Date: Wed, 20 Aug 1997 16:17:07 -0700 (PDT) From: Sean Eric Fagan <sef@Kithrup.COM> To: jkh@time.cdrom.com, terry@lambert.org Cc: hackers@FreeBSD.ORG, julian@whistle.com Subject: Re: [Fwd: Re: Please Help Me Understand dlopen()] Message-ID: <199708202317.QAA10222@kithrup.com>
next in thread | raw e-mail | index | archive | help
>> And not much of a hardship since one of the major arguments in favor >> of going to ELF is that it'd let us abandon some of our older, molding >> items in the toolchain in favor of the more actively maintained GNU >> code (which also dropped support for our a.out format some time >> back). ;-) >That is Sean's point, I think. To handle the transition piecemeal, >as some seem to want to do (it;s a compromise position, but if a >compromise is necessary to let some people remain backwards a bit >longer so everyone can go foraward, so be it), would require adding >the a.out support back into binutils. Actually, that was only *part* of my point. (I've got a binutils here. It configures for freebsd, but doesn't seem to recognize the a.out. I'll keep poking at it.) My main point was that we would pretty much be wedded to the GNU binutils in order to produce both formats. This, as most of you know, doesn't bother me at all -- they're there, they work, they're freely distributable. Other people have their own problems. One problem is that libbfd is pretty darned large. (As one would expect from a library that has the guts of nm, ar, ranlib, and a few others in it.) Making it a shared library will, of course, help considerably, and that's fine. But I want people to be aware of this. Sean.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199708202317.QAA10222>