Further pstops corruption due to lack of linelen assignment

[michaelrsweet]

Version: 1.2-current
CUPS.org User: nix

[Replicated against 1.2rc1, which does not appear in the version pulldown]

Output from dvips and probably other software is corrupted by the pstops filter. (I have only seen additional garbage added immediately after %%BeginSetup:, but other corruption should be possible.)

The cause is a failure to propagate linelen up to copy_dsc() after copying the prolog; the setup section also lacks the same. So the first line in the setup section is assumed to be as long as the first line in the prolog, and... mess results.

Patch attached.

[michaelrsweet]

CUPS.org User: mike

Fixed in Subversion repository.

[michaelrsweet]

"linelen-fix.diff":

Index: filter/pstops.c

--- filter/pstops.c (revision 5343)
+++ filter/pstops.c (working copy)
@@ -738,14 +738,14 @@
*/

fprintf(stderr, "DEBUG: Before copy_prolog - %s", line);

• copy_prolog(fp, doc, ppd, line, linelen, linesize);
• linelen = copy_prolog(fp, doc, ppd, line, linelen, linesize);

/*

• Then the document setup section...
  */

fprintf(stderr, "DEBUG: Before copy_setup - %s", line);

• copy_setup(fp, doc, ppd, line, linelen, linesize);
• linelen = copy_setup(fp, doc, ppd, line, linelen, linesize);

/*

• Then process pages until we have no more...