Re: ccache, MSVC, and meson
От | Justin Pryzby |
---|---|
Тема | Re: ccache, MSVC, and meson |
Дата | |
Msg-id | 20220701191841.GH13040@telsasoft.com обсуждение исходный текст |
Ответ на | Re: ccache, MSVC, and meson (Andres Freund <andres@anarazel.de>) |
Ответы |
Re: ccache, MSVC, and meson
|
Список | pgsql-hackers |
On Tue, May 24, 2022 at 01:30:39PM -0700, Andres Freund wrote: > On 2022-05-24 14:52:02 -0500, Justin Pryzby wrote: > > > The spurious message should be fixed, of course. I suspect you dont need a > > > wrapper, you can just set CC='ccache cl.exe' or similar? Afaics it's not > > > meaningful to do 'CC=ccache.exe' alone, because then it'll interpret arguments > > > as ccache options, rather than compiler options. > > > > if meson didn't crash CC=ccache.exe might have worked, because I had set > > CCACHE_COMPILER. > > Did you report the issue? Should be simple enough to fix. > > I seriously doubt it's a good idea to use CCACHE_COMPILER - there's no way > meson (or autoconf or ..) can rely on the results of compiler tests that way, > since CCACHE_COMPILER can change at any time. This updated patch doesn't use CCACHE_COMPILER. cache misses are several times slower (12 minute build time vs 2:30 with ninja, without ccache), so it's possible that can be *slower* if the hit ratio is inadequate. ninja on cirrus builds 3x faster with ccache, but msbuild is only ~2x faster, so I recommend using it only with ninja. There's still warts requires using "plain" with /Z7 /MDd.
Вложения
В списке pgsql-hackers по дате отправления: