F10 Cups / Samba / Vista [SOLVED?]

Joel Gomberg obligor11-fedora at yahoo.com
Fri Jan 30 18:33:21 UTC 2009


Joel Gomberg wrote:
> Craig White wrote:
>> On Wed, 2009-01-28 at 22:28 -0800, Joel Gomberg wrote:
>>> Craig White wrote:
>>>> On Thu, 2009-01-29 at 12:08 +1030, Tim wrote:
>>>>> On Wed, 2009-01-28 at 10:56 -0500, Brian C. Huffman wrote:
>>>>>> the alternative that someone suggested about just making it a network
>>>>>> printer and pointing to the http address for cups (bypassing samba)
>>>>>> does work.  I'm just thinking that I might want to use this 
>>>>>> machine as
>>>>>> a fileserver at some point and there's no need to use two points 
>>>>>> of entry if one works.
>>>>> It's worth thinking about that putting Samba in between CUPS and 
>>>>> remote
>>>>> access to it, and you are putting an *extra* thing in the middle, *is*
>>>>> complicating how printing works.
>>>> ----
>>>> sure, when you're talking about one system.
>>>>
>>>> also, experience tells me that things don't occur in isolation and
>>>> eventually you figure out that whatever is broken also affects other
>>>> things down the line that you haven't discovered yet.
>>>>
>>>> The theory here is fairly simple...If cups is functioning properly,
>>>> computer can print. If samba is functioning properly, Windows computers
>>>> can print to samba shared printer through cups. Since PCL drivers on
>>>> Windows systems do not speak postscript, you must permit 'raw' printing
>>>> to allow the PCL through to the printer unmolested by cups. It really
>>>> isn't that complicated.
>>> I've been experiencing a similar problem with my wife's new Vista 
>>> computer.  The printer is recognized.  I can print a test page.  
>>> Printing anything else has failed 99% of the time.  The only relevant 
>>> error I see in my cups error_log is:
>>>
>>> cupsdAuthorize: No authentication data provided.
>> ----
>> sounds as if you added...
>>         guest ok = yes
>> to the 'printers' section of smb.conf, this would go away though it
>> defies logic that you can print a test page from Windows but not from
>> other applications on Windows because of authentication.
>>
>> Also, samba authentication is for Windows to samba and once samba has
>> the job, it should automatically authenticate itself to cups so it
>> doesn't make much sense for cups to complain about authentication.
>> ----
>>> The same setup worked on W2k and continues to work on a VMware W2k 
>>> virtual machine.  The relevant user has the same user name and 
>>> password on both machines.  Only printing is affected.  Both 
>>> computers can see the shares of the other.  The shares can be 
>>> mounted, files transferred, etc.  This must be something 
>>> Vista-specific.  And, no, I can't explain why one or two jobs did 
>>> print.  I haven't posted about this issue before, because it doesn't 
>>> really seem to be a Fedora problem.
>> ----
>> There may be something Vista specific - I can't personally test because
>> I'm not eager to install Vista on any of my systems. I would suspect
>> that it's a Windows print driver issue.
> 
> That appears to be it.  I opened the task manager and changed the 
> compatibility setting on all three running Epson driver processes from 
> Vista to XP Service Pack 2. [This is done by right-clicking on the 
> process, selecting "Properties" and then the Compatibility tab.] 
> Printing now works.  Thanks for sharing your suspicion about the 
> underlying cause of the problem.

I'm replying to my own post, because the successful print was a one-time 
anomaly.  I do, now, however, seem to have found the problem.  In Vista, 
right-click on the printer, select properties -> advanced -> print processor. 
In my case, there were two processors:  Epson Inkjet and Winprint.  I changed 
the selection from Winprint to Epson Inkjet [raw] and I have now been able to 
print several jobs in a row without any problems.  My printer is an Epson Stylus 
Photo R380.  I haven't seen this documented anywhere.  I just decided to look 
more deeply at the printer properties after checking the Vista Event Viewer, 
which suggested the driver was the problem.

-- 
Joel




More information about the fedora-list mailing list