Thread: pgsql: Try to unbreak MSVC builds for pg_waldump
Try to unbreak MSVC builds for pg_waldump remedy an omission in commit 7d8219a444 Branch ------ master Details ------- https://git.postgresql.org/pg/commitdiff/bbec50de16d2bf6e1a2878ff0f7e39dbd86ecff8 Modified Files -------------- src/tools/msvc/Mkvcbuild.pm | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
On Sat, Apr 8, 2023 at 8:22 AM Andrew Dunstan <andrew@dunslane.net> wrote: > Try to unbreak MSVC builds for pg_waldump > > remedy an omission in commit 7d8219a444 Thanks for taking care of this. I was under the impression that issues like this would be caught by CI. We used to have an MSVC CI target, but that's no longer the case -- not since commit e6b6ea02. I'll need to remember that. -- Peter Geoghegan
Peter Geoghegan <pg@bowt.ie> writes: > I was under the impression that issues like this would be caught by > CI. We used to have an MSVC CI target, but that's no longer the case > -- not since commit e6b6ea02. I'll need to remember that. I thought we were planning to drop the MSVC build scripts for v16. Is that no longer happening? If they're going to be around for awhile, maybe that CI job needs to be resurrected. regards, tom lane
On Sat, Apr 8, 2023 at 10:53 AM Tom Lane <tgl@sss.pgh.pa.us> wrote: > I thought we were planning to drop the MSVC build scripts for v16. > Is that no longer happening? If they're going to be around for > awhile, maybe that CI job needs to be resurrected. It seems as if commit e6b6ea02 had very good reasons to replace the MSVC build scripts with a meson build; it's really slow, and wastes a lot of cycles, while testing much less than meson. So I doubt that we'd want to restore exactly what was in place prior to that commit. If we can't just drop the MSVC build scripts, perhaps we could add something akin to the existing CompilerWarnings target. Something that's purely intended to detect these kinds of oversights. Even that seems excessive to me, though. I hope that we can just get rid of the scripts before long. -- Peter Geoghegan
Hi, On 2023-04-08 13:53:34 -0400, Tom Lane wrote: > Peter Geoghegan <pg@bowt.ie> writes: > > I was under the impression that issues like this would be caught by > > CI. We used to have an MSVC CI target, but that's no longer the case > > -- not since commit e6b6ea02. I'll need to remember that. > > I thought we were planning to drop the MSVC build scripts for v16. > Is that no longer happening? If they're going to be around for > awhile, maybe that CI job needs to be resurrected. I was holding off submitting a patch until the BF supported meson - which only happened recently. I was actually planning to start a thread about the topic in the next few days. As far as I can tell, the meson build has more than feature parity with the "homegrown" build - the only two dependencies I haven't been able to test is building with are ossp-uuid (doesn't build) and kerberos (quite complicated, last release for windows is ages ago) Greetings, Andres Freund
On 2023-04-08 Sa 14:13, Andres Freund wrote:
Hi, On 2023-04-08 13:53:34 -0400, Tom Lane wrote:Peter Geoghegan <pg@bowt.ie> writes:I was under the impression that issues like this would be caught by CI. We used to have an MSVC CI target, but that's no longer the case -- not since commit e6b6ea02. I'll need to remember that.I thought we were planning to drop the MSVC build scripts for v16. Is that no longer happening? If they're going to be around for awhile, maybe that CI job needs to be resurrected.I was holding off submitting a patch until the BF supported meson - which only happened recently. I was actually planning to start a thread about the topic in the next few days. As far as I can tell, the meson build has more than feature parity with the "homegrown" build - the only two dependencies I haven't been able to test is building with are ossp-uuid (doesn't build) and kerberos (quite complicated, last release for windows is ages ago)
I am still fighting a few issues, was intending to work more on it this week, so I'm not signing off on buildfarm support for meson quite yet. I think we need to support the old system until meson support is fully baked.
cheers
andrew
-- Andrew Dunstan EDB: https://www.enterprisedb.com