Thread: [PATCH]: Building PostgreSQL 7.0.2 on Cygwin 1.1.4 (Take 2)
Please ignore my previous post... On Tue, Sep 26, 2000 at 07:08:53PM +0200, Joost Kraaijeveld wrote: > I read you patch etc. but I do not agree with the linking of msvcrt.lib. > I think that it should be solved with an #ifdef (see below). That way we > do not need any Microsoft libraries. As Joost was kind enough to point out, there was a deficiency with my previous patch. Please ignore the previous patch and use this one instead. Sorry for the confusion. Jason -- Jason Tishler Director, Software Engineering Phone: +1 (732) 264-8770 x235 Dot Hill Systems Corporation Fax: +1 (732) 264-8798 82 Bethany Road, Suite 7 Email: Jason.Tishler@dothill.com Hazlet, NJ 07730 USA WWW: http://www.dothill.com
Attachment
Jason Tishler writes: > [part of patch] > * src/win32/README: Change wording to indicate that un.h is no longer > [...] > $ rm src/win32/un.h # Note that this step is optional > [...] > + $ cp endian.h /usr/include > [...] You can remove endian.h as well if you wish. Pete Forman wrote in thread: [PORTS] NT FAQ needs updating: > endian.h only serves to include sys/param.h, which is in B20 and > 1.1. backend/libpq/pqformat.c is the only file to reference > endian.h and should get its BYTE_ORDER from sys/param.h. BTW, I've just applied your patch (minus endian.h) and successfully built PostgreSQL 7.0.2 from scratch on Cygwin 1.1.4. -- Pete Forman -./\.- Disclaimer: This post is originated Western Geophysical -./\.- by myself and does not represent pete.forman@westgeo.com -./\.- the opinion of Baker Hughes or http://www.crosswinds.net/~petef -./\.- its divisions.
Peter,
On Wed, Sep 27, 2000 at 04:25:46PM +0100, Pete Forman wrote:
> BTW, I've just applied your patch (minus endian.h) and successfully
> built PostgreSQL 7.0.2 from scratch on Cygwin 1.1.4.
Thanks for trying and validating my patch.
Browsing CVS via:
http://www.postgresql.org/cgi/cvsweb.cgi/pgsql/
I am very happy to see last log entries like:
Fixes for Cygwin, with help from Pete Forman
It seems like you trying to get the necessary Cygwin 1.1.4 mods into
CVS. Can you help get the relevant portions of my patch into CVS too?
Thanks,
Jason
--
Jason Tishler
Director, Software Engineering Phone: +1 (732) 264-8770 x235
Dot Hill Systems Corporation Fax: +1 (732) 264-8798
82 Bethany Road, Suite 7 Email: Jason.Tishler@dothill.com
Hazlet, NJ 07730 USA WWW: http://www.dothill.com
I ran the regression tests on the Cygwin PostgreSQL build. All tests except for two passed. So following the instructions found near the beginning of regress.out, I am reporting "any apparent problems" found to the list. The two tests that failed are horology and random. See attached for diffs between the expected and results. My guess is that the horology failures can be safely ignored. Unfortunately, I'm not quite so sure for random. Can someone more knowledgeable with PostgreSQL, in particular running the regression tests, help put my mind at ease? BTW, I'm happy to report that the Cygwin PostgreSQL build passed the plpgsql test. I was most concerned about this one since plpgsql.dll had the most difficult build issue for me to resolve. Thanks, Jason -- Jason Tishler Director, Software Engineering Phone: +1 (732) 264-8770 x235 Dot Hill Systems Corporation Fax: +1 (732) 264-8798 82 Bethany Road, Suite 7 Email: Jason.Tishler@dothill.com Hazlet, NJ 07730 USA WWW: http://www.dothill.com
Attachment
Jason Tishler writes:
> I ran the regression tests on the Cygwin PostgreSQL build. All tests
> except for two passed. So following the instructions found near the
> beginning of regress.out, I am reporting "any apparent problems" found
> to the list.
>
> The two tests that failed are horology and random. See attached for diffs
> between the expected and results. My guess is that the horology failures
> can be safely ignored. Unfortunately, I'm not quite so sure for random.
The horology result is a known failure on platforms that don't properly
support daylight savings time before 1970. Add the following line to the
file resultmap (at the end, for example):
horology/*-*-cygwin*=horology-no-DST-before-1970
then that test should pass. I will do the same thing in the development
branch.
The random test can indeed be safely ignored, the point being that this
test may fail randomly.
--
Peter Eisentraut peter_e@gmx.net http://yi.org/peter-e/
Jason Tishler writes: > It seems like you trying to get the necessary Cygwin 1.1.4 mods into > CVS. Can you help get the relevant portions of my patch into CVS too? Some of the stuff you touched has moved around a bit so evaluation might take a bit, but rest assured that it will be done. -- Peter Eisentraut peter_e@gmx.net http://yi.org/peter-e/
Peter, On Thu, Sep 28, 2000 at 10:42:19AM +0200, Peter Eisentraut wrote: > Jason Tishler writes: > > > It seems like you trying to get the necessary Cygwin 1.1.4 mods into > > CVS. Can you help get the relevant portions of my patch into CVS too? > > Some of the stuff you touched has moved around a bit so evaluation might > take a bit, but rest assured that it will be done. Thanks for your reply and assurances -- I'm quite willing to be patient. Thanks, Jason -- Jason Tishler Director, Software Engineering Phone: +1 (732) 264-8770 x235 Dot Hill Systems Corporation Fax: +1 (732) 264-8798 82 Bethany Road, Suite 7 Email: Jason.Tishler@dothill.com Hazlet, NJ 07730 USA WWW: http://www.dothill.com
On Thu, Sep 28, 2000 at 10:37:32AM +0200, Peter Eisentraut wrote: > The horology result is a known failure on platforms that don't properly > support daylight savings time before 1970. Add the following line to the > file resultmap (at the end, for example): > > horology/*-*-cygwin*=horology-no-DST-before-1970 > > then that test should pass. I will do the same thing in the development > branch. > > The random test can indeed be safely ignored, the point being that this > test may fail randomly. Thanks for the above -- I appreciate you putting my mind at ease. Jason -- Jason Tishler Director, Software Engineering Phone: +1 (732) 264-8770 x235 Dot Hill Systems Corporation Fax: +1 (732) 264-8798 82 Bethany Road, Suite 7 Email: Jason.Tishler@dothill.com Hazlet, NJ 07730 USA WWW: http://www.dothill.com
What do people want done with this? Does it go into the main PostgreSQL tree? > Please ignore my previous post... > > On Tue, Sep 26, 2000 at 07:08:53PM +0200, Joost Kraaijeveld wrote: > > I read you patch etc. but I do not agree with the linking of msvcrt.lib. > > I think that it should be solved with an #ifdef (see below). That way we > > do not need any Microsoft libraries. > > As Joost was kind enough to point out, there was a deficiency with my > previous patch. Please ignore the previous patch and use this one > instead. Sorry for the confusion. > > Jason > > -- > Jason Tishler > Director, Software Engineering Phone: +1 (732) 264-8770 x235 > Dot Hill Systems Corporation Fax: +1 (732) 264-8798 > 82 Bethany Road, Suite 7 Email: Jason.Tishler@dothill.com > Hazlet, NJ 07730 USA WWW: http://www.dothill.com [ Attachment, skipping... ] [ Attachment, skipping... ] [ Attachment, skipping... ] -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Bruce, On Tue, Oct 17, 2000 at 12:27:47AM -0400, Bruce Momjian wrote: > What do people want done with this? Does it go into the main PostgreSQL > tree? On Thu, Sep 28, 2000 at 10:42:19AM +0200, Peter Eisentraut wrote: > Jason Tishler writes: > > > It seems like you trying to get the necessary Cygwin 1.1.4 mods into > > CVS. Can you help get the relevant portions of my patch into CVS > > too? > > Some of the stuff you touched has moved around a bit so evaluation > might take a bit, but rest assured that it will be done. I was under the impression that my patch was being hand merged into the CVS since the tree has changed since 7.0.2. Jason -- Jason Tishler Director, Software Engineering Phone: +1 (732) 264-8770 x235 Dot Hill Systems Corporation Fax: +1 (732) 264-8798 82 Bethany Road, Suite 7 Email: Jason.Tishler@dothill.com Hazlet, NJ 07730 USA WWW: http://www.dothill.com
Jason Tishler writes: > I was under the impression that my patch was being hand merged into the > CVS since the tree has changed since 7.0.2. I was under the impression that part of your patch was going to abandon compatibility with Cygwin 1.0 and B20. We can't do that. -- Peter Eisentraut peter_e@gmx.net http://yi.org/peter-e/
Peter,
On Tue, Oct 17, 2000 at 07:13:05PM +0200, Peter Eisentraut wrote:
> Jason Tishler writes:
>
> > I was under the impression that my patch was being hand merged into the
> > CVS since the tree has changed since 7.0.2.
>
> I was under the impression that part of your patch was going to abandon
> compatibility with Cygwin 1.0 and B20.
The above was certainly not my intention. My patch only changed the
source just enough to get it to build with Cygwin 1.1.4 -- there were no
functional changes (except preventing the possible psql crashes).
Unfortunately I do not have any Cygwin 1.0 and b20 setups, so I can't
determine whether or not my patched version will build under those
environments. If the source builds, then it should should function as
before.
Are there any kind souls out there still using Cygwin 1.0 or b20 and
willing to help? If so, please apply my patch to PostgreSQL 7.0.2 and
then try to build it. If it builds, then run the regression tests and
post the results.
> We can't do that.
I can appreciate the above, but I must point out the following:
Cygnus no longer supports b20
Cygnus no longer sells the Cygwin 1.0 CD
My suggestion (and hope) is that the current PostgreSQL CVS source should
build OOTB with a "current" Cygwin. And that patches are produced for
the older Cygwin versions, if necessary.
Jason
--
Jason Tishler
Director, Software Engineering Phone: +1 (732) 264-8770 x235
Dot Hill Systems Corporation Fax: +1 (732) 264-8798
82 Bethany Road, Suite 7 Email: Jason.Tishler@dothill.com
Hazlet, NJ 07730 USA WWW: http://www.dothill.com
If you can, please check our current snapshot to see how it works on your platform. Thanks. > Please ignore my previous post... > > On Tue, Sep 26, 2000 at 07:08:53PM +0200, Joost Kraaijeveld wrote: > > I read you patch etc. but I do not agree with the linking of msvcrt.lib. > > I think that it should be solved with an #ifdef (see below). That way we > > do not need any Microsoft libraries. > > As Joost was kind enough to point out, there was a deficiency with my > previous patch. Please ignore the previous patch and use this one > instead. Sorry for the confusion. > > Jason > > -- > Jason Tishler > Director, Software Engineering Phone: +1 (732) 264-8770 x235 > Dot Hill Systems Corporation Fax: +1 (732) 264-8798 > 82 Bethany Road, Suite 7 Email: Jason.Tishler@dothill.com > Hazlet, NJ 07730 USA WWW: http://www.dothill.com [ Attachment, skipping... ] [ Attachment, skipping... ] [ Attachment, skipping... ] -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Bruce, On Fri, Jan 19, 2001 at 04:48:40PM -0500, Bruce Momjian wrote: > If you can, please check our current snapshot to see how it works on > your platform. Thanks. Both the snapshot dated Sun Jan 21 09:04:00 2001 and a fresh CVS checkout built OOTB and passed all regression tests. Thanks, Jason -- Jason Tishler Director, Software Engineering Phone: +1 (732) 264-8770 x235 Dot Hill Systems Corp. Fax: +1 (732) 264-8798 82 Bethany Road, Suite 7 Email: Jason.Tishler@dothill.com Hazlet, NJ 07730 USA WWW: http://www.dothill.com