This one is for the LaTeX gurus in the group:
I have 18 Beamer presentations generated the usual way (created in LyX,
output using pdflatex). To service a student request <sigh>, I loaded
each one into Acrobat Pro 7 and used the menu option to enable
commenting by Acrobat Reader. This leads Acrobat Pro to do arcane
things (measured by a progress bar) and then write out a modified copy
of the file.
On 14 files, this worked fine; on four, it aborted midway with an error
message that Acrobat could not read the file. The message is probably a
bit misleading, as both Acrobat Pro and Acrobat Reader displayed all
pages of each file correctly on screen. Poking around with Pro's
"preflight" tool (which I think is more or less analogous to chklatex),
I got a message on the four recalcitrant files that the document
structure was corrupt.
So I returned to LyX and generated those four presentations using the
straight PDF output option, rather than pdflatex. All four files passed
visual muster, and Acrobat Pro was perfectly happy to enable commenting
on all four. Other than considerable time wasted hunting the bug and
work-around, the only (minor) problem is that the non-pdflatex versions
are *much* bigger than their pdflatex cousins.
I can't find anything that differentiates the winners from the losers
here. It's not the presence of graphics (at least one of the four had
none), it's not the presence of math insets (several of the 14 that
worked had math insets), ...
Does anyone have any idea what pdflatex might be doing to offend the
Gods of Document Structure (particularly in a way that does not stop
Acrobat or Acrobat Reader from displaying the document correctly)?
TIA,
Paul