Date: Tue, 4 Feb 2003 14:42:05 +0200 From: Alexey Zelkin <phantom@FreeBSD.org.ua> To: "Georg-W. Koltermann" <g.w.k@web.de> Cc: Max Khon <fjoe@iclub.nsu.ru>, freebsd-java@freebsd.org Subject: Re: jdk 1.4.1 patchset2 is coming ... Message-ID: <20030204144205.A77607@phantom.cris.net> In-Reply-To: <1044312220.281.15.camel@bat.localnet>; from g.w.k@web.de on Mon, Feb 03, 2003 at 11:43:27PM %2B0100 References: <20030203134010.A69472@phantom.cris.net> <20030204012148.A23926@iclub.nsu.ru> <20030203213233.A72514@phantom.cris.net> <1044312220.281.15.camel@bat.localnet>
next in thread | previous in thread | raw e-mail | index | archive | help
hi, On Mon, Feb 03, 2003 at 11:43:27PM +0100, Georg-W. Koltermann wrote: > What does this mean for the -current branch? Has the libc_r stuff been > included in 5.0-R, or will I need to CVSup to current -current? There's simple (but important) patch present. It allowed me to make FreeBSD stack threads management simple and clean (+35, -420 in code lines, ~3-5% in performance numbers on heavy java threads applications). It was committed to main source tree yesterday, therefore it won't appear in 5.0-RELEASE. Actually if you're running post-07-Jan-2003 system (which include 5.0-RELEASE) you'll need to apply patch distributed with patchset2 and recompile/install libc_r only. cd /usr/src/lib/libc_r/uthread patch < /path/patch-pthread_get_attr_np.c make make install And then go to build/use JDK. > P.S.: Hey guys I am just upgrading my -stable system to 5.0-R *because* > JDK1.4 was reported not to run well on -stable. Now you announce you > have done all further porting work on -stable and only lightly tested on > -current. Grumble. > > A while earlier, last year, I ran -current because billh did all of his > HotSpot work on -current. Then I converted to -stable as life in > -current became too hot (too much work, I mean) to me and as the libc_r > patches for HotSpot had been backported. > > I feel like chasing a permanent current-stable-current-stable spiral of > Java development in FreeBSD. Shouldn't we all agree on *one* primary os > branch for Java, announce that, and stick with it? All my work was always primarily targeted for production quality system (currently 4.x-STABLE). I also paid attention to -CURRENT too, at least at level of compile test and basic functionality testing (and will pay more since it's going to become -STABLE relatively soon). Previous 1.4.x patchset was directed to both these system, but have had unresolved problems with -STABLE. While my work on switching to 1.4.1 I have found reasons of these problem and (hopefully) fixed them. > Am Mo, 2003-02-03 um 20.32 schrieb Alexey Zelkin: > > hi, > > > > On Tue, Feb 04, 2003 at 01:21:49AM +0600, Max Khon wrote: > > > > > On Mon, Feb 03, 2003 at 01:40:10PM +0200, Alexey Zelkin wrote: > > > > > > > KEEP IN MIND: 1.4.1-p2 depends on version of pthread_attr_get_np() function > > > > functionality which hit tree very recently. You'll need to update your > > > > libc_r with patch distributed along with 1.4.1-p2 before start to build > > > > and use (or cvsup/update your system to 04-Feb-2003 state) > > > > > > 04 Jan 2003, actually > > > > No. I actually meant today or later. I.e. accounting my today's commit > > to libc_r). JVM/1.4.1 heavily depend on functionality added in this commit. > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-java" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030204144205.A77607>