Date: Thu, 12 Feb 1998 06:03:41 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: ac199@hwcn.org Cc: jkh@time.cdrom.com, hoek@hwcn.org, hackers@FreeBSD.ORG Subject: Re: ftp://freebsd.org - RIP Message-ID: <199802120603.XAA03786@usr02.primenet.com> In-Reply-To: <Pine.BSF.3.96.980211234712.197A-100000@localhost> from "Tim Vanderhoek" at Feb 11, 98 11:49:05 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > > The last will get a little-to-significantly easier when we stop > > > telling people to upload new ports there... > > > > Where should they upload them then? ;-) > > They should put them in the pr so that 1) there is a permanent > record, 2) incoming doesn't get over-cluttered, 3) they're easy > to access and the committer doesn't have to worry about making > sure some silly 2k file in incoming/ is removed. > > Most ports are pretty small. What about ports that aren't small? Whistle has acapd and ldap both working now. Each required significant fixes to the user space pthreads. The acapd required modifications to the SGI/HP/MOSCOW-SPARC-COMPUTING STL to support non-static mutex initializers. The threads modifications for per threads exception stack registration require gcc 2.8.0 to allow STL exception handling work per thread; otherwise you must #ifdef the code that makes them work out (and then acapd will not work on FreeBSD). The egcs (Cygnus) compiler is kind of broken in the way the do thread exception stacks: they require you to choose at the time you create the compiler, instead of at runtime. The ldap (which I worked on, along with the minor Pthreads draft 4 support in STL, and a tiny amount on the ioctl and _thread_sys_* error returns in pthreads) requires significant patches. In effect, it requires *all* the patches from Critical Angle, Inc., except the SSL stuff which is non-exportable, and then patches over and above that because of the Pthreads draft 4 create_thread argument 2 being a pointer to an attr pointer instead of an attr pointer, as in the final draft. And additional changes because of some invalid assumptions about FD_SET/FD_ZERO. These are *not* 1-2k patches. The stuff Jeremey did on making signal handling (mostly) work in pthreads, along, is a rather large diff (though Julian has been committing the changes to pthreads as fast as they are made, so they wouldn't need rolled in). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199802120603.XAA03786>