| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | pgsql-hackers(at)postgreSQL(dot)org |
| Subject: | Still more #ifdef fun: struct sockaddr_storage |
| Date: | 2006-07-15 16:22:03 |
| Message-ID: | 16940.1152980523@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
pgcheckdefines reports
src/port/getaddrinfo.c references HAVE_STRUCT_SOCKADDR_STORAGE_SS_LEN, defined in src/include/libpq/pqcomm.h
I think this is an actual bug: if we have a platform that calls its
struct sockaddr_storage member __ss_len, and we are using our own
getaddrinfo code, we will fail to initialize the length member of
the returned ai_addr struct.
It would be safe to include src/include/libpq/pqcomm.h into
getaddrinfo.c, but it seems a tad ugly. Perhaps the machinations
in pqcomm.h to normalize the sockaddr_storage names ought to be
moved into a separate header file? If so, where to put it?
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Florian G. Pflug | 2006-07-15 16:22:12 | RFC: A dead simple implementation of pg_rotate_wal |
| Previous Message | Sergey E. Koposov | 2006-07-15 16:12:58 | PG_RETURN_INT64 vs PointerGetDatum & ANALYZE |