FAQ
I have finished updating about half of Postgres' builtin functions to
the new style of fmgr interface. At this point, everything that accepts
any pass-by-value datatypes is converted; the remaining work is for
functions that use only pass-by-reference datatypes, and therefore
receive only pointer arguments.

This should already take care of function-call-related portability
problems on many platforms. In particular these changes should
eliminate the need for Ryan Kirkpatrick's Linux/Alpha patches, and
should also solve the known problems on PPC builds with optimization
higher than -O0. We might be able to increase the optimization level
on other platforms that have had trouble with function-call
optimizations, too.

I'd like to get the current code tested by some people with Alphas
(or other 64-bit platforms), PPCs, and anything else that has had
optimization-related problems. You can get the 7.1 development code
from our CVS server, or use the current daily-snapshot tarball (see
ftp://ftp.postgresql.org/pub/dev/).

regards, tom lane

Search Discussions

  • Bruce Momjian at Jun 14, 2000 at 5:02 pm

    eliminate the need for Ryan Kirkpatrick's Linux/Alpha patches, and
    should also solve the known problems on PPC builds with optimization
    higher than -O0. We might be able to increase the optimization level
    I have change the -O0 to -O2.

    --
    Bruce Momjian | http://www.op.net/~candle
    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
  • Ryan Kirkpatrick at Jun 16, 2000 at 1:14 pm

    On Wed, 14 Jun 2000, Tom Lane wrote:

    I have finished updating about half of Postgres' builtin functions to
    the new style of fmgr interface. At this point, everything that accepts
    any pass-by-value datatypes is converted; the remaining work is for
    functions that use only pass-by-reference datatypes, and therefore
    receive only pointer arguments.
    Cool! I need to study up a bit on fmgr, but anything that help
    Linux/Alpha is a good thing. :)
    This should already take care of function-call-related portability
    problems on many platforms. In particular these changes should
    eliminate the need for Ryan Kirkpatrick's Linux/Alpha patches, and
    Not all of my Linux/Alpha patches are related to fmgr, there are a
    few related to s_lock and such (which can probably be put into the source
    tree with #ifdefs). But this should take care of the majority of the
    Linux/Alpha patch.
    I will download the snapshot today at work (where I have "real"
    bandwidth :) and test things out this weekend. I should have a report by
    Monday. Maybe I will even have a patch that can be safely applied to
    the development source tree. :) TTYL.

    ---------------------------------------------------------------------------
    "For to me to live is Christ, and to die is gain." |
    --- Philippians 1:21 (KJV) |
    ---------------------------------------------------------------------------
    Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |
    ---------------------------------------------------------------------------
  • Tom Lane at Jun 16, 2000 at 2:46 pm

    Ryan Kirkpatrick writes:
    Not all of my Linux/Alpha patches are related to fmgr, there are a
    few related to s_lock and such (which can probably be put into the source
    tree with #ifdefs).
    Right-o. Up to now we haven't worried about that, but now that the
    fundamental problem is fixed (I hope), we can start cleaning up the
    smaller details so that Linux/Alpha will be supported out-of-the-box.

    regards, tom lane
  • Ryan Kirkpatrick at Jun 18, 2000 at 10:43 pm

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
grouppgsql-ports @
categoriespostgresql
postedJun 14, '00 at 4:55p
activeJun 18, '00 at 10:43p
posts5
users3
websitepostgresql.org
irc#postgresql

People

Translate

site design / logo © 2022 Grokbase