https://bugzilla.novell.com/show_bug.cgi?id=628233 https://bugzilla.novell.com/show_bug.cgi?id=628233#c24 Johannes Meixner <jsmeix@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED --- Comment #24 from Johannes Meixner <jsmeix@novell.com> 2010-10-20 10:04:41 UTC --- Unfortunately you did not provide information which printer driver you use but "Canon i865" indicates that you use the Gutenprint driver. "RIPCache 1024m" matches to this mail on the gimp-print-devel list: http://sourceforge.net/mailarchive/message.php?msg_name=79617DCE-FA3D-462C-8... -------------------------------------------------------------------------- There has been some recent discussion on this - recent versions of Ghostscript changed the internal banding API but the CUPS raster "device" was never updated, leading to issues with the default RIPCache setting of 8MB. A quick workaround is to set the cupsd.conf RIPCache value to something larger, e.g. 1024MB: cupsctl RIPCache=1024MB -------------------------------------------------------------------------- This also explains that the root cause is not in CUPS but a non-backward-compatible change within Ghostscript. See comment #16: ------------------------------------------------------------------- cups-1.4.4-set_default_RIPCache_128m.patch enlarges the CUPS upstream default RIPCache from 8m to 128m to avoid various kind of printout failures. ------------------------------------------------------------------- i.e. the changed default RIPCache does of course not solve any kind of printout failures but it can avoid some printout failures. See comment #0 what I wrote in http://www.cups.org/str.php?L3535 ------------------------------------------------------------------ I don't like the idea to let the default RIPCache size depend on the amount of main memory on a particular machine because the RIPCache size must match to the printing requirements and those are independent of the main memory. (Actually it is the other way round: The main memory must be sufficient for the printing requirements.) In other words: I think that printing should work the same on a 256MB main memory machine and on a 2GB main memory machine. The 256MB main memory machine is probably already under full load when processing one single large format photo print job but for each job it should work the same. ------------------------------------------------------------------ and where "jimjensen" wrote on 11:48 Mar 24, 2010 ------------------------------------------------------------------ A RIPCache value of 32m is adequate for printing 13" x 19" (according to my test on a HP B8550), ------------------------------------------------------------------ and what I wrote directly in comment #0 ---------------------------------------------------------------- 128 MB is half of the minimum required RAM (256 MB) and a quater of the recommended RAM (512 MB) for a system where openSUSE is to be installed (according to what I found on a openSUSE 11.2 box) so that "RIPCache 128m" should be a safe default which can be set on any user's system. ---------------------------------------------------------------- A default "RIPCache 1024m" (i.e. 1 GB!) is totally impossible. Summary: The issue is fixed as good as currently possible in CUPS. In some cases the current default RIPCache value of 128m is still insufficient (probably only when printing at high resolution with the Gutenprint driver) so that you may have to specify much more. I enhanced the section about RIPCache accordingly at http://en.opensuse.org/Portal:Printing -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.