Bugs item #606, was opened at 2006-09-16 20:28

Status: Closed Priority: 2 Submitted By: Hartmut Henkel (hhenkel) Assigned to: Hartmut Henkel (hhenkel) Summary: Forbidden ... while scanning ??? Category: Primitives Group: v1.40.0 Resolution: Fixed

Initial Comment: \pdftrailer {abc \bye gives ! Forbidden control sequence found while scanning text of \pdftrailer. but \pdfximage {abc \bye or \pdfliteral {abc \bye give ! Forbidden control sequence found while scanning text of ???. Instead of these three question marks there should be the name of the primitive. ----------------------------------------------------------------------

Comment By: The Thanh Han (hanthethanh) Date: 2009-12-01 21:29

Message: fixed in 1.40.10 ---------------------------------------------------------------------- Comment By: Taco Hoekwater (taco) Date: 2009-04-07 06:28 Message: Attached is what I would do: patch scan_keyword() to restore cur_cs in the 'false' case. ---------------------------------------------------------------------- Comment By: Martin Schröder (oneiros) Date: 2007-05-27 15:40 Message: Logged In: YES user_id=421 Hartmut, can you please check this again for 1.50? ---------------------------------------------------------------------- Comment By: Martin Schröder (oneiros) Date: 2006-09-21 12:22 Message: Logged In: YES user_id=421 I've backed out the change; this is low priority and probably needs more changes. ---------------------------------------------------------------------- Comment By: Hartmut Henkel (hhenkel) Date: 2006-09-17 19:10 Message: Logged In: YES user_id=929 yes, the patch works, thanks a lot! But there are in fact many more \pdf... primitives that suffer from this problem, actually any with a scan_keyword call before the scan_pdf_ext_toks. After the way how it's done in @=, it seems that a new variable save_cur_cs is not needed if one uses k for this purpose. Needs change of a few k uses to j. But problematic is \pdfoutline as it has another scan_pdf_ext_toks for the attr keyword. Wouldn't these many changes be too risky for 1.40.0? ---------------------------------------------------------------------- Comment By: Martin Schröder (oneiros) Date: 2006-09-17 16:34 Message: Logged In: YES user_id=421 The patch was missing. Fixed. ---------------------------------------------------------------------- Comment By: Martin Schröder (oneiros) Date: 2006-09-17 14:30 Message: Logged In: YES user_id=421 The cause seems to be that some scan_functions loose the meaning of cur_cs, which leads to the ??? when reaching . The attached (IMHO crude) patch fixes this for the cases you mentioned, but there are probably other cases not found... ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=606&group_id=106