Hi again, bug squashers, This is a new issue, unrelated to the August 19th report.
I think there may be a logical parsing error in the following circumstance:
When a '@w{TEXT}' packet extends beyond the 'fill-column' limit, then if
there is a word break ahead of the packet on that line, then logically, the
entire packet should wrap to the next line. This was the case in makeinfo
v:4.13, but not in v:5.1.
The documentation says that “'@w{TEXT}' outputs TEXT, while prohibiting
line breaks *within* TEXT.”
To me, this means that the entire packet should be processed as if it were
a single word.
What do you think?
What might give a clue is that there is an inconsistency in parsing a
'@w{TEXT}' packet at the end of a line, as shown in the example with plain
text contents. This could indicate that the same parsing sequence is
happening in more than one place, and that they don't match, OR that a
search for a word break is (inappropriately) happening within the @w{}.
I have included a simple test file that encloses an xref and then some
plain text within the '@w{}' to demonstrate these problems.
See lines 101-118 of bug_report.texi
Sorry I can't help with the Perl, I'm only an embedded-systems guy. :-)
Mahlon
--
*“You will do foolish things, but do them with enthusiasm.”*
*--Sidonie Gabrielle Colette*
Mahlon R. Smith (马伦 )
Associate Professor, Foreign Languages Department
Beijing University of Technology
Email: [email protected]
makeinfo_bug_report2.tar.bz2
Description: BZip2 compressed data
