On 11/11/2009 06:38 PM, rgheck wrote:
There's still a problem, at least in trunk, with the elyxer detection
code, though the actual bug, I think, is somewhere else.
The problem with the code is that we should NOT be refusing to check
for LaTeX-->HTML converters simply because we have found a LyX-->HTML
converter. Once we have the ability to allow the user to choose among
various options, then we want to know if there is an alternative, even
if elyxer becomes the default. Indeed, though it's hard to imagine,
the LaTeX-->HTML converter could possibly be used in some other
converter chain, i.e., other than LyX-->HTML, e.g.:
DocBook-->LaTeX-->HTML. Not that I'd recommend it.
It's easy enough to re-instate such detection, but then the problem is
that elyxer is not used for HTML output. Instead, the LaTeX-->HTML
converter gets used. This seems wrong, because LyX ought to be using
the shortest path it can, which would be LyX-->HTML, not
LyX-->LaTeX-->HTML.
Does anyone have any idea why the LyX-->HTML converter isn't being
found first in this case?
OK, I do see why this is happening now. It's because we look for a way
to convert the file to FMT from the various backend formats we can
produce, with the order being: File Format (LaTeX/DocBook), Text, XHTML,
LyX. Since we find a way to convert it to HTML from LaTeX, the
LyX-->HTML route never gets checked.
Something needs to be done about this, as I said. But I don't know what.
We can't change the order of the backend checks, because then (if elyxer
is installed) we will find the LyX-->HTML-->LaTeX-->PDF route and export
PDF that way, which is stupid. Even checking all the backends and
finding the shortest isn't guaranteed to work.
Any thoughts?
Richard
