Re: HEAD build failure

[Andreas Gustafsson <gson%gson.org@localhost>]

Luke Mewburn wrote:
> The patch seems fine, and if it has resolved the problem, then commit it.

Done.

It's hard to know conclusively that the problem is resolved because
it's a random failure, but I have done a number of builds with the
patched version and the problem hasn't recurred.

> (As a matter of style I'd check the return values "!= -1" instead of "< 0",
> but that's a personal preference.)

I didn't write the "< 0", I copied it from mkdir(1).  If it's good enough
for mkdir, it's good enough for install. :)

Note that there is still another known race condition in install(1)
which I reported in PR bin/41512.  I don't have a patch for that one.
-- 
Andreas Gustafsson, gson%gson.org@localhost