Log of #ghostscript at irc.freenode.net.

 <<<Back 1 day (to 2021/02/23)Fwd 1 day (to 2021/02/25) >>>20210224 
Robin_Watts test00:02.08 
artifexirc-bot <RayJohnston> test passed00:05.22 
docbrown I have some enterprise-y software that's using an old version of CUPS to send PCL to an HP laser printer. The printer handles the PCL just fine, but if I run it through the latest GhostPCL or EscapeE, I get an interpreter error about "operator 0x4c". EscapeE at least tries to render the page, but it's very garbled. Are there any PCL experts here14:43.42 
  that might be able to help? Unfortunately the vendor refuses to upgrade their print system, so I guess I just need to figure out some way to patch the PCL output to make it work.14:43.43 
artifexirc-bot <KenSharp> docbrown not right at the moment, but to be honest I don't understand hte problem, if the printer works, why do you care about what Ghostscript deos with the PCL ?14:44.49 
  <KenSharp> docbrown not right at the moment, but to be honest I don't understand hte problem, if the printer works, why do you care about what GhostPCL/ deos with the PCL ?14:45.25 
docbrown I'm capturing the PCL and need to convert it to PDF for post-processing.14:45.33 
artifexirc-bot <KenSharp> Well, you could open a bug report and attach the PCL file to the bug report, that way someone will look at it. I can't guarantee that we'll accept it as a bug though.14:46.19 
chrisl docbrown: Firstly, "PCL" covers a *lot* of ground. You'll need to be a lot more specific about *what* PCL14:47.04 
artifexirc-bot <KenSharp> You should probably be aware that PCL and PDF do not share the same imaging model, PCL RasterOPs will not come out right on a PDF unless you render the PCL to an image and then wrap the image up as a PDF.14:47.07 
  <KenSharp> Which probably defeats any 'post processing' you had in mind14:47.33 
docbrown KenSharp: Ok. Even if I have to OCR it, that will be fine.14:48.48 
  chrisl: PCL/XL is what EscapeE recognizes it as14:49.08 
  I'll try to put together a bug report. Thanks all.14:49.20 
artifexirc-bot <Robin_Watts> docbrown: Find us the smallest example of a file that shows the problem, if you can.14:49.26 
  <KenSharp> I'd be surprised if we get PCL/XL wrong. But PCL/XL can include legacy PCL of earlier versions, and that could be your problem14:49.40 
  <Robin_Watts> All too frequently people give us many megs of data, with loads of pages and say "it does wrong", and then wonder why we don't prioritise it πŸ™‚14:50.01 
chrisl Emoticons don't work too well over the IRC bridge.....14:50.42 
docbrown My files are one page at about 110KB.14:50.43 
artifexirc-bot <KenSharp> one page is OK I expect14:50.56 
  <Robin_Watts> docbrown: That's a sane size.14:51.01 
  <Robin_Watts> emoticon test πŸ™‚14:51.54 
  <Robin_Watts> emoticons work fine. You just need a decent client (and/or unicode font)14:52.22 
  <KenSharp> The log is screwy14:52.38 
  <chrisl> Γ°ΕΈβ„’14:52.49 
  <KenSharp> thornYdieresistrademark14:52.55 
  <Robin_Watts> but yes, it's annoying that when I type 2 chars : then ), discord converts it to a unicode thing.14:53.16 
  <Robin_Watts> hexchat displays it correctly.14:53.31 
  <KenSharp> docbrown I see the bug report.15:03.02 
  <KenSharp> There's an awful lot of warnings/errors in the back channel there15:05.38 
  <KenSharp> Possibly because it should really have stopped at the first error of course15:06.08 
  <chrisl> Hmm "HP-PCL XL;1;1;Comment Copyright Artifex Sofware, Inc. 2005"15:07.13 
  <KenSharp> That's in the PCL ? Not too surprising I guess, since its CUPS15:08.08 
  <chrisl> Yeh, but 2005??15:08.29 
  <KenSharp> Hey that's only 15 years....15:08.55 
  <KenSharp> docbrown did say it was a legacy system, presumably he's been back to check on it in the DeLorean15:09.23 
  <chrisl> Well, actually, the sources still have that comment 😦15:09.24 
  <KenSharp> They do ? I guess we've serially missed that when updating copyright notices15:10.01 
docbrown To be clear, it's not my legacy system. My stuff runs the latest CUPS. ;)15:10.13 
artifexirc-bot <chrisl> Ah, even worse, then!! πŸ˜‰15:10.33 
  <KenSharp> Well if we're producing it, then we ought to be able to read it. So one for Henry15:10.57 
  <chrisl> We *were* producing it.....15:11.11 
  <KenSharp> I suppose it could be the PCL outptu device whose name I have temporarily forgotten (possibly in self defence)15:11.27 
  <chrisl> Something like pxlcolor/pxlmono15:12.31 
  <KenSharp> That's the puppy15:12.38 
  <chrisl> That code has had considerable work over the last 5-10 years....15:16.09 
docbrown Even if it turns out that this was invalid output from CUPS that got fixed in a later version, that's fine. I don't expect GhostPCL to be able to read blatantly incorrect stuff. I'd just like to know what exactly it doesn't like so I can patch the PCL files. But I know basically nothing about PCL, so...15:16.32 
artifexirc-bot <KenSharp> Yeah but if the user won't upgrade then the answer may simply be 'yes it was a bug, it got fixed'.15:16.37 
  <chrisl> I might be hard to tell with just the PCL to work from15:17.15 
  <KenSharp> Can you run pxldis chrisl ?15:17.41 
  <chrisl> Probably...15:18.08 
  <KenSharp> Might tell you something, I can't seem to find my copy and I don't think it runs on Windows anyway15:18.25 
  <chrisl> It errors out pretty quick, without much meaninful info15:20.02 
  <KenSharp> Oh well15:20.15 
  <chrisl> Another consumer from a well known printer manufacturer throws a bunch of "$print - ERROR: (Illegal_Tag_Error) Unknown Tag (0x4C)" in it's disassembly15:24.14 
  <KenSharp> Seems like the PCL is pretty badly broken, potentially irrevocably so. Though it does raise the question of how the printer works15:24.49 
  <chrisl> Even XL printers appear to have quite some variations between them15:25.35 
  <chrisl> I'm hesitant to throw it at my HP printer in case it barfs a thousand pages of garbage at me15:26.46 
  <KenSharp> Yeah that's always a pain15:26.57 
  <KenSharp> And very wasteful15:27.02 
docbrown We print hundreds of these a day on at least four different models of HP printers.15:27.49 
artifexirc-bot <chrisl> That doesn't prove it's correct....15:28.24 
docbrown I know. I'm just saying it probably won't screw up your printer.15:28.41 
artifexirc-bot <KenSharp> I'm not sure we'd learn much by printing it, other than waht it's supposed to look like I guess. If we already know that at least some HP printers can deal with it15:29.52 
docbrown It works on a Lexmark, too. But I don't remember what model it was before we threw it out for other reasons.15:30.00 
artifexirc-bot <chrisl> For ppstat.pcl, my HP printer returns: PCL XL Error15:33.13 
  <chrisl> 15:33.14 
  <chrisl> Subsystem: KERNEL15:33.15 
  <chrisl> Error: IllegalTag15:33.17 
  <chrisl> 15:33.18 
  <chrisl> File Name: kernel.c15:33.19 
  <chrisl> 15:33.20 
  <chrisl> Line Number: 192115:33.22 
  <chrisl> 15:33.23 
  <chrisl> PCL XL Error15:33.26 
  <chrisl> 15:33.27 
  <chrisl> Subsystem: KERNEL15:33.28 
  <chrisl> Error: IllegalTag15:33.30 
  <chrisl> 15:33.31 
  <chrisl> File Name: kernel.c15:33.32 
  <chrisl> 15:33.34 
  <chrisl> Line Number: 192115:33.35 
  <KenSharp> Well that pretty much confirms the file is broken I guess15:33.38 
  <chrisl> And prints a page with that on it15:34.08 
docbrown chrisl: What model of printer did you use?15:34.30 
artifexirc-bot <KenSharp> Which is a lot nicer than the usual PCL error garbage15:34.31 
chrisl docbrown: HP LaserJet 400 color M451dn15:34.52 
artifexirc-bot <KenSharp> same as mine, or very very similar15:35.38 
chrisl And same error message with pprec.pcl15:35.51 
artifexirc-bot <KenSharp> Still, the question really is can the content be fixed and that's a question I certainly can't answer, one for Henry15:36.34 
docbrown Hmmm, that's very similar to one that we have that works. I wonder if CUPS is somehow sending different output when I use netcat.15:36.51 
artifexirc-bot <KenSharp> Who either isn't here or is hiding15:36.52 
  <KenSharp> To be fair it is only 8:30 am for him15:37.13 
  <chrisl> I'm using netcat for this. No cups involvement15:37.48 
docbrown I'm using netcat to capture the PCL from CUPS, though.15:39.17 
chrisl docbrown: Have you looked at: https://wiki.ubuntu.com/DebuggingPrintingProblems ?15:40.16 
  Specifically the "Capturing print job data" and "Getting the data which would go to the printer" sections15:40.46 
docbrown Unfortunately, I have zero access to the system running CUPS. I'm just taking the printer offline, assigning its IP to my desktop, and using netcat to capture what comes in on port 9100.15:42.05 
chrisl Okay, that is going to make things extremely awkward15:43.11 
docbrown Believe me, I know. lol15:43.24 
  I've been fighting with this vendor for ages, but they are very set in their ways.15:43.39 
artifexirc-bot <KenSharp> description of the entire printing industry15:44.30 
chrisl docbrown: Silly question: you're not using "-C" with netcat?15:45.09 
docbrown ncat -o pprec.pcl -l 910015:46.03 
chrisl Then I am at a loss. GhostPCL is producing a pretty good analogue of what my HP printer does with the samples we have15:49.27 
docbrown I actually haven't tried netcat-ing the file directly myself... If I do that, I also get the IllegalTag errors.15:52.47 
artifexirc-bot <KenSharp> Somehow that sounds like what you are sending to the printer is not wehat you are actually capturing15:54.07 
docbrown Right. I just don't see how that could be happening. But I'll do some more investigating.15:55.14 
artifexirc-bot <HenryStiles> is this the bug report I just responded to on Bugzilla?15:55.48 
docbrown Yes15:55.53 
chrisl docbrown: Can you post a comment on the bug saying things are on hold while you investigate further at your end?15:55.53 
docbrown chrisl: Sure thing15:56.17 
  I appreciate everyone's help. Sorry to take up your time.15:59.59 
artifexirc-bot <HenryStiles> no problem16:00.42 
docbrown Well, I've made some progress. If I capture the PCL with Wireshark and netcat that to the printer, it at least doesn't print it an error. But it does hang until I cancel the job on the front panel. Not sure if that's better or worse...16:46.09 
artifexirc-bot <KenSharp> Its still different from what happens with CUPS though ?16:46.38 
docbrown Yeah, CUPS prints it fine.16:46.55 
artifexirc-bot <KenSharp> I suppose its possible CUPS terminates the network connection which causes the printer to stop or something, no idea really16:47.35 
  <KenSharp> Ah, but the printer doesn't print anything either ? That doesn't seem like a good thing16:48.01 
docbrown I'll try to do a binary diff of the netcat-captured file and the Wireshark-captured one.16:49.09 
artifexirc-bot <HenryStiles> docbrown you can post that result on the bug if you like, and I'll compare it if you like. I have an XL disassembler.16:50.00 
  <HenryStiles> docbrown you can post that result on the bug , and I'll compare it if you like. I have an XL disassembler.16:50.13 
docbrown Success! The Wireshark output is rendered by GhostPCL and the printer successfully.17:02.53 
  The GhostPCL output is rather hard to read, but I probably need to tweak some options.17:03.16 
  Should I change the bug status to INVALID since it turned out to not even be a bug in Ghostscript?17:07.06 
artifexirc-bot <HenryStiles> I'll do it, thanks17:08.04 
chrisl docbrown: If you need visually accurate representation of the PCL, you might want to look at the pdfimage device family (not sure if that was mentioned before)17:13.10 
artifexirc-bot <HenryStiles> oh of course netcat doesn't support binary directly, it has to be escaped.17:14.50 
  <chrisl> Hmm, I use netcat to send PCL, PXL, Postscript to my printer regularly....17:16.24 
docbrown The pdfimage8 device produces perfect output. You all just made my day!17:17.32 
artifexirc-bot <chrisl> Probably as of the next release, the pdfimage devices will have an (optional) OCR function, too. So you can keep the ability to copy/paste/search for text17:18.50 
  <HenryStiles> @chrisl ah okay I misread something.17:19.30 
  <chrisl> @HenryStiles I did wonder if there might be some inter-platform interaction going on.17:20.21 
docbrown Maybe it escapes in listen mode but not client mode for some reason?17:20.46 
artifexirc-bot <HenryStiles> @chrisl but the comment in the bug indicating a problem in netcat seems unlikely. We are missing something.17:21.25 
  <HenryStiles> anyway back to the salt mine.17:21.50 
  <chrisl> @HenryStiles With not having access to the "source", it's pretty hard to tell.... and it's cups, so black magic abound!17:22.14 
  <RayJohnston> I empty the input tray (except for one sheet) for those kinds of tests (or more often add paper because someone used it up and didn't)17:38.55 
  <chrisl> @RayJohnston I do that when I remember - I rarely remember 😦17:39.54 
  <RayJohnston> well, I rarely send random files to my HP to test πŸ˜•17:40.32 
 <<<Back 1 day (to 2021/02/23)Forward 1 day (to 2021/02/25)>>> 
ghostscript.com #mupdf