Hi,
On 2023-09-28 22:53:09 -0400, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > On 2023-09-28 19:20:27 -0700, Andres Freund wrote:
> >> Thus the easiest fix looks to be to use this:
> >> - export_fmt = '-exported_symbols_list=@0@'
> >> + export_fmt = '-Wl,-exported_symbols_list,@0@'
> >> I don't have anything older than Ventura to check though.
>
> I don't have meson installed on my surviving Catalina box, but
> I tried the equivalent thing in the Makefile universe:
>
> diff --git a/src/Makefile.shlib b/src/Makefile.shlib
> index f94d59d1c5..f2ed222cc7 100644
> --- a/src/Makefile.shlib
> +++ b/src/Makefile.shlib
> @@ -136,7 +136,7 @@ ifeq ($(PORTNAME), darwin)
> BUILD.exports = $(AWK) '/^[^\#]/ {printf "_%s\n",$$1}' $< >$@
> exports_file = $(SHLIB_EXPORTS:%.txt=%.list)
> ifneq (,$(exports_file))
> - exported_symbols_list = -exported_symbols_list $(exports_file)
> + exported_symbols_list = -Wl,-exported_symbols_list,$(exports_file)
> endif
> endif
>
> That builds and produces correctly-symbol-trimmed shlibs, so I'd
> say it's fine.
Thanks for testing!
I'll go and push that 16/HEAD then.
> (Perhaps we should apply the above to HEAD alongside the meson.build fix, to
> get more test coverage?)
The macos animals BF seem to run Ventura, so I think it'd not really provide
additional coverage that CI and your manual testing already has. So probably
not worth it from that angle?
> > Attached is the above change and a commit to change CI over to Sonoma. Not
> > sure when we should switch, but it seems useful to include for testing
> > purposes at the very least.
>
> No opinion on when to switch CI. Sonoma is surely pretty bleeding edge
> yet.
Yea, it does feel like that...
Greetings,
Andres Freund