Hi, this is upstream - I agree with Martin-Éric that you should definitely file this bug against gs-esp, too, since this is where the problem originates.
Nevertheless thanks for this hint at the gs options. I will review them and change the CUPS-PDF defaults accordingly where necessary. Volker On Wed, 2007-05-02 at 11:28 +0300, =?UTF-8?Q? Martin-=C3=89ric?= Racine wrote: > Thanks for your report. > > "Critical: causes serious data loss, or introduces a security hole on > systems where you install the package." > > Did the above issue introduce a security hole or cause serious data loss? > I seriously doubt that it did. Thus, I'm bringing severity down to important. > > PS: I would tremendously appreciate it if you filed a bug against > gs-esp for this issue, as this seems to be where the real problem is. > Thanks! > > On 5/2/07, Christopher Zimmermann <[EMAIL PROTECTED]> wrote: > > Package: cups-pdf > > Version: 2.4.2-3 > > Severity: critical > > Tags: patch > > Justification: breaks the whole system > > > > Hi! > > > > When printing certain files via cups-pdf, gs runs forever using all cpu > > and slowly eating all memory. In consequence the system becomes > > unresponsive after a few seconds and therefore unusable. > > If you think this bug is not critical, please tell me, so I know better > > next time. > > A typical example in our environment would be a MS Excel user printing > > to the cups-pdf printer using dim borderlines. If you'd like an example > > .ps file I can provide one. > > Actually this seems to be a bug in gs-esp, but since I managed to fix it > > in cups-pdf, I file it against cups-pdf. > > > > I could fix it by adding -c .setpdfwrite in the config: > > GSCall %s -q -dCompatibilityLevel=%s -dNOPAUSE -dBATCH -dSAFER > > -sDEVICE=pdfwrite -sOutputFile="%s" -dAutoRotatePages=/PageByPage > > -dAutoFilterColorImages=false -dColorImageFilter=/FlateEncode > > -dPDFSETTINGS=/prepress -c .setpdfwrite -f %s > > > > I removed the -c save pop, since /usr/share/doc/gs-esp/Use.htm > > [Improving Performance] suggests that it is not needed. > > > > I don't really understand why this fix works and probably this has to be > > fixed in gs, too. But the default of gs should be changed anyway as far > > as I understand the -c .setpdfwrite option. > > I don't think it is the same bug as #267423 since for me gs never > > finishes. > > > > > > Kind Regards, > > > > Christopher Zimmermann > -- Volker Christian Behr Experimentelle Physik V (Biophysik), Physikalisches Institut Universitaet Wuerzburg, Am Hubland, 97074 Wuerzburg, Germany Office: Room F-069a +49-931-888-5766 (phone) +49-931-888-5851 (fax)