Re: Meson and Numa: C header not found - Mailing list pgsql-general

From Tom Lane
Subject Re: Meson and Numa: C header not found
Date
Msg-id 526422.1744301420@sss.pgh.pa.us
Whole thread Raw
In response to Re: Meson and Numa: C header not found  ("Daniel Westermann (DWE)" <daniel.westermann@dbi-services.com>)
Responses Re: Meson and Numa: C header not found
List pgsql-general
"Daniel Westermann (DWE)" <daniel.westermann@dbi-services.com> writes:
> Tomas Vondra <tomas@vondra.me> writes:
>> On 4/10/25 08:00, Daniel Westermann (DWE) wrote:
>>> ../postgresql/meson.build:957:12: ERROR: C header 'numa.h' not found

> It is there:

> postgres@pgbox:/home/postgres/ [DEV] ls /usr/include/numa.h
> /usr/include/numa.h

Just to confirm, do you have all of these in /usr/lib64?

$ ls -l /usr/lib64/libnuma*
lrwxrwxrwx. 1 root root    16 Nov  4 19:00 /usr/lib64/libnuma.so -> libnuma.so.1.0.0
lrwxrwxrwx. 1 root root    16 Nov  4 19:00 /usr/lib64/libnuma.so.1 -> libnuma.so.1.0.0
-rwxr-xr-x. 1 root root 62424 Nov  4 19:00 /usr/lib64/libnuma.so.1.0.0

As of HEAD, line 957 of meson.build is

  if not cc.has_header('numa.h', dependencies: libnuma, required: libnumaopt)

I'm not familiar at all with meson, but I wonder exactly what the
dependencies: bit does.  That kind of looks like it might cause the
check to fail if the libnuma library isn't there, which is something
we do need to check, but doing it this way seems like it'd produce
a pretty misleading failure message if numa.h exists but the library
doesn't.

            regards, tom lane



pgsql-general by date:

Previous
From: Adrian Klaver
Date:
Subject: Re: Interesting case of IMMUTABLE significantly hurting performance
Next
From: "Daniel Westermann (DWE)"
Date:
Subject: Re: Meson and Numa: C header not found