3.3.0 build problem on VMS IA64 -- Defect in <stdlib.h> about openssl HOT 4 CLOSED

It is the same sort of thing as #24075, indeed. The fix should be this, please try it:

--- a/Configurations/10-main.conf
+++ b/Configurations/10-main.conf
@@ -2061,7 +2061,8 @@ my %targets = (
                               ? "/WARNINGS=DISABLE=(".join(",",@warnings).")" : (); }),
         cflag_incfirst   => '/FIRST_INCLUDE=',
         lib_defines      =>
-            add("OPENSSL_USE_NODELETE", "_XOPEN_SOURCE", "_XOPEN_SOURCE_EXTENDED",
+            add("OPENSSL_USE_NODELETE",
+                "_XOPEN_SOURCE", "_XOPEN_SOURCE_EXTENDED=1",
                 sub {
                     return vms_info()->{def_zlib}
                         ? "LIBZ=\"\"\"".vms_info()->{def_zlib}."\"\"\"" : ();

I'm submitting a fixup PR in a minute.

from openssl.

It is the same sort of thing as #24075, indeed. [...]

Ah, yes. I was confusing "CC/DEFINE=FRED", which is equivalent to
"#define FRED 1", with "#define FRED", which is not.

[...] The fix should be this, please try it: [...]

Yup. Given the impracticality of ever getting all the bad <stdlib.h>
in the universe fixed, adding the "=1" work-around here looks to me like
the right way. Sorry to waste the time.

Off topic: My "mms test" (IA64, x86_64) hangs after the second test:
01-test_fipsmodule_cnf.t ................ skipped: Test only [...]
Already known or worth reporting?

from openssl.

Sorry to waste the time.

Not at all, I'm glad it was an easy fix

from openssl.

Off topic: My "mms test" (IA64, x86_64) hangs after the second test:
01-test_fipsmodule_cnf.t ................ skipped: Test only [...]
Already known or worth reporting?

This is nothing I've seen. Might be worth an mms test /macro=(V=1) for verbosity to see a bit better what's going on, and reporting if you can't figure it out.

from openssl.