Re: [PATCH] history locking with fcntl

[Vincent Lefevre <vincent@vinc17.org>]

On 2008-04-18 19:31:55 -0700, Wayne Davison wrote:
> > I'll try it.
> 
> Much appreciated.

I've done a few tests and don't get history corruption.

But my patch avoided a possible problem in rename() when there is a
tmpfile. With your patch, one still has:

        if (fclose(out) < 0 && ret >= 0)
            ret = -1;
        if (ret >= 0) {
            if (tmpfile) {
                if (rename(tmpfile, unmeta(fn)) < 0)
                    zerr("can't rename %s.new to $HISTFILE", fn);
                free(tmpfile);
                tmpfile = NULL;
            }

Note that since out has been closed, tmpfile is no longer locked.
The target file isn't locked either.

Now, I don't think that rename() can be cached. So, I'd say that there
should be no problems under "normal" conditions.

-- 
Vincent Lefèvre <vincent@vinc17.org> - Web: <http://www.vinc17.org/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.org/blog/>
Work: CR INRIA - computer arithmetic / Arenaire project (LIP, ENS-Lyon)