From: Erik Faye-Lund Date: Tue, 13 Dec 2011 18:52:34 GMT Subject: Re: [PATCH/RFC 1/2] wrapper: supply xsetenv Message-ID: In-Reply-To: <7vwra0uxqo.fsf@alter.siamese.dyndns.org> On Tue, Dec 13, 2011 at 7:34 PM, Junio C Hamano wrote: > Jeff King writes: > >> On Tue, Dec 13, 2011 at 01:10:26PM +0100, Erik Faye-Lund wrote: >> >>> +int xsetenv(const char *name, const char *val, int overwrite) >>> +{ >>> +    if (setenv(name, val, overwrite)) >>> +            die_errno("setenv failed"); >>> +} >> >> It probably doesn't matter, because the error condition is almost >> certainly just "oops, we ran out of memory". But you could also print >> the name of the variable being set, which may give the user a clue to >> some misconfiguration (e.g., trying to put some extremely long value >> into the environment). > > Do we have enough memory to format that message in that situation ;-)? We could. Running out of environment space is not the same as running out of memory. For instance, Windows has a maximum environment size of 32 kB. Older Linux kernels maxed out at 128 kB. So I think it's a good idea to at least try. The worst that can happen is another, less descriptive error, no?