Hrvoje Niksic writes:
Colin Rafferty <craffert(a)ms.com> writes:
> > This scheme will fail to work for people who use `-q -no-site-file'.
> > Maybe you should modify the code to initialize
> > `internal-doc-file-name' from C to the correct value.
>
> Ugh. I don't really want to start messing with the code that
> heavily.
I wouldn't call that "heavy." That's what we ship
the sources for,
after all. :-)
Seriously, the change is pretty trivial, and results in a more
robust
executable. If you extend it a little bit further, your changes might
be of use to others -- mail us the patch, and why not apply it?
I hadn't thought of it that way. Now you've got me thinking about a
good way to specify the doc-file name.
Maybe a configure option? There are a lot of options already, but I
really don't want to generate a unique doc-file name unnecessarily.
On the other hand, most people don't even realize this is a problem,
so maybe this really should be a forced option.
So I create a bitmask of all the configure options set, generate a
uuencoded string, and the doc-file name is XL4J.DOC. This may
actually work. If I'm only allowed [A-Z0-9], I can sqeeze in 5
options per character. With an 8+3 filename limit, that's 40 options.
> I guess if our users start with `-q -no-site-file', they get
what
> they deserve.
Uh-oh.
Oh well.
--
Colin