| | <<<Back 1 day (to 2014/12/11) | 20141212 |
| mvrhel_laptop | windstorm coming in the next couple hours so I may be without power | 01:57.31 |
| kens | chrisl I'm already writing an (extended) response to the Lilypond ranter | 09:32.57 |
| chrisl | kens: fair enough... just no point having it assigned to me | 09:35.08 |
| kens | No, I was going to change it anyway, I just wish Hin-Tak would shut up on it | 09:36.18 |
| | He's really not contributing anything useful | 09:36.37 |
| chrisl | And he's still using a ghostscript.com address :-( | 09:41.22 |
| kens | I know, I thought we had agreed to kill those (especially his) | 09:41.38 |
| | I'll take it up with Marcos again when he's back | 09:41.56 |
| yukam | what is preferred place for discussing jbig2dec development? I have some major (more than 2X) speedup patches; and assorted error fixes, questions, etc | 12:28.53 |
| tor8 | yukam: this is the place | 12:29.04 |
| Robin_Watts | yukam: Possibly the best way to do this is for you to open a bug on bugs.ghostscript.com and to attach your patch there. | 12:35.17 |
| | We're happy to discuss it here of course, but that's the best way to submit a patch that we can all look at. | 12:35.40 |
| | yukam: What you say on the bug is not quite true. | 12:54.57 |
| | When MEMENTO is not defined, memento.h defines various things away. | 12:55.36 |
| | For example once you have included memento.h, it will always be safe to call Memento_label. In non MEMENTO builds, it will do nothing. | 12:56.28 |
| yukam | Is it called by jbig2dec? Does it belong to public interface of jbig2dec? | 12:57.38 |
| Robin_Watts | yukam: Memento is a tool for checking for memory leaks etc. | 13:08.42 |
| kens | yukam I don't think I understand your question. Memento is a memory checking/squeezing utilty. Its not exactly called by jbig2dec | 13:08.46 |
| | collision :-D | 13:09.00 |
| Robin_Watts | If it's enabled, then yes, it will affect jbig2dec. | 13:09.11 |
| kens | Affect, yes, but its not (exactly) called by jbig2dec, and its certainly not part of any public interface. | 13:09.42 |
| Robin_Watts | And if it's enabled for one part of your code, you need to enable it for all of it - we use it in gs and mupdf, hence it has to be part of jbig2dec. | 13:09.47 |
| | The only problem, (AIUI), is that memento.h is not installed as part of the jbig2dec headers. | 13:10.26 |
| kens | That seems to be thrust of the people commenting on the (closed)_ bug | 13:10.54 |
| | Hmm, I can't remember enough of JBIG2, I may need ot go do some rereading. | 13:11.58 |
| | yukam, if you have a file which causes an error, whcih that patch fixes, it would be really useful to attach it to the bug report | 13:12.52 |
| sandy | Is there a method available to get deep copy of gs device. I want to pass a cloned copy to gdev_prn_output_page_aux for some of my processing. | 13:14.30 |
| kens | What do you mean by 'deep copy of gs device' ? | 13:14.49 |
| | Why would you want to copy the device ? Surely you would simply declare your own ? | 13:15.10 |
| | Frightened another one off :-D | 13:18.59 |
| sandy | a clone of gx_device | 13:35.07 |
| kens | Why would you want such a thing ? And if you do then allocate a structure and copy the contents. | 13:35.30 |
| | Note that, a device is a complex object, and while the base class may be gx_device, the real device is probably (almost certainly) a superclass. Copying the basic device won't be sufficient. | 13:36.14 |
| | Perhaps it would be easier if you explain the problem you are trying to solve, and we can tell you how to do that instead. | 13:36.51 |
| sandy | Yes i was thinking the same but at runtime I will no be knowing the size of various struct pointer so i was looking if there is any method which does that | 13:37.23 |
| kens | Simply put; no. | 13:37.47 |
| | But with the caveats that it might be yes if I understood your aims | 13:37.59 |
| sandy | OK, i want to do interpreter and rasterization in paralle by multithreading. | 13:38.08 |
| | in lock-step fashion | 13:38.23 |
| kens | You can't do multi-threaded interpretation. | 13:38.25 |
| | If you want to interpret and render in a separate thread, then Gs can alredy to that, the background pritn stuff | 13:38.56 |
| | <sigh> Typing gone again..... | 13:39.08 |
| sandy | I mean once the intrepreter is done for pdf page, i will launch seprate thread for zoutputpage in detached mode (will have some way to synchronize it) so main thread will start the interp on next page | 13:40.56 |
| kens | Here's the commit that introduced background printing: | 13:41.05 |
| | http://git.ghostscript.com/?p=ghostpdl.git;a=commit;h=c947a5c62b9eef2b2b2afcab986a5b7bbb8fc3ad | 13:41.07 |
| | You can find the BGPrint parameter documented here : | 13:43.59 |
| | http://www.ghostscript.com/doc/9.15/Language.htm | 13:43.59 |
| sandy | Thank you. I will look in the link for more information. | 13:45.32 |
| kens | You need to be aware that for non clist devices, all the rendering is done prior to the page being output, the output_page device method simply transfers the rendered bitmap to the final device (disk for file formats). So you cna only separate interpretation and rendering by using a clist device. Trying to override a clist device is a distinclty non-trivial task. I speak as someone who has spent much of the last 2 months doing a somewhat equivalent task. | 13:45.59 |
| | If you decide to try BGPrint you should use hte *latest* sourc e code, not the 9.15 release there are at least 3 crashes which have been fixed since 9.15 was released. | 13:47.48 |
| avih | tor8: btw, your mujs makefile change which includes ar -> $(AR) is pretty useful, since the official windows build are cross compiled on linux, and the $(AR) becomes useful. thx | 13:48.20 |
| | the official mpv* windows builds | 13:48.37 |
| | (i already cross compiled mpv with mujs successfully. so far i've been compiling mpv an mujs on cygwin and osx natively) | 13:49.11 |
| | and* | 13:49.18 |
| sandy | @kens Thank you for mentioning revision as I was using 9.15. | 13:49.49 |
| kens | NP | 13:50.19 |
| | 9.15 does work, but there are some categories of problems which have been fixed after the release. | 13:50.35 |
| | The 3 commits at the top of this list: | 13:51.09 |
| | http://git.ghostscript.com/?p=ghostpdl.git&a=search&h=HEAD&st=commit&s=BGPrint | 13:51.09 |
| sandy | ohh I see. | 13:55.15 |
| zeniko | tor8: there are 4 minor bug fixes for recent commits on zeniko/mupdf for review | 15:01.44 |
| | Robin_Watts: related to one of the bug fixes: why does pdf_xref_find_subsection solidify subsection 0 instead of the subsection with a partial overlap in your "case 3" (in commit e767bd783d91ae88cd79da19e79afb2c36bcf32a)? | 15:03.05 |
| kens | Anyone know when Henry gets back ? | 15:19.45 |
| Robin_Watts | zeniko: I can't remember offhand. | 15:30.53 |
| | subsection 0= "the newest subsection" | 15:31.05 |
| | zeniko: possibly it's a mistake. | 15:31.24 |
| kens | Crumbs, I hope Henry wasn't coming back to London today | 15:36.19 |
| | All London air space is closed until 19:00 | 15:36.38 |
| chrisl | Holy crap that's pretty major | 15:36.58 |
| kens | "Computer says no" apparently | 15:37.16 |
| chrisl | Clearly "back up systems" is phrase that hasn't entered their vocabulary...... | 15:38.14 |
| kens | They do have a backup, its called pencil and paper :-) | 15:38.33 |
| avih | surely it's called afternoon tea... | 16:38.36 |
| McErroneous | Hi, i intend to learn about printing in linux, and i am having problems to use ghostscript, "gs -sDevice=hpdjplus -sOutputFile Test.prn sample.ps" gives an "Error: /undefinedfilename" (saying a filename in Test.prn could not be found ??) | 17:30.48 |
| kens | You need an '=' for OutputFIle | 17:31.05 |
| McErroneous | ohhh , +kens thank you... | 17:31.31 |
| kens | NP | 17:31.36 |
| rayjj | for the logs (what "sandy" and kens were discussing). The function setup_device_and_mem_for_thread clones a device (and other stuff) for multi-threaded rendering and BGPrint. The synopsis of the cloning method is: get the device prototype, use gs_copydevice to get a writable version, get the params from the source device and 'put_params' those to the clone | 18:19.52 |
| McErroneous | Hi, i do not manage to convert a <ps-file> to a printer-specific format, i type "gs -sDevice=hpdj500 -sOutputFile=Test.prn sample.ps", but the file "Test.prn" is not created and , instead i end up in gs-shell... , how do i manage to tell gs to store the file in present directory ? | 18:20.32 |
| Robin_Watts | McErroneous: Try using: gs -sDevice=hpdj500 -o Test.prn sample.ps | 18:21.12 |
| rayjj | McErroneous: command line parameters for gs are case sensitive -sDEVICE=___ | 18:21.28 |
| | McErroneous: and using -o Test.prn (as Robin_Watts suggested) automatically does -dBATCH -dNOPAUSE | 18:22.10 |
| | in particular the -dBATCH prevents going to the GS> command line prompt | 18:22.36 |
| McErroneous | +Robin_Watts : using the -o option did not work either... | 18:22.40 |
| Robin_Watts | D'Oh. What ray said. | 18:22.51 |
| rayjj | McErroneous: no error messages ? | 18:22.57 |
| Robin_Watts | McErroneous: Try using: gs -sDEVICE=hpdj500 -o Test.prn sample.ps | 18:22.58 |
| rayjj | McErroneous: with -sDevice=hpdj500 it will go to the default device (on linux usually x11, on Windows the 'display' device) which will process the page to the screen | 18:24.47 |
| Robin_Watts | rayjj: Did you get VisualGDB working in the end? | 18:25.30 |
| McErroneous | no error messages.... | 18:25.33 |
| rayjj | McErroneous: because all the -sDevice=___ does is to define a string in systemdict with the name Device with the contents to whatever (hpdl500) | 18:26.37 |
| | Robin_Watts: I didn't mess with it any more -- I am working on the cust 532 issues | 18:27.01 |
| Robin_Watts | rayjj: Fair enough. It would seem sensible to at least try it within the next couple of weeks though, otherwise the trial will expire and you'll have no idea whether you liked it or not. | 18:27.39 |
| rayjj | Robin_Watts: good point :-) | 18:29.01 |
| Robin_Watts | I'm sure I've used it with peeved or peeves before. Let me try... | 18:29.36 |
| rayjj | McErroneous: also, your "sample.ps" may not contain a 'showpage'. | 18:29.37 |
| | Robin_Watts: peeves has an old gcc/gdb that it griped about | 18:30.03 |
| McErroneous | well , i get a hp 3745 installed, and "gs -h " did not show a hp3745-printer.., so i just guessed hpdj500, but all i intent is to produce a converted file from ps-format (to any printer-specific format : not important to which Printer-Languange) | 18:30.10 |
| rayjj | McErroneous: try: gs -sDEVICE=hpdj500 -o Test.prn sample.ps -c showpage | 18:30.59 |
| McErroneous | no file is beeing created... | 18:32.34 |
| | gs says : done... | 18:33.02 |
| rayjj | McErroneous: even with that command line ? | 18:33.04 |
| | McErroneous: that isn't Ghostscript then. We never say "done" | 18:33.26 |
| Robin_Watts | gs --help | 18:33.51 |
| rayjj | McErroneous: what is the first line of output following the gs ? | 18:34.02 |
| McErroneous | GPL Ghostscript 8.71 (2010-02-10) | 18:34.42 |
| rayjj | McErroneous: OK. So something is funky with your sample.ps | 18:35.18 |
| Robin_Watts | McErroneous: Try: locate tiger.eps | 18:35.47 |
| McErroneous | i tried a different ps-file..., same problem | 18:35.48 |
| Robin_Watts | does that show anything? | 18:35.52 |
| | rayjj: OK, I can visualgdb to peeved no problem. | 18:37.05 |
| McErroneous | no , i does not show anything... | 18:37.13 |
| | i guess there is no tiger.eps on my system | 18:37.42 |
| rayjj | McErroneous: try: gs --sDEVICE=ppmraw -o test.ppm -c " /Helvetica 24 selectfont 100 500 moveto (It Works) show showpage" | 18:37.55 |
| | this creates "test.ppm" | 18:38.33 |
| | with the It Works string as the only thing on the page | 18:38.49 |
| McErroneous | GPL Ghostscript 8.71 (2010-02-10) | 18:39.40 |
| | Copyright (C) 2010 Artifex Software, Inc. All rights reserved. | 18:39.41 |
| | This software comes with NO WARRANTY: see the file PUBLIC for details. | 18:39.41 |
| | Error: /undefinedfilename in (-o) | 18:39.41 |
| | Operand stack: | 18:39.41 |
| | Execution stack: %interp_exit .runexec2 --nostringval-- --nostringval-- --nostringval-- 2 %stopped_push --nostringval-- --nostringval-- --nostringval-- false 1 %stopped_push | 18:39.43 |
| Robin_Watts | -sDEVICE, not --sDEVICE | 18:39.43 |
| McErroneous | Dictionary stack: --dict:1150/1684(ro)(G)-- --dict:0/20(G)-- --dict:71/200(L)-- | 18:39.47 |
| | Current allocation mode is local | 18:39.50 |
| | Last OS error: 2 | 18:39.52 |
| | GPL Ghostscript 8.71: Unrecoverable error, exit code 1 | 18:39.55 |
| rayjj | oops | 18:40.04 |
| Robin_Watts | McErroneous: Paste back the actual command you sent. | 18:40.06 |
| McErroneous | gs --sDEVICE=ppmraw -o test.ppm -c " /Helvetica 24 selectfont 100 500 moveto show showpage" | 18:40.29 |
| rayjj | McErroneous: try: gs -sDEVICE=ppmraw -o test.ppm -c " /Helvetica 24 selectfont 100 500 moveto (It Works) show showpage" | 18:40.30 |
| | I had --sDEVICE due to a cut and paste error | 18:41.00 |
| McErroneous | okay... | 18:41.50 |
| | now.. a file has been created... | 18:42.02 |
| rayjj | McErroneous: OK. Now try the -sDEVICE=hpdj500 | 18:42.23 |
| | in that same command line. | 18:42.33 |
| McErroneous | that works.. | 18:43.24 |
| rayjj | McErroneous: OK, so as I mentioned, there is something funky with your "sample.ps" file | 18:43.50 |
| | McErroneous: what does "head -2 sample.ps" give you ? | 18:44.32 |
| McErroneous | %!PS-Adobe-2.0 | 18:45.14 |
| | %%Creator: dvips(k) 5.85 Copyright 1999 Radical Eye Software | 18:45.14 |
| rayjj | McErroneous: well, I can't really tell more without seeing the file. If you put it on pastebin, I'll have a look | 18:46.19 |
| McErroneous | could i upload the file straight to you ? here in irc ? | 18:48.05 |
| Robin_Watts | No. | 18:50.36 |
| | You can put it on pastebin or dropbox and give us a link to it. | 18:50.54 |
| | or some other ftp/web site. irc file transfer never works :( | 18:51.22 |
| McErroneous | thank you.., for your time , and your afford... :-) | 18:55.39 |
| rayjj | Robin_Watts: I found the typo that was causing VisualGDB to not start, but I get "The following GDB command is taking to long to complete: -data-evaluate-expresion "\$pc" | 18:59.34 |
| Robin_Watts | rayjj: Hmm. | 18:59.51 |
| | In 'Debug->Quick Debug Setup'... | 19:00.12 |
| rayjj | That's on peeves where it complains about an old gdb that has "known issues" | 19:00.14 |
| | Robin_Watts: let me try peeved... | 19:00.27 |
| Robin_Watts | I have: Program to debug: "/home/robin/sauce/ghostpdl.git/gs/debugbin/gs" | 19:00.48 |
| | Working directory: "/home/robin/sauce/ghostpdl.git/gs" | 19:01.02 |
| | Arguments and additional environment are both empty. | 19:01.18 |
| | I don't tick "open a core dump". | 19:01.25 |
| | Initial breakpoint in main. | 19:01.38 |
| | And it works beautifully for me on peeves. | 19:02.00 |
| | peeved! sorry. | 19:03.44 |
| | let me try peeves for comparison. | 19:03.49 |
| rayjj | I tried your executable with those same settings and I get the "taking too long" on "info shared" | 19:04.03 |
| | Robin_Watts: I'll try peeved ... | 19:04.16 |
| Robin_Watts | yes, I get the taking too long on peeves too. peeved is fine. | 19:05.05 |
| rayjj | Robin_Watts: peeved works for me as well. | 19:05.29 |
| Robin_Watts | Ok, so we'll have to remember not to use peeves :) | 19:05.56 |
| rayjj | Robin_Watts: unfortunately, the problem I need to work on only fails on peeves :-( | 19:06.30 |
| Robin_Watts | rayjj: Oh. Crap. | 19:06.50 |
| | Can you copy the gdb binary from peeved to peeves as gdb7 and then use gdb7 in the Quick Debug window? | 19:07.19 |
| rayjj | Robin_Watts: that's OK. I'm used to gdb anyway, and the problem appears to be a race condition | 19:07.23 |
| | Robin_Watts: Ill try that later | 19:08.46 |
| | Robin_Watts: that doesn't work: gdb: error while loading shared libraries: libtinfo.so.5: cannot open shared object file: No such file or directory | 19:13.49 |
| Robin_Watts | rayjj: Ok, so you'd need to copy some dependencies too, but this might be a pandoras box. | 19:14.31 |
| rayjj | Robin_Watts: yeah, I'll just stick to peeved with VisualGDB, and use gdb on peeves | 19:15.10 |
| mvrhel_laptop | arg. I have to get these image issues fixed with the xpswrite device | 20:13.02 |
| | trying to beat on printing for gsview and it is clear that there are a few problems | 20:13.32 |
| | aha. I see that marcos had opened a bug for me about the missing images | 21:13.38 |
| | testing fix for this now | 21:14.05 |
| | cool it worked | 21:31.09 |
| | however Bug 695698 is a different case. | 21:39.52 |
| | will beat on this in a bit. bbiaw | 21:40.04 |
| | Forward 1 day (to 2014/12/13)>>> | |
IRC Logs