Horton

Re: [RFC] building postgres with meson - v11

Peter Eisentraut <peter.eisentraut@enterprisedb.com>

View thread View on postgr.es
From: Peter Eisentraut <peter.eisentraut@enterprisedb.com>
To: Andres Freund <andres@anarazel.de>, pgsql-hackers@postgresql.org, samay sharma <smilingsamay@gmail.com>, Nazir Bilal Yavuz <byavuz81@gmail.com>, Andrew Dunstan <andrew@dunslane.net>
Date: 2022-08-17T13:50:23Z
Lists: pgsql-hackers

Attachments

On 11.08.22 02:20, Andres Freund wrote:
> Attached is a new version of the meson patchset. Plenty changes:

I have various bits of comments on this.

- There are various references to "pch" (pre-compiled headers).  Is
   there more discussion anywhere about this?  I don't know what this
   would entail or whether there are any drawbacks to be aware of.  The
   new *_pch.h files don't have any comments.  Maybe this should be a
   separate patch later.

- About relativize_shared_library_references: We have had several
   patches over the years for working around SIP stuff, and some of
   them did essentially this, but we decided not to go ahead with them.
   We could revisit that, but it should be a separate patch, not mixed
   in with this.

- postgresql-extension.pc: Similarly, this ought to be a separate
   patch.  If we want people to use this, we'll need it in the makefile
   build system anyway.

- -DFRONTEND is used somewhat differently from the makefiles.  For
    example, meson sets -DFRONTEND for pg_controldata, but the
    makefiles don't.  Conversely, the makefiles set -DFRONTEND for
    ecpglib, but meson does not.  This should be checked again to make
    sure it all matches up.

- Option name spelling should be make consistent about underscores
   versus hyphens.  Built-in meson options use underscores, so we
   should make the user-defined ones like that as well (some already
   do).  (wal-blocksize krb-srvnam system-tzdata tap-tests bsd-auth)

- I have found the variable name "cdata" for configuration_data() to
   be less than clear.  I see some GNOME projects to it that way, is
   that where it's from?  systemd uses "conf", maybe that's better.

- In the top-level meson.build, the "renaming" of the Windows system
   name

       host_system = host_machine.system() == 'windows' ? 'win32' : 
host_machine.system()
       build_system = build_machine.system() == 'windows' ? 'win32' : 
build_machine.system()

   seems unnecessary to me.  Why not stick with the provided names?

- The c99_test ought to be not needed if the c_std project option is
   used.  Was there a problem with that?

- Is there a way to split up the top-level meson.build somehow?  Maybe
   just throw some stuff into included files?  This might get out of
   hand at some point.

- The PG_SYSROOT detection gives different results.  On my system,
   configure produces
 
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX12.3.sdk,
   meson produces
 
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk.
   src/template/darwin goes out of its way to get a version-specific
   result, so we need to carry that over somehow.  (The difference does
   result in differences in the built binaries.)


Then, some patches from me:

0001-Change-shared-library-installation-naming-on-macOS.patch

This changes the makefiles to make the shared library file naming on
macOS match what meson produces.  I don't know what the situation is
on other platforms.

0002-meson-Fix-installation-name-of-libpgfeutils.patch

This was presumably an accidental mistake.

0003-meson-Libraries-need-to-be-built-with-DSO_MAJOR_VERS.patch

This is needed to make NLS work for the libraries.

0004-meson-Add-darwin_versions-argument-for-libraries.patch

This is to make the output match what Makefile.shlib produces.

0005-meson-Fix-link-order-of-support-libraries.patch
0006-meson-Make-link-order-of-external-libraries-match-ma.patch
0007-WIP-meson-Make-link-order-of-object-files-match-make.patch

I have analyzed the produced binaries between both build systems to
make sure they match.  If we link the files and libraries in different
orders, that becomes difficult.  So this fixes this up a bit.  0005 is
needed for correctness in general, I think.  0006 is mostly cosmetic.
You probably wanted to make the library order alphabetical in the
meson files, which I'd support, but then we should change the
makefiles to match.  Similarly, 0007, which is clearly a bit messy at
the moment, but we should try to sort that out either in the old or
the new build files.


And finally some comments on your patches:

meson: prereq: Don't add HAVE_LDAP_H HAVE_WINLDAP_H to pg_config.h.

This can go ahead.

meson: prereq: fix warning compat_informix/rnull.pgc with msvc

-   $float f = 3.71;
+   $float f = (float) 3.71;

This could use float literals like

+   $float f = 3.71f;

Commits

  1. meson: Add 'running' test setup, as a replacement for installcheck

  2. meson: Add support for building with precompiled headers

  3. windows: Adjust FD_SETSIZE via commandline define

  4. meson: docs: Add xml{lint,proc} wrapper to collect dependencies

  5. meson: ecpg: Split definition of static and shared libraries

  6. windows: Set UMDF_USING_NTSTATUS globally, include ntstatus.h

  7. ci: Add hint about downloadable logs to README

  8. meson: Set up absolute rpaths to libdir

  9. meson: Include CFLAGS/c_args in summary and pg_config output

  10. windows: remove date from version number in win32ver.rc

  11. meson: Add initial version of meson based build system

  12. ci: windows: set error mode to not include SEM_NOGPFAULTERRORBOX

  13. Refactor PG_TEST_EXTRA logic in autoconf build

  14. Split TESTDIR into TESTLOGDIR and TESTDATADIR

  15. Don't hardcode tmp_check/ as test directory for tap tests

  16. Extend gendef.pl in preparation for meson

  17. Include c.h instead of postgres.h in src/port/*p{read,write}*.c

  18. Remove DLLTOOL, DLLWRAP from configure / Makefile.global.in

  19. Run xmllint validation only once

  20. Bump minimum Perl version to 5.14

  21. Move gramparse.h to src/backend/parser

  22. Run perltidy over Catalog.pm

  23. Parse catalog .dat files as a whole when compiling the backend

  24. Build all Flex files standalone

  25. Move private declarations shared between guc.c and guc-file.l to new header

  26. Preparatory refactoring for compiling guc-file.c standalone

  27. Move darwin sysroot determination into separate file

  28. Fix MSVC warning in compat_informix/rnull.pgc

  29. solaris: Use versioning scripts instead of -Bsymbolic

  30. Change shared library installation naming on macOS

  31. regress: allow to specify directory containing expected files, for ecpg

  32. Don't add HAVE_LDAP_H HAVE_WINLDAP_H to pg_config.h

  33. Refactor dtrace postprocessing make rules

  34. Add output directory option to gen_node_support.pl

  35. Add output directory argument to generate-unicode_norm_table.pl

  36. Add output file argument to generate-errcodes.pl

  37. Add output path arg in generate-lwlocknames.pl

  38. Move snowball_create.sql creation into perl file

  39. ecpg: Output dir, source dir, stamp file argument for preproc/*.pl

  40. psql: Output dir and dependency generation for sql_help

  41. Deal with paths containing \ and spaces in basebackup_to_shell tests

  42. Remove LLVM_CONFIG from Makefile.global.in

  43. Make update-unicode target work in vpath builds

  44. Refactor DLSUFFIX handling

  45. ldap tests: Add paths for openbsd.

  46. ldap tests: Don't run on unsupported operating systems.

  47. Remove check for accept() argument types

  48. Move our p{read,write}v replacements into their own files.

  49. Adjust yywrap macro for non-reentrant scanners for MSVC.

  50. Remove any -arch switches given in ExtUtils::Embed's ldopts from our

  51. Change PL/Perl and Pg interface build to use configured compiler and

  52. Apparently, on some systems, ExtUtils::Embed and MakeMaker are slightly