| | <<<Back 1 day (to 2020/06/03) | Fwd 1 day (to 2020/06/05) >>> | 20200604 |
| zdohnal | Hi all, I'm debugging filter chain during printing for specific printer model, and gs is run during filtering like this: | 10:13.15 |
| | gs -sstdout=%stderr -dBATCH -dPARANOIDSAFER -dQUIET -dNOPAUSE -sDEVICE=ijs -sIjsServer=hpijs<</Duplex false/Tumble false>>setpagedevice -dDEVICEWIDTHPOINTS=612 -dDEVICEHEIGHTPOINTS=792 -sDeviceManufacturer=\"HEWLETT-PACKARD\" -sDeviceModel=\"deskjet 5600\" -r300 -sIjsParams=Quality:Quality=0,Quality:ColorMode=2,Quality:MediaType=0,Quality:PenSet=2,PS:MediaPosition=7 -dIjsUseOutputFD -sOutputFile=out.prn out2.ps | 10:13.19 |
| | there is an error in cmdline after hpijs, but I don't know what is wrong with it | 10:13.59 |
| | would anyone mind looking into the command line and tell me what is the problem there? | 10:14.32 |
| | when I ran the command in shell it shows it is uncomplete command, but I don't know what to add | 10:15.34 |
| kens | There is what appears to be an incomplete setpagedevice command following the -sDEVICE=hpijs | 10:17.35 |
| | Actually ti is complete, but it doesn't have -c to introduce PostScript so GS will try to treat it as an input filename | 10:18.15 |
| | The section <<...>> setpagedevice | 10:18.49 |
| | That's standard PostScript, which you can send to GS but not like that | 10:19.04 |
| zdohnal | kens: thanks! do you know where '-c' should be put? | 11:10.08 |
| Robin_Watts_ | zdohnal: after hpijs and before << | 11:43.42 |
| | surrounded by spaces. | 11:44.01 |
| zdohnal | Robin_Watts_: unfortunately it doesn't work for me - the part of cmd looks now like this - -sIjsServer=hpijs -c <</Duplex false/Tumble false>>setpagedevice -dDEVICEWIDTHPOINTS=612 | 11:45.35 |
| | oh quotes | 11:46.18 |
| | Robin_Watts_: so the correct form should be: -sIjsServer=hpijs -c '<</Duplex false/Tumble false>>setpagedevice' -dDEVICEWIDTHPOINTS=612 ? | 11:47.41 |
| Robin_Watts_ | Whatever style of quoting works for your shell. | 11:52.26 |
| | I'd have guessed at " rather than ' but anything is possible. | 11:52.40 |
| chrisl | zdohnal: The -c stuff is documented: https://ghostscript.com/doc/9.52/Use.htm#General_switches | 11:57.15 |
| zdohnal | chrisl: IIUC it should work without quotes, but it still doesn't (command line is not complete, expects a ending character). I tried to move -c + ps code to other locations, but without any effect | 12:28.19 |
| kens | If I were you I'd start by simplifying the command line, remove bits until the problem goes away and then you know what the problem is. | 12:29.21 |
| | I'd start by removing the sequence <<....>> setpagedevice | 12:29.32 |
| zdohnal | chrisl: those postscript commands are added by foomatic-rip, if I'll change the postscript code to '-dDuplex=false' which is in PPD, it works | 12:29.55 |
| kens | Yeah becasue (almost certainly) that doesn't do anything | 12:30.30 |
| | the -d and -s switches end up ebing passed into PostScript, and potentially to the device code. | 12:30.51 |
| | If you define a value in PostScript, and nothing is looking for it, then it won't do anything | 12:31.04 |
| | It won't cause any problems but it won't have any effect either | 12:31.15 |
| | Essentially that's the same as what I suggested, just removing it | 12:31.29 |
| | The need for quotes depends on the shell you are using (or scripting language) it has to be right for the scripting language or shell | 12:32.39 |
| | double quotes usually work but I believe some people find single quotes fine on Linux | 12:32.56 |
| | And I notice your command line has hte double quotes escaped | 12:33.10 |
| | So presumably that's a requirement of whatever you are running | 12:33.22 |
| | You might find -c \"....\" -f works | 12:33.43 |
| | But that's a pure guess | 12:33.51 |
| zdohnal | kens: ok, thank you! then it seems hpijs is not capable to eat postscript commands as -c... since it is old and ijs is not much used, I'll let it go | 12:40.21 |
| | anyway kens, chrisl, Robin_Watts__: thanks for the help! | 12:40.37 |
| kens | Its not hpijs that interprets the PostScript, that is Ghostscript | 12:40.49 |
| | The Tumble and Duplex commands are used to turn duplex on and off, and to corrrectly orient the back side when duplexing is enabled | 12:41.27 |
| chrisl | zdohnal: Give me a half hour or so and I'll try it properly here | 12:41.35 |
| kens | If you don't get those right, then you will get pages where the back is upside down wrt to the front | 12:41.52 |
| zdohnal | kens, chrisl: that's the only way I can any 'output' of gs, when I use -c: https://paste.centos.org/view/f0ccfc9b | 12:44.24 |
| | it reports an error in hpijs too, so I thought it was due error in hpijs that gs failed | 12:44.49 |
| kens | In that case it appears to be the hpijs service which is giving you an error yes | 12:45.11 |
| | Its saying it can't set 8.26x11.69 inch media | 12:45.33 |
| | Very possibly yuour PostScript file is doing that (out.prn) | 12:45.51 |
| | You may need to set -dFIXEDMEDIA after setting the medi size in points | 12:46.34 |
| | And you might want to do that *before* executing setpagedevice | 12:47.25 |
| | Because I suspect that your default media (in Ghostscript) is A4, so the device is intially loaded with A4, then you do a setpagdevice, and it merges the Duplex and Tumble requests with the existing device setup (which includes A4) and sends that to the ijs server | 12:48.19 |
| | So after -sIjsServer=hpijs, put -dDEVICEHEIGHTPOINTS=792 -dDEVICEWIDTHPOINTS=612 -dFIXEDMEDIA -c "..... | 12:49.40 |
| | I'd have to actually test this to be sure, and I don't happen to have an IJS server running right now. Chris did say he'd test it if you want to hang around for a little while | 12:50.26 |
| chrisl | "STATE: -marker-supply-low-warning" no idea what that means..... | 12:50.45 |
| | "STATE: -marker-supply-low-warning" no idea what that means..... | 12:50.51 |
| kens | I'm guessing maybe its a low toner warning | 12:51.01 |
| zdohnal | chrisl: I would say it comes from hpijs, where it is default for toner level... | 12:53.15 |
| kens | HP help forums suggest its a low toner/ink warning | 12:53.52 |
| zdohnal | kens: yep, it is kind of funny to get it when you don't print at all :) and only run their ijs server | 12:54.37 |
| kens | I guess it can't communicate with a printer and panics | 12:55.12 |
| chrisl | It seems odd to complain about toner when the output is going to a file | 12:56.17 |
| zdohnal | kens: tried to switch the sequence as you said, but still the same error https://paste.centos.org/view/9b93baad | 12:57.19 |
| kens | Well, that's coming back from the IJS server isn't it ? So that's HP.... | 12:57.32 |
| | zdohnal: I'd have to set up an IJS server to experiment, but basically it looks to me like there's an attempt going on to set A4 and teh IJS server is saying 'you can't do that' | 12:58.28 |
| chrisl | zdohnal: I'm not sure how to point hpijs at my printer.... or whether it will work with a laserjet either | 12:58.34 |
| | Maybe try: https://pastebin.com/E1HA5ybr | 12:59.08 |
| zdohnal | if the problem is in hpijs, I would say let it go - because they mostly don't accept any patches on their launchpad and hpijs is old type - they prefer pdf/ps/hpcups over hpijs... | 12:59.21 |
| | chrisl: the last one works! | 13:00.19 |
| | chrisl: thanks! | 13:00.27 |
| | chrisl++ | 13:00.30 |
| chrisl | Cool :-) | 13:00.35 |
| zdohnal | no karma bot here :( | 13:00.38 |
| kens | Thanks are enough for us :-) | 13:01.00 |
| zdohnal | either way, thanks kens++ and Robin_Watts__++ too :) | 13:01.09 |
| kens | You're welcome | 13:01.16 |
| chrisl | zdohnal: For the most part, gs command line options are order agnostic, but a very few aren't. And "-c" is awkward because it actually requires interpreting Postscript | 13:01.44 |
| | I think, as my reward, I'm going to get some lunch - since it's already past time.... :-) | 13:02.59 |
| zdohnal | chrisl: so the problem is the ppd from hplip :( | 13:03.02 |
| | chrisl: so enjoy the meal :) I'll get my snack | 13:03.20 |
| chrisl | zdohnal: I don't know what hpijs does wrt to ppd files, but if it's just the ppd missing some supported page sizes, it shouldn't be too hard to resolve | 13:50.56 |
| ray_laptop | I usually tell people to put the -c "..." right before the input file since it causes us to run 'init2' and options that are handled during initialization may not have an effect if they follow the -c | 14:09.55 |
| | and it's entirely too hard to track which options those are | 14:10.25 |
| chrisl | Yeh, normally I would do the same, but I wasn't sure if the input was coming from file name or from a pipe (stdin). | 14:17.26 |
| | <<<Back 1 day (to 2020/06/03) | Forward 1 day (to 2020/06/05)>>> | |
Log of #ghostscript at irc.freenode.net.