Date: Thu, 20 Mar 1997 16:39:03 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: imp@village.org (Warner Losh) Cc: terry@lambert.org, james@wgold.demon.co.uk, hackers@FreeBSD.ORG Subject: Re: Barb problem, FOUND Message-ID: <199703202339.QAA14590@phaeton.artisoft.com> In-Reply-To: <E0w7oof-0003K9-00@rover.village.org> from "Warner Losh" at Mar 20, 97 01:52:04 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> At the time we were doing OI, nearly *ALL* of the compilers were so > afflicated. Sun's, Centerline's, Lucid's, cfront, Microsoft's, Dec's, > HP's and IBM's. They all sucked at doing inline virtuals, creating > multiple copies for them. There were *NO* compilers that we could > have used that were compainle with the Sun compiler on SunOS (our > primary market for this library). Oh, and g++ wouldn't even compile > OI. Oregon Software's C++ as well? > When all or nearly all of the compilers you have to deal with don't > grok a construct, it is a bad construct. Sometimes it isn't as simple > as you paint thing Terry. The compilers weren't emitting them into seperate segments? You could post-process the object files (a "prelink") if it was. Otherwise, what can I say... you live with the overhead of the emitted static inlines and complain to the vendor. It seems very similar to the issues of statically linked Motif binaries and other large space wasters... ie: NetScape, et. al.. Regards, Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199703202339.QAA14590>