IRC Logs

Log of #ghostscript at irc.freenode.net.

Search:
 <<<Back 1 day (to 2015/01/08)20150109 
jluc hello10:14.11 
ghostbot Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line.10:14.11 
jluc i wish to create PDF without images out of PDFs with images10:15.05 
  how can i do so with ghostcript ?10:15.39 
chrisl You mean just drop sampled images from the input?10:18.55 
jluc yes it should be as if there were no images, or as if images were replaced with white or blank areas10:20.14 
chrisl There's no easy way to do that with Ghostscript. It could be done, but it would take a fair amount of Postscript programming to make it happen10:21.01 
jluc ok thanks10:21.38 
  do you know of any tool to do so ?10:22.00 
  keeping PDF layout (apart for images)10:22.21 
chrisl Acrobat Pro lets you delete objects10:22.24 
jluc i'm looking for a command line solution10:23.34 
  i first imagined this as a very basic functionnality10:23.52 
  then i realised it seems its not10:24.05 
chrisl Why? Most people are intent on maintaining the appearance as accurately as possible10:24.24 
  I can't really fathom why you'd want to do it.....10:25.11 
jluc printing images on home quality printers is often very ugly10:25.11 
  costs lots of ink money10:25.17 
chrisl So, print in grayscale....10:25.35 
jluc and produced pages car hardly be read most of the time when there is an image background10:25.44 
  can10:25.53 
  i want to have orthograph proofread10:26.17 
chrisl Hmm, just dropping the image won't guarantee to help that situation10:26.33 
jluc images hinder the process10:26.34 
chrisl What if the text over the image is white text?10:26.48 
jluc true that'd be a problem10:27.01 
  but that is rare and a much less worse problem then having the images 10:27.26 
chrisl Besides, if the text isn't easy to discern over the image, isn't that the first "fail" in the proofing process?10:27.39 
jluc readability is adjusted to printing process10:28.09 
  final printers are professional printers10:28.14 
  proofreader's printer are home shit10:28.23 
chrisl Also, just dropping the image doesn't ensure there aren't other objects under the image10:28.38 
jluc true, its not perfect, but it'd be usefull in most of usecases10:29.21 
  ok so no hints ?10:30.36 
chrisl Ghostscript does allow you to specify ICC profiles for different object types. If you could create an ICC profile that mapped all colours to white, and apply that to images..... that *might* be viable. But that's *way* outside my knowledge base.....10:30.41 
jluc lol that's an idea10:30.59 
chrisl What platform are you running on?10:31.18 
jluc ubuntu + scribus10:31.24 
  solution as for now is to rename all images folders to images.bak and re-produce the PDF with missing images10:32.06 
  i do it by hand as for now10:32.26 
chrisl Actually, what I would be tempted to do would be to try to create an ICC profile to give a very washed out look to images, thus the image is still visible, but very light10:32.31 
jluc so i can write a script to do that10:32.31 
chrisl You could look into http://sourceforge.net/projects/pdfedit/10:32.50 
  It has scripting capabilities, so you might be able to script the removal of image objects....10:33.20 
jluc ok thank you chrisl 10:33.37 
chrisl disclaimer: I've never found pdfedit to work terribly well, so YMMV!10:33.42 
jluc ok10:33.50 
chrisl But personally, I like the "washed out" ICC profile solution - I think it would fit better into a proof reading workflow10:34.55 
tor8 chrisl: wouldn't hacking the GS pdf interpreter to just drop the showimage at the end work as well?10:38.46 
  or is all that tied to consuming the data?10:39.01 
chrisl tor8: as I said, Postscript programming required.....10:39.12 
tor8 chrisl: true...10:39.23 
  robin's pdf filter thing for mupdf should be able to do it as well, but that requires C programming and recompiling10:39.43 
chrisl I'm not sure how much, but knowing the PDF interpreter, "trivial" rarely applies.....10:40.03 
tor8 yeah. as you said, a washed out color profile is probably the best.10:40.23 
  but if all you want is the text, the text extraction device might be enough?10:41.06 
chrisl He wants to keep the layout10:41.19 
  Hacking the GS PDF interpreter *probably* wouldn't be a massive hassle, but ensuring it works with all cases could well be10:41.36 
jluc with complex layout, text extraction makes the text quite often not possible to be read10:45.04 
  to be read as a text (not as a bunch of glyphs)10:45.28 
tor8 chrisl: jluc: the mupdf text extraction ought to be possible to preserve the layout if using absolute HTML positioning or svg or something similar10:45.56 
  but again, requires programming10:46.00 
chrisl The txtwrite device maintains sentences, paragraphs etc, but not the page layout.....10:46.30 
  WTF is the point of automake?!?! What a pile of utter sh*te...... :-(10:54.35 
tor8 uh oh... I think we finally broke chrisl.10:54.57 
sebras chrisl: welcome! I have been here for a long time. ;)10:55.22 
chrisl Well, the thing is, I get the point of autoconf - although I think it's overblown, bonkers and stupidly over complex..... but automake is just... ugh :-(10:56.33 
sebras chrisl: from my pov it is just a tool that is more difficult to configure to have it produce the makefile you really want.10:57.37 
  chrisl: what exactly did you get stuck on?10:57.44 
chrisl I'm not stuck, exactly.... the ijs automake stuff creates soft links for 'compile', 'config.guess' etc which link to files specific to the automake version in use. So you end up with a configuration that relies on the same automake version being installed at build time10:59.36 
  Aha, looks like I want to add the "--copy" argument, that copies the relevant scripts rather than linking them......11:06.25 
sebras chrisl: right, I think there is always a way to have a package's autotools stuff be independent of the version installed on your system.11:07.31 
  chrisl: don't forget --add-missing11:08.02 
chrisl sebras: but really, that *should* be the default - that is, after all, supposed to the be the point of autotools11:08.07 
  --add-missing is already there, but thanks11:08.22 
sebras chrisl: agreed. I have no idea why the default is to symlink, that makes no sense.11:09.22 
chrisl sebras: presumably it's just to make life difficult for people like us!11:10.24 
sebras chrisl: that does indeed sound like GNU.11:10.38 
Robin_Watts paulgardiner, chrisl: I booked the snowmobile trip yesterday. No charges to my CC unless we don't turn up. And we can cancel anytime with 24 hours notice.14:30.58 
  They will supply snow suits/boots/helmets. They can't guarantee to supply googles. so we need to get our own.14:31.39 
  I'm seeing if the Hotels Ski place will rent them now.14:31.55 
chrisl Robin_Watts: sounds good. Are the goggles required, or would something like sports sun glasses work?14:49.07 
Robin_Watts chrisl: They seemed to imply that we really ought to have them.15:37.10 
  They look to be about 20 quid each to buy on amazon.15:38.06 
  I'd be wanting a pair I can wear glasses underneath.15:39.19 
chrisl Robin_Watts: Okay. I was just thinking that I really need some shades for cycling (stop the flies getting in my eyes), so would have been two birds, one stone.....15:40.29 
Robin_Watts chrisl: Feel free to mail them and ask.15:40.55 
  They say that we should bring googles/scarf/waterproof gloves.15:41.29 
  So I guess it's a matter of covering as much exposed skin as possible.15:41.46 
chrisl Yeh, so goggles are going to work better for that15:42.41 
rayjj You *definitely* want goggles for snowmobiling. For skiing you can get by with sun glasses, but on the vehicle it's different. Double layer glass is good to prevent fogging, and sometimes can be had for near the same cost15:50.16 
  and AFAIK you can wear glasses under all ski goggles15:50.38 
Robin_Watts Yeah, Amazon do single layer skiing googles for 3 quid, but ones you'd actually want start about 20 quid (and go up to 100 or so)15:52.49 
  Anyone have any experience of renting such things?15:53.37 
graphicore mvrhel_laptop: Can you have a look at: http://bugs.ghostscript.com/show_bug.cgi?id=69498215:55.26 
rayjj graphicore: I had a look, and I tend to concur that this doesn't work as I expect. I zapped the tint transform in tritone.pdf (to make it totally wrong) and with or without the -sDeviceNProfile=artifex_tritone.pdf it looks the same15:58.26 
  graphicore: and different to the original (mostly correct) tint transforms in tritone.pdf15:59.08 
  graphicore: mvrhel_laptop: In theory (AIUI) with an incorrect tint transform, using -sDeviceNProfile=artifex_tritone.pdf should restore the original (correct) appearance16:01.12 
graphicore +rayjj: I thought the icc profile should define the colors rather than the tint transform in the pdf/eps at least in the eps, the deviceN to cmyk transformation is meant as a fallback.16:01.34 
  rayjj: also, the test.eps has for that reason messed up color definitions (something else could still be wrong)16:02.27 
rayjj graphicore: OK, that's what I thought (why the Orange was more green than orange)16:03.07 
mvrhel_laptop oh someone actually using a DeviceNProfile16:04.12 
rayjj graphicore: on the off chance that setcmykcustomcolor was the problem (defining Separation colorspace instead of DeviceN) I modified that and still the same which is why I switched to trying tritone.pdf16:04.43 
  mvrhel_laptop: yeah, and it seems to be ignored16:05.00 
graphicore eventually I plan like to use the display device for a color managed workflow to create deviceN (duotone, tritone etc. images)16:05.02 
rayjj I have to do a school run. bbiab16:05.11 
mvrhel_laptop graphicore: glad to here you want to make use of it. I will look this all over either today or this weekend16:05.50 
graphicore mvrhel_laptop: great THANKS! :-) here is the project, btw: https://github.com/graphicore/multitoner16:06.59 
mvrhel_laptop cool16:07.58 
Robin_Watts resists saying "I believe Amazon sells a large range of was to suit all budgets." on bug 695763.16:19.37 
mvrhel_laptop bbiaw17:26.30 
 Forward 1 day (to 2015/01/10)>>> 
ghostscript.com
Search: