[cups] Printed output has umlaut instead of capital P

Peter Flynn peter at silmaril.ie
Thu Jun 26 14:12:20 PDT 2014


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

This is a weird one. I'm just about to print a very long document (a
thesis), and I printed a selection of test pages first.

After five pages, all capital P characters are substituted with an
umlaut (diaeresis) character (ยจ). The document is set in T1 Charter, and
the error occurs in the default roman only (not italic, bold, or
anything else). The characters in the document display correctly on the
screen: the error occurs only when printing.

I did another test of the first 20 pages and the same thing happened. I
printed it from qpdfview and evince: same thing. I can't use acroread,
as it is stuck fast in lo-res (150dpi) monochrome for some unknown reason.

The system is running Xubuntu 14.04 and CUPS 1.7rc1. The printer is a HP
K7100 attached to an Xubuntu 13.04 machine in the next room, connected
over the wired network, running CUPS 1.7.2. No other printout has
demonstrated this error, and I've been printing large and complex docs
for a long time.

If I copy the file to one of the iMacs and print it from there to the
same printer, it seems to work OK. The iMac runs CUPS 1.7.2.

This is a ~500pp PDF generated with pdflatex. There were no errors in
processing except a few trivial overfull \hboxes. The end of the log
file says:

>> Here is how much of TeX's memory you used: 24254 strings out of
>> 493307 339189 string characters out of 6139906 583714 words of
>> memory out of 5000000 26952 multiletter control sequences out of
>> 15000+600000 210125 words of font info for 303 fonts, out of
>> 8000000 for 9000 981 hyphenation exceptions out of 8191 
>> 59i,20n,124p,1106b,2552s stack positions out of
>> 5000i,500n,10000p,200000b,80000s
[snip font log]
>> Output written on thesis.pdf (493 pages, 12062434 bytes). PDF
>> statistics: 4758 PDF objects out of 5155 (max. 8388607) 2583
>> compressed objects within 26 object streams 0 named destinations
>> out of 1000 (max. 500000) 862 words of extra memory for PDF
>> output out of 10000 (max. 10000000)

Yes, 303's a lot of fonts: the document has a lot of complex formatting
examples.

My gut feeling is that something, somewhere, is running out of memory.
Could this be CUPS or one of its rasterisers? Does CUPS or even PDF have
any kind of "conserve memory at the expense of speed" switch that
pdflatex could set?

///Peter
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.14 (GNU/Linux)

iQEcBAEBAgAGBQJTrIyzAAoJEHt9ZfbX6inQ02IH/Rfh1xrxfl+zAMxOSCnox1wx
vOp/n1xPwDT460o7LeVQM8aidoFKIccSkesU2vAGFgTJHSVzSZacE18nLfkeaN90
zYJHpSLiDmf/ZUdwq7oGjlx6xyrafswKPtBTfMIvPhhLHvDP4bDlGiAvvz1AdFlg
vJTJkk+CAErBa4TN2f8vCwHtVLmSVmdhFaTZ4wRJngUdG/V/ub6qo7tkP9WEb4fX
zwQxMzB01Uh9pbat6N9w9AVHiSHMqEW/XjNaWMZDdlgjWG2EcnwjnXqA89/KJavJ
FdyXwrjwXQCXEy3QINpMuMHRwaYx2QQFnd1sldWowlpmKIohads0o+xNVg8u4ac=
=kLZC
-----END PGP SIGNATURE-----



More information about the cups mailing list