Thread: Resultmap for FreeBSD 4.7
It appears i386 FreeBSD 4.7, along with the rest of the 4.x series requires the positive-zeros geometry output. Attached is the switch. I assume that 4.6 and recent alpha releases also require positive-zeros. -- Rod Taylor <rbt@rbt.ca>
Attachment
Rod Taylor <rbt@rbt.ca> writes: > It appears i386 FreeBSD 4.7, along with the rest of the 4.x series > requires the positive-zeros geometry output. Larry Rosenman submitted a directly contradictory report less than a month ago ... http://fts.postgresql.org/db/mw/msg.html?mid=1007251 regards, tom lane
--On Monday, November 25, 2002 22:17:48 -0500 Tom Lane <tgl@sss.pgh.pa.us> wrote: > Rod Taylor <rbt@rbt.ca> writes: >> It appears i386 FreeBSD 4.7, along with the rest of the 4.x series >> requires the positive-zeros geometry output. > > Larry Rosenman submitted a directly contradictory report less than > a month ago ... > http://fts.postgresql.org/db/mw/msg.html?mid=1007251 I thought I remembered this...... Rod, Do you have any CPU optimization flags set in /etc/make.conf? What processor are you running on? (my answers are: no P-IIIm). LER > > regards, tom lane -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 972-414-9812 E-Mail: ler@lerctr.org US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749
> Rod, > Do you have any CPU optimization flags set in /etc/make.conf? > > What processor are you running on? > > (my answers are: > no > P-IIIm). This was a Celeron based laptop (Thinkpad X20). 4.7 release upgraded from 2.2.2 -- so I imagine a little bit of cruft has been left around. For the curious, I have done several disk-copies for laptop upgrades. No special compile settings. (./configure; make install) Ok.. Got a few more 4.7 boxes laying around: Compaq Quad Xeon (700Mhz) [fresh install two weeks ago] - positive-zeros Low end Compaq (900Mhz PIII) [upgraded from 4.3] - positive-zeros FreeBSD 4.6 on Dual 400Mhz celerons [upgraded from 4.0]- negative-zeros -- no-diffs -- Looks like 4.7 reverses a patch that went into 4.6. Is it possible you were testing 4.7-STABLE rather than RELEASES? It may have gone back in by that point. -- Rod Taylor <rbt@rbt.ca>
Attachment
Yes, I'm running -STABLE. --On Monday, November 25, 2002 23:16:04 -0500 Rod Taylor <rbt@rbt.ca> wrote: >> Rod, >> Do you have any CPU optimization flags set in /etc/make.conf? >> >> What processor are you running on? >> >> (my answers are: >> no >> P-IIIm). > > This was a Celeron based laptop (Thinkpad X20). 4.7 release upgraded > from 2.2.2 -- so I imagine a little bit of cruft has been left around. > > For the curious, I have done several disk-copies for laptop upgrades. > > No special compile settings. (./configure; make install) > > > Ok.. Got a few more 4.7 boxes laying around: > > Compaq Quad Xeon (700Mhz) [fresh install two weeks ago] - positive-zeros > > > Low end Compaq (900Mhz PIII) [upgraded from 4.3] - positive-zeros > > > FreeBSD 4.6 on Dual 400Mhz celerons [upgraded from 4.0]- negative-zeros > -- no-diffs -- > > > > Looks like 4.7 reverses a patch that went into 4.6. Is it possible you > were testing 4.7-STABLE rather than RELEASES? It may have gone back in > by that point. > > -- > Rod Taylor <rbt@rbt.ca> -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 972-414-9812 E-Mail: ler@lerctr.org US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749
Rod Taylor writes: > Looks like 4.7 reverses a patch that went into 4.6. Is it possible you > were testing 4.7-STABLE rather than RELEASES? It may have gone back in > by that point. No, the behavior was fixed (as in: it correctly shows negative zeros now) in the stable branch after the 4.7 release. Unfortunately, there is no version number distinguishing the release and the branch. -- Peter Eisentraut peter_e@gmx.net
Actually, you can tell from __FreeBSD_Version in /usr/include/sys/param.h LER --On Wednesday, November 27, 2002 00:51:58 +0100 Peter Eisentraut <peter_e@gmx.net> wrote: > Rod Taylor writes: > >> Looks like 4.7 reverses a patch that went into 4.6. Is it possible you >> were testing 4.7-STABLE rather than RELEASES? It may have gone back in >> by that point. > > No, the behavior was fixed (as in: it correctly shows negative zeros now) > in the stable branch after the 4.7 release. Unfortunately, there is no > version number distinguishing the release and the branch. > > -- > Peter Eisentraut peter_e@gmx.net -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 972-414-9812 E-Mail: ler@lerctr.org US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749