRC channel, or the #ghostscript stackoverflow.com (http://stackoverflow.com/questions/tagged/ghostscript)01:32.03 
psycho_oreos ray_laptop, I did if you like I can paste you my blog entry. I probably need to tidy up the mini-HOWTO because I was sort of intoxicated when I wrote it up. :D01:33.08 
  Err at least I mentioned the channel and IRC network Freenode. :D01:33.23 
ray_laptop ken sharp often helps folks out on S.O., and he's in the UK timezone. I am often here, but don't spend nearly as much time on S.O. and am in the PST timezone (West Coast of CA)01:33.27 
psycho_oreos Ahh yeah no wonder, I was wondering how I think sags was around when I was. Hehe.01:34.07 
ray_laptop and Robin_Watts helps folks out with mupdf and GS01:34.12 
  psycho_oreos: what timezone are you ?01:34.31 
psycho_oreos AEST (GMT+10h).01:34.45 
  Yeah I couldn't believe my luck, I hanged around a bit in postscript channel but it was dead there. I asked a question without any response and I thought to try my luck (possibly a far fetched idea) in #ghostscript.01:35.39 
ray_laptop psycho_oreos: Thanks. We have customers in Japan and a few other places near that timezone. I'm usually on late enough to interact with their morning.01:36.46 
psycho_oreos Right now it is close to midday in my timezone. I am guessing it would be night time, almost midnight there where you are.01:36.50 
  Hehe yeah I was thinking, though I think I also did stay up (in my timezone though) because I presumed most people would be based in US.01:37.33 
  May as well show my blog entry, if need be I can add extra references/credits to stackoverflow.com (wouldn't hurt either way I guess) :D http://eidolon-sapphire.blogspot.com.au/2014/01/rendering-japanese-postscript-files.html01:38.34 
  Be warned, may seem weird haha.01:38.43 
ray_laptop psycho_oreos: it's almost 6pm here (Sunday night).01:43.25 
  psycho_oreos: I'll have a look at the blog. TIA01:43.38 
psycho_oreos o.O late for sure. Coolies! let me know if there are things that need changing but yeah it does seem weird. Almost midday Monday morning here.01:44.16 
  6pm doesn't seem too bad I guess. :D01:44.28 
  Also, thank you along with others for helping. Without any support I would have probably sort of given up entirely and would have sought other ways to obtain a properly rendered output. :))01:47.08 
ray_laptop Blog's not bad. Thanks. BTW, there is a file gs/doc/Use.htm that has a lot of newbie info. This can be read with a browser from our git repo: http://git.ghostscript.com/?p=ghostpdl.git;a=blob_plain;f=gs/doc/Use.htm;hb=HEAD01:49.04 
  and the Font section follows the "How Ghostscript finds files" http://git.ghostscript.com/?p=ghostpdl.git;a=blob_plain;f=gs/doc/Use.htm;hb=HEAD#Finding_files in http://git.ghostscript.com/?p=ghostpdl.git;a=blob_plain;f=gs/doc/Use.htm;hb=HEAD#Font_lookup01:49.55 
psycho_oreos Thanks! I should add those into my blog entry (along with credits of course).01:50.35 
  Ok, thanks for the links. Updated those extra links as further reading and mentioned stackoverflow.com #ghostscript page as well.01:58.23 
  Also curiously, Ghostscript is not related to TeX is it?01:59.03 
ray_laptop psycho_oreos: AIUI (I've never used it) Tex (latex) outputs PostScript that Ghostscript is often used to print or convert to PDF02:20.31 
  psycho_oreos: BTW, I'm here off and on, particularly outside normal work hours. But few of the Artifex folks and other contributors (sebras, zeniko, sags, tkamppeter, Gigs) also keep non-standard hours. The Artifex staff that are here (besides myself) are: chrisl, henrys, kens, marcosw, mvrhel, paulgardiner, Robin_Watts, and tor.02:26.48 
psycho_oreos ray_laptop, ahh that would make sense. I was at one point getting really desperate to get that old issue sorted out. I found some reference talking about TeX and went installing a whole bunch of TeX add-on.02:32.49 
  ray_laptop, all good. :)) gotta love timezones in hindsight. :D02:33.10 
stanv hello08:10.43 
ghostbot que tal, stanv08:10.43 
stanv Does, anybody know, is there solution for : https://access.redhat.com/site/solutions/7638308:11.22 
  ?08:11.23 
kens First thisng is 'which version of Ghostscript is being used' ?08:11.51 
  The URL contains the solution 'use a recent version'08:12.24 
  8.70 is really quite old08:12.40 
stanv ghostscript-9.1008:13.30 
kens THe URL says (I think) 8.7008:13.39 
stanv okay08:14.16 
  thanks08:14.17 
kens If you beliieve you have found a bug then please report it at http://bugs.ghostscript.com, don't forget to attach a specimen file and give the command line08:14.20 
stanv I just have same message error08:14.29 
  okay08:14.34 
kens It sounds like you simply have an invalid file but we would need to check it to srr08:14.59 
stanv Okay. I'm going to make a check.08:18.26 
sebras tor8: I think Robin mistakenly thought you were on fb when he posted this one (and I _need_ to ensure that you receive it): http://goo.gl/1ZtHwx09:44.20 
kens ROFL09:44.41 
sebras kens: I think it's a good way for him to start the week. ;)09:45.58 
kens sebras it makes my eyes hurt ot look at that poster....09:49.05 
chrisl kens: do you need to talk about that build issue now? Or is not urgent? (I need to run out for a while this morning)10:38.52 
kens chrisl its not urgent, I'm looking at a resource problem atm10:39.07 
chrisl Okay. TBH, the easiest thing is probably if you get it working for you (keep path assignments etc in msvc.mak), stick it on a branch in your repo, and I'll look at "generalising" it from there.10:40.51 
kens Hmm OK I'll see what I have to do for that, the build system baffles me usually, at the moment I'm trying to figure out how to add teh new cpp file....10:41.34 
  I cna do it unconditionally of course, so maybe I'll do that and leave it to you.10:41.52 
chrisl Yes, just do it unconditionally, don't worry about other configurations, I'll sort that out10:42.29 
kens OK then I'll get ot it if I can figure out how to fix pdfwrite for my current problem.10:42.51 
Robin_Watts Robins tip for the week - don't damage your back. It hurts a lot.10:58.57 
kens WHat did you do ?10:59.20 
Robin_Watts I stood up from the sofa to answer the phone on saturday night, and was in pain before I reached the phone. No idea other than that.10:59.46 
kens :-(11:00.22 
paulgardiner Nasty. I haven't done that for a while, but I remember how painful it is.11:02.17 
  Robin_Watts: the epage test shell seems to be producing pdf files now. I should try out your new build settings and see if I can get it working on Android.11:06.17 
Robin_Watts paulgardiner: OK, cool.11:06.47 
  Let me get you a git format-patch11:06.59 
paulgardiner You could push to the repo maybe?11:07.29 
Robin_Watts paul.patch in my home dir.11:07.38 
paulgardiner okay ta11:07.46 
Robin_Watts It may need tweaking for you.11:07.49 
  Like, the NDK location is hardwired.11:07.59 
  I'd like to fix that kind of thing first.11:08.08 
paulgardiner okay11:08.10 
  yeah. Makes sense11:08.17 
Robin_Watts Damn. There is a problem.11:12.39 
  New version of patch there now, that actually seems to start building :)11:15.19 
paulgardiner I can see the advantage in that. :-)11:16.11 
  patch overwritten?11:16.20 
  or new name?11:16.41 
Robin_Watts same name.11:16.49 
paulgardiner ta11:16.53 
  using VS2005? That probably shouldn't surprise me, but...11:19.04 
Robin_Watts paulgardiner: The alternative was MSVC 611:26.22 
  actually, I think the latest version of the FBS was updated to include support for 2008, 2010, 2012, but the version I started on (your git version) only knew about vcpp and vs2005.11:27.33 
paulgardiner My surprise was that VC/VS was used at all, which I think means I've forgotten a lot more than you concerning how the FBS works.11:28.36 
Robin_Watts Step 1) Build tools like the message compiler, error generator etc.11:28.58 
  Step 2) Build the actual code, making use of the tools built in step 1.11:29.18 
paulgardiner Ah of course11:29.56 
Robin_Watts oops. spotted another problem.11:47.45 
paulgardiner Should I wait for another patch?11:50.35 
Robin_Watts There were a couple of linux specific files that need tweaking. I forgot to copy them into the new tree.11:52.43 
  Give me 10 mins.11:52.46 
  OK, I've pushed 2 commits to the repo. That should replace the patch.11:54.43 
paulgardiner ta11:55.32 
Robin_Watts and I'm still getting a build problem here :(12:00.02 
  I copied stuff from one tree into the git tree to be able to commit it. I must have missed something :(12:00.23 
paulgardiner easily done12:00.54 
  I have arm-linux-androidabi-gcc is not recognised. Is that related?12:01.32 
Robin_Watts You've probably not set NDK_HOME12:01.47 
paulgardiner oh not I didn't. So that should be the same as NDK_BASE_PATH in tools-windows-x86.py? I did change that because my copy is on E:12:06.52 
Robin_Watts I've probably got it set in more than 1 place.12:09.26 
  My plan is that it should only be set in the 1 place, as NDK_HOME, but I'm clearly not there yet.12:09.46 
  It might be worth you giving me a couple of hours at this. Do you have a bug you can look at while I get this sorted?12:10.14 
  I don't want to waste your time.12:10.18 
  I'm having problems with an include not being found at the moment.12:10.53 
paulgardiner Robin_Watts: fine, if that's okay with you. I'll see if I can get the iOS problem fixed12:11.02 
Robin_Watts Somehow a c++ file has snuck into the build between the version of the tree we had before, and the version we have now.12:25.16 
  I've not got the C++ flags set right. I'll need to do some fiddling with the ndk to find out what the proper flags should be.12:25.45 
  ok, just tidying some stuff up now.12:58.07 
  paulgardiner: OK. I've force pushed what I believe is the fixed version to the repo.13:08.36 
  I'm just doing a complete build now to see. This will take ~20-30 mins.13:08.54 
paulgardiner Okay ta13:09.13 
Robin_Watts paulgardiner: Gah. Failed at the final link because of a pthreads error.13:31.16 
  but I've seen this before, so bear with me.13:31.23 
paulgardiner No hurry. I'm getting somewhere with iOS, so quite pleased to have the chance13:31.55 
Robin_Watts and with that fix, it wants to build everything again. bah.13:38.00 
  build complete.13:53.31 
  patches force pushed.13:53.42 
kens Hehe I have a PDF file that GS processes fine, but Acrob at crashes (yes actual seg fault) on13:54.15 
  MuPDF likes the file too13:54.34 
chrisl We're clearly wrong, we must fix GS so it crashes, too!13:55.42 
kens ANd MuPDF also :-)13:55.52 
chrisl Indeed yes - we all *have* to be "bug compatible" with Acrobat.....13:56.17 
kens Its a slight worry, because the PDF file is made by pdfwrite...13:56.42 
  So I'mgoing to try a few differnt versions of Acrobat13:56.56 
chrisl Nevertheless, a seg fault can't be our fault.....13:57.28 
kens Absolytleu not.13:57.37 
  Or even absolutely13:57.47 
paulgardiner Robin_Watts: thanks.13:58.04 
kens Reader 4 can open and display the PDF file without complaint. I reckon its an Acrobat problem14:00.46 
  Anyone have a copy of reader XI ?14:00.56 
chrisl No, but I could install it if you like14:02.06 
kens Its ok I'm getting a copy now14:02.17 
  HmmI see Reader can convert PDF into Word, Excel and RTF now.14:07.29 
  And my PDF file kills it as well :-)14:07.56 
paulgardiner tor8: Hi. There's a fix for the iOS release build on paul/master14:11.56 
tor8 paulgardiner: LGTM14:30.17 
paulgardiner Thanks tor8 14:30.40 
henrys It's a holiday in the U.S.14:46.34 
chrisl henrys: which one?14:46.52 
kens explains the quiet14:46.53 
henrys martin luther king14:47.07 
Robin_Watts chrisl: the one below canada and north of mexico.14:47.18 
chrisl LOL14:47.31 
tor8 Robin_Watts: maybe you mean above canada? the U.S. has higher mountains ;)15:25.00 
Robin_Watts Mt Baker is the biggest, right? That's only just in the states :)15:25.56 
  well, Denali, but that's in Alaska.15:26.15 
  No, Mount Whitney, apparently.15:27.14 
kens What, nothing in Texas ?15:27.55 
  chrisl any idea where the link step is performed in VS makefiles ?I need to add a library include directive as well...15:40.50 
chrisl Erm, yeh, hang on.....15:41.10 
  kens: I think you need to add the lib to the end of base/msvctail.mak15:42.36 
kens Right I'll go loook there, thanks15:42.46 
chrisl Not sure what to do for the METRO build, though.....15:43.14 
kens I don't need METRO right at the moment, this will be a windows 8 command shell version of GS15:43.37 
  Hmm, not sure I know the name of the library being included :-)15:44.30 
  Wow, it seems to have built....16:08.12 
  Now i need to go add the parameter parsing16:08.34 
Robin_Watts paulgardiner: Fixed version force pushed.16:37.50 
paulgardiner ts16:38.11 
  ta even16:38.17 
ray_laptop bah! I see an object being traced by the GC, but then it is getting freed. What am I doing worng ? :-(16:55.52 
kens Using the garbage collector :-P16:56.12 
ray_laptop kens: I have to. It's a device16:56.30 
  at least it is in stable memory, immovable (in its own chunk)16:56.58 
  chrisl: have you any insight. It's been a while since I had to dig into GC issues since you've been handling things.16:58.04 
chrisl ray_laptop: if you run with -Z@ does it show it being freed, or being gc'ed?16:59.31 
ray_laptop chrisl: -Z@ doesn't show whether or not it is being traced by the GC.17:02.06 
chrisl ray_laptop: no, but it will show *why* it was freed17:02.27 
ray_laptop but I set a data BP on the address so I see when it is freed (when -Z@ does the fill)17:02.41 
  it's freed by the GC. I can see that on the call stack17:03.07 
  maybe it's not being traced. I was looking in "do_validate_chunk", but that's not actually the tracing -- let me check...17:04.14 
mmacaulay hi all, I'm having some issues embedding jpegs on OS X with MuPDF - I think it's probably an issue with the homebrew package, but just wondering if anyone has seen this before: "Wrong JPEG library version: library is 80, caller expects 90"17:04.47 
chrisl ray_laptop: Okay, pointer enumerations can be ignored if the garbage collector can't find the chunk in the allocator that's being gc'ed - that can happen with global vs local mismatches.17:05.09 
kens mmacauley sounds like you are using shared libraries, and your system shared library version of OpenJPEG is too old.17:05.51 
mmacaulay kens, yeah I'm pretty sure that is the case, as the HB package has deps on jpeg, openjpeg, and jbig2dec17:07.56 
Robin_Watts kens: That's jpeg, not openjpeg.17:08.45 
kens You should be able to build MuPDF with the static versions of the 3rd party librasries (I htink, I'm not the expert here)17:08.50 
  Robin_Watts : one of those (waves hand)17:09.01 
ray_laptop chrisl: I can't get VS to breakpoint on my enum function defined with gs_public_st_suffix_add1_final. Do you know any tricks ?17:09.14 
chrisl ray_laptop: expand the macro(s) so the enum function is declared "normally" :-(17:10.08 
ray_laptop chrisl: that's what I was afraid you were going to say. I tried setting a breakpoint at the supposed entry point in the disassembly window, but that bombed (I got an illegal instruction)17:14.12 
chrisl ray_laptop: doesn't gs_public_st_suffix_add1_final() use some of the general purpose enumerator/relocator methods, rather than ones of your own?17:15.50 
Robin_Watts ray_laptop: Make the first line of enum function call rays_funky_breakpoint.17:15.54 
  (if that's possible)17:16.23 
kens NIght all17:19.21 
chrisl ray_laptop: It looks like gs_public_st_suffix_add1_final() will direct your enum/reloc calls to basic_enum_ptrs/basic_reloc_ptrs17:22.48 
henrys marcosw: you around today? looking for a replacement for henrysx6 and you've been buying machines lately what do you recommend?17:45.54 
Robin_Watts henrys: So you'll take mvrhel?17:50.05 
henrys Robin_Watts: yes17:50.49 
Robin_Watts ok, I will not upgrade my reservation beyond the 5 seater then.17:51.08 
henrys Robin_Watts: I don't know what the other are doing though - I timed my itinerary to correspond with mvrhel_laptop 's17:51.41 
  s/other/others17:51.49 
Robin_Watts heads to pick up helen. bbs.17:52.38 
chrisl ray_laptop: (in case you missed it earlier) It looks like gs_public_st_suffix_add1_final() will direct your enum/reloc calls to basic_enum_ptrs/basic_reloc_ptrs17:55.54 
ray_laptop chrisl: that'd be fine if it got there. I have the st_device_printer as the 'supstname' name17:58.40 
chrisl ray_laptop: if it's any help, for the X device, gs_public_st_suffix_add1_final() expands like this: http://pastebin.com/e20gwbEg18:05.03 
ckorzhik hello.19:59.47 
  I want to download source code and try to compile it under android. Which repo i need?20:01.36 
  there are so many of them http://git.ghostscript.com/20:02.07 
  hm, it seems to me, that i have found it20:05.40 
ray_laptop chrisl_away: (for the logs). Thanks for the expansion, but I still didn't get my breakpoint. Turns out that I was filling my "custom" printer device using prn_device_body which sets the stype to st_device_printer. I needed prn_device_stype_body with my sturcture descriptor.21:43.12 
  as I mentioned earlier I was fooled at first because the do_validate_chunks was seeing the device, but not the trace. As with most things, obvious in retrospect.21:44.26 
  have I mentioned that the device management in gs is a *mess* ?22:59.43 
marcosw henrys: I'm around now. you want to talk about computers?23:05.01 
henrys marcosw: just thinking of updating henrysx623:05.35 
  marcosw: also can you check if you can login to my network as you usually do, I believe I've fixed everything the comcast modem is now just a bridge.23:06.58 
  ?23:07.03 
marcosw I would buy a hex core i7; hold on and I'll find the product page on tigerdirect that I was thinking of getting.23:07.11 
  and yes. I can log into macpro now. thanks.23:07.23 
henrys marcosw: noise needs to be within reason23:07.50 
marcosw you mean idle noise or noise when running at full speed? the first won't be a problem, the second you can deal with water cooling.23:09.24 
  here's the motherboard/cpu I was thinking about to replace some of the slower cluster nodes when I instead decided to buy another PowerEdge C6100 for my garage: http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=8548559&CatId=738223:10.12 
henrys marcosw: I've never really trusted that water stuff - you know water evaporates ;-) But I guess I could give it a go.23:11.06 
marcosw I've used water cooling for 3 or 4 years and never had any problems. They current systems are all closed loop.23:11.40 
  Here's the water cooling I use on my i7 hex core: http://tinyurl.com/pfahokt23:12.33 
  You do need a big enough case for it.23:13.02 
ray_laptop of course, it isn't really water23:14.01 
marcosw And I don't know if the memory in henrsyx6 is the right type, it probably is since most motherboards adapt to pretty much anything reasonable.23:14.08 
  it's a "distilled water/propylene glycol/anti-bacterial coolant"23:15.47 
  so apparently some water.23:15.57 
henrys marcosw: yeah that cpu looks to have 3x the performance of henrysx6 http://www.cpubenchmark.net/high_end_cpus.html23:16.01 
ray_laptop it mentions DDR3 2400 8Gb Corsair is only $11423:16.02 
  compared to the CPU, MB ($829) and the water cooling ($119) you might as well get new memory23:17.06 
  the older memory in the x6 is probably DDR3, but not 240023:18.00 
marcosw henrys: one warning about water cooling setups, some of them require a software driver to operate in an intelligent manner and of course they only support Windows (the url I sent saying this was the one I used my have been wrong, I have a Corsair Hydro series with two fans but it might be a different version).23:18.36 
  ray_laptop: I have great trouble getting memory to run at the advertised speed. maybe it's because I'm overclocking the cpu but I never see the numbers that are on the package. i.e. my i7 has 2400 ram but it only runs at 1600. otoh, memory speed seems to have little or no effect on cluster node performance. kilometers runs its memory at 1600 and miles at 1066 but the machines are identical in actual performance (the machines ar23:23.03 
henrys it looks like 133323:24.25 
  marcosw: how are you detecting the speed?23:25.42 
  ray_laptop:maybe I'll live with the fan, I don't mind henrysx6 and I have the case open on my desk.23:26.45 
marcosw it looks like the link I sent you is the exact water cooling setup I have, the H100i. Under Linux it defaults to operating at the quietest profile, so that's good.23:29.36 
  you can check the memory speed from the linux command line, hold on will I figure out the command...23:29.58 
  sudo dmidecode --type 1723:30.58 
henrys marcosw: I use dmidecode but I think that tells you what it is supposed to run at.23:31.05 
  marcosw: typing at the same time.23:31.20 
marcosw you can boot and run the grub based memory tester. i think that gives actual speeds. I'm going to try it on one of my other machines.23:33.32 
henrys marcosw: yeah I think that works.23:34.11 
mvrhel_laptop henrys: looks good (your schedule)23:34.17 
  just got back from skiing. snow is still really low here23:34.29 
henrys mvrhel_laptop: I think we're okay here in the mountains but it's warm 60's23:35.05 
  in town23:35.11 
mvrhel_laptop wow23:35.14 
  we have been having inversions here. Friday it as in the 60s on mount rainer and low 40s here23:35.40 
  s/as/was/23:35.47 
henrys mvrhel_laptop: lots of football excitement there?23:36.25 
mvrhel_laptop yes of course23:36.43 
marcosw the memory test says it the memory speed is 6806 MB/s. Than it goes on to say DDR3-640. Which is less than dmidecode reports, it says 1066 on that machine.23:37.00 
henrys mvrhel_laptop: I watched both games yesterday, I only really watch the playoffs, seattle looks really good.23:37.03 
mvrhel_laptop we will see how they match up to denver. should be a good game23:37.21 
henrys mvrhel_laptop: at least they seem to have a more violent group of felons than Denver, so I'm going with them.23:38.13 
mvrhel_laptop henrys: yes. I agree. these are guys I would not want to meet up with in a dark alley23:39.33 
henrys mvrhel_laptop: I have an EZ rent a car so that might be the best place to meet.23:40.16 
mvrhel_laptop ok sounds good. I will meet you there henrys 23:40.44 
  you have my cell phone number too. it is in the contact list23:40.59 
henrys mvrhel_laptop: right23:43.08 
 Forward 1 day (to 2014/01/21)>>> 
ghostscript.com
Search: