inn-stable on HP/UX

[Russ Allbery]

Alex Kiernan <alexk at alexk.eng.demon.net> writes:

> If the tradindexed code didn't mmap & lock didn't it might work. One
> thought I guess is what code paths is the locking protecting against?
> Just innd and expireover?

And tdx-util, and nnrpd will acquire a lock if the first time it tries, it
ends up seeing a data file that doesn't match the index file.  That lets
it cleanly wait for expireover if expireover is currently rebuilding that
group.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>

    Please send questions to the list rather than mailing me directly.
     <http://www.eyrie.org/~eagle/faqs/questions.html> explains why.