Printing from WinXP -> Puppy - Strange Problem [SOLVED]

[rcrsn51]

Nice work.

Regarding the Raspberry Pi: What OS is it running?

In mainstream Linux, the old methods of CUPS printer sharing as used by Puppy don't work any more. You need additional infrastructure.

This is discussed in How to Install your Printer in Puppy. The recommendation is to share a printer via the p910nd server instead of CUPS.

Unfortunately, that won't help you. With p910nd, the client still has to supply the driver, which the SP1 box cannot do.

In any case, using CUPS to share a printer that is already shareable by virtue of being a standalone networked printer is a rather convoluted process!

[mikeb]

Glad you have a solution... 'tricking it' seems the way with printers and windows sometimes and your fix sounds all too familiar.
I once had to get a driver from xerox's japanese site as all english versions had been removed....microsoft don't like supporting older systems lol..
(or HP)..but fortunately a foreign language protects us free spirits at times

Sorry if my input was fragmented but I don't get notifications here and the weather was good for sneaking away on a boat.

mike

[p310don]

The PI is running Raspbian, I'm pretty sure from April this year.

In mainstream Linux, the old methods of CUPS printer sharing as used by Puppy don't work any more. You need additional infrastructure.

I've come up against that, even in newer Pups with later / latest CUPS. I really like the setup that I have in that it "just works" - except when it doesn't as per this thread.

I've read the p910nd thread thinking it might have been my solution, but that windows driver thing is the stickler.

In any case, using CUPS to share a printer that is already shareable by virtue of being a standalone networked printer is a rather convoluted process!

No doubt, but has the advantage of being able to use the driver that comes with Windows (HP Color Laserjet PS) so as to avoid bloat & resource overhead that some windows drivers have.
In that vein, there are two XP computers that commonly use this printer. They are identical hardware wise (P4, 2.8ghz, 512mb ram), one has SP3 and the other is SP1. The SP3 machine prints via the Windows driver. It loads a preview screen first, and then prints to the printer - a process that takes perhaps 30 seconds at best. The SP1 machine prints via CUPS (now) and takes maybe 10 seconds to do the job.

[rcrsn51]

I suspect that mikeb is correct about the SP1 box. Buried somewhere in its registry is information about the Samsung printer that ties it to the old IP, which it keeps wanting to use.

By using a different name for the Samsung printer, you were able to bypass this problem.

[rcrsn51]

I was curious about how your situation works with the newer CUPS.

I have a Stretch-live setup with CUPS 2.2.1. Its installed printer is a stand-alone HP Laserjet via the socket:// protocol over Ethernet.

I made this printer shareable with the usual two CUPS settings.

I then booted an old XP netbook. Using the printer wizard, I installed a networked printer like you did:

Code: Select all

http://192.168.2.10:631/printers/HP_Laserjet_Pro..."

I selected "HP Color Laserjet PS" for the driver and printed a test page.

But it failed, and CUPS reported "can't detect file type".

So I tried a different driver in XP - "IBM Network printer 24 PS". This one worked!

So the system still works in new CUPS, but you need a more compatible Postscript driver in the XP box.

[jamesbond]

I suspect that mikeb is correct about the SP1 box. Buried somewhere in its registry is information about the Samsung printer that ties it to the old IP, which it keeps wanting to use.

By using a different name for the Samsung printer, you were able to bypass this problem.

No that's not it. I run through the registry and remove anything that has a samsung name in it. And rebooted too to make sure. It didn't change anything. We tried connecting using hostname, IP address, etc. Nothing worked.

The machine that has SP1 also has Samsung driver installed there although it doesn't work. My thinking is, when WinXP tried to add the printer, it sent a query about the printer and found out that it was a Samsung, and thus attempted to pair it with the Samsung driver (which doesn't work). And then instead of saying bad driver or something, it produced a very helpful message of "cannot connect to printer".

The trick that p310don worked because when he first created BrotherSam share, it was a Brother printer (whose driver isn't installed in the machine), so Windows didn't attempt to pair it to an existing driver and displayed a "choose driver" dialog instead (and choose HP Color Laserjet PS). Which worked for the Brother printer.

Then from CUPS we "modify printer" and changed that share name to point to the Samsung printer (but keep the share name of "BrotherSam").

From Windows point of view, nothing changed; as far as it is concerned it thinks that it is still printing to the Brother printer using that HP Color Laserjet PS driver.

But of course now CUPS actually prints to Samsung. Which works.

[jamesbond]

But it failed, and CUPS reported "can't detect file type".

That's exactly the error message that we saw.

If you see this: https://unix.stackexchange.com/question ... st-upgrade it said that CUPS worked before upgrade, and failed after that.

I think it's a bug somewhere in the CUPS printing chain. I didn't spend enough time on this to figure out what's wrong.

I'm inclined to this conclusion because there is similar bugs here: https://bugs.launchpad.net/ubuntu/+sour ... ug/1718394.

It said the filter failed. But the filter didn't fail using CUPS "print test page". So whatever doing Postscript rasterization isn't doing a good enough job.

So I tried a different driver in XP - "IBM Network printer 24 PS". This one worked!

So the system still works in new CUPS, but you need a more compatible Postscript driver in the XP box.

Curious! That adds weigh to the case of PS rasterisation bug.

As a side note: we're currently building a new Fatdog and the new CUPS have new elements in its rasterisation; mupdf. We haven't tested anything yet (we've got it to boot to desktop but many of the components aren't there yet), so it will be interesting to see what happens.

[rcrsn51]

I notice that some of those examples refer to Gutenprint drivers.

I suspect that some drivers, like the HP Laserjet PS, have lost compatibility with how CUPS detects the format of the incoming job - PS, PDF, binary, etc.

But it's hard to know who's at fault.

[p310don]

I selected "HP Color Laserjet PS" for the driver and printed a test page.

But it failed, and CUPS reported "can't detect file type".

So I tried a different driver in XP - "IBM Network printer 24 PS". This one worked!

So the system still works in new CUPS, but you need a more compatible Postscript driver in the XP box.

For interest's sake, I'll give this a go printing to the Pi's version of the printer and see what happens (it has a later CUPS). I have a couple of days off work, so won't have results for a little bit (unless curiosity gets the better of me and I go into work to check it out on my days off)

[p310don]

Great news. Someone was sick so I had to go to work. Not normally great news, but it gave me a chance to try to IBM Network printer driver to the PI's shared printer. Worked flawlessly first try.

So now the question is for rcrsn51, is it a good idea to change some of your how to's to use that driver rather than the HP Laserjet Color PS?

[rcrsn51]

Great news. Someone was sick so I had to go to work. Not normally great news, but it gave me a chance to try to IBM Network printer driver to the PI's shared printer. Worked flawlessly first try.

Excellent.

So now the question is for rcrsn51, is it a good idea to change some of your how to's to use that driver rather than the HP Laserjet Color PS?

Actually, I removed that info from How to Install your Printer in Puppy a while back.

I don't remember why - possibly because CUPS port 631 shared printing wasn't working the same way anymore and running a p910nd server seemed like a better choice.

Or maybe because the XP method no longer works in Win7. [Edit] I tried to make it work in Win7 and gave up.

But I will add a link to this thread.