X-Authentication-Warning: delorie.com: mail set sender to djgpp-workers-bounces using -f X-Recipient: djgpp-workers AT delorie DOT com Date: Mon, 21 Apr 2014 15:53:54 +0300 From: Eli Zaretskii Subject: Re: ctime.c changes add about 4.5k more size In-reply-to: X-012-Sender: halo1 AT inter DOT net DOT il To: djgpp-workers AT delorie DOT com Message-id: <83mwfedesd.fsf@gnu.org> References: <83vbu3cc6b DOT fsf AT gnu DOT org> Reply-To: djgpp-workers AT delorie DOT com Errors-To: nobody AT delorie DOT com X-Mailing-List: djgpp-workers AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk > Date: Mon, 21 Apr 2014 15:39:35 +0300 > From: Ozkan Sezer > > FWIW, here is the long diff between versions 1.10 and 1.11 of ctime.c: > http://www.delorie.com/bin/cvsweb.cgi/djgpp/src/libc/ansi/time/ctime.c > http://www.delorie.com/bin/cvsweb.cgi/djgpp/src/libc/ansi/time/ctime.c.diff?r1=1.10&r2=1.11 > ... which adds more size to an already long time code. The question is where do 5KB of extra code come from. Some changes in the diffs are just reshuffling of the same code, others add what cannot possibly generate so much more instructions. One way of finding the potential culprits is to compare the output of "nm -A ctime.o" on both versions of compiled ctime.c. It's possible that some static function (those marked with the lower-case "t") either grew in size or were added in the new version. Armed with this knowledge, we might take a good look at those few functions, and see if we can do without them, or maybe simplify them. However, in general computing time_t value is a tricky business, so it's quite possible the extra code is just some added feature (e.g., I see some kind of quoting that is now being supported) or a bugfix, and you cannot really do without those, if you want to be able to read the latest time-zone files.