new db_lock.c (was: Re: QDBM and transactions)

[Matthias Andree]

On Sat, 06 Nov 2004, Pavel Kankovsky wrote:

> On Tue, 2 Nov 2004, Matthias Andree wrote:
> 
> > I believe I've fixed this now. Could you have another look at the
> > current db_lock.c file to see if the signal handler is safe now?
> 
> It appears to ok now.
> 
> Two additional minor imperfections:
> 1. there is no check for read() errors (e.g. EIO) in check_zombies(),

There is one: on read() errors, the read() return value will not
match sizeof(cell), so we break out of the loop, with r still -1.
This will then trigger a return value of -1, unless the lseek already
returns -1. Looks fine to me.

> 2. "\n" is missing at the end of check_lock() error text.

Fixed now in CVS.

-- 
Matthias Andree