Thread

1. Use HostsFileName everywhere

   Zsolt Parragi <zsolt.parragi@percona.com> — 2026-06-18T06:10:14Z

   Hello!
   
   I noticed that 3 error messages hardcode pg_hosts.conf instead of
   using the user-configured filename.
   
   Attached patch fixes these occurrences.
   

2. Re: Use HostsFileName everywhere

   surya poondla <suryapoondla4@gmail.com> — 2026-07-06T23:43:53Z

   Hi Zsolt,
   Thanks for reporting the bug and the patch. If an operator changes the
   hosts_file GUC, without the patch those three messages don't get updated
   and it makes debugging a lot harder. The fix is clearly right.
   
   A couple of observations:
   1. Small typo in the commit message "harcoded" → "hardcoded".
   2. While reviewing I noticed a related NULL-safety question. This isn't
   caused by your patch, I was just digging to see whether it could turn into
   a crash or segfault after the change.
   guc_parameters.dat declares hosts_file with boot_val => 'NULL', so
   HostsFileName can be NULL if the GUC is never assigned. After the patch,
   the three messages pass HostsFileName straight to errmsg("... \"%s\" ...",
   HostsFileName), and passing NULL to %s is technically undefined behaviour.
   I traced the code path and ran a small reproducer, if HostsFileName is NULL
   and ssl_sni is on, execution flows into load_hosts() ->
   open_auth_file(NULL, ...) -> fopen(NULL, "r"), which on macOS returns NULL
   with errno=EFAULT (glibc behaves the same way).
   The subsequent errmsg(..., "%s", NULL) prints (null) rather than crashing.
   So the observable difference from this patch in the NULL corner case is:
   before the patch we would see: 'could not load "pg_hosts.conf": ...'
   after the patch we would see: 'could not load "(null)": ...'
   Both are misleading but neither is fatal. Also the corner case itself is
   hard to reach with configuration alone as SelectConfigFiles() in guc.c
   fills in configdir/pg_hosts.conf if hosts_file is unset.
   
   Regards,
   Surya Poondla
   

3. Re: Use HostsFileName everywhere

   Michael Paquier <michael@paquier.xyz> — 2026-07-07T06:36:09Z

   On Mon, Jul 06, 2026 at 04:43:53PM -0700, surya poondla wrote:
   > Hi Zsolt,
   > Thanks for reporting the bug and the patch. If an operator changes the
   > hosts_file GUC, without the patch those three messages don't get updated
   > and it makes debugging a lot harder. The fix is clearly right.
   
   Looks like an oversight of 4f433025f666.  This is not critical, still
   a nice life improvement if setting a custom file value for these error
   messages.
   
   > A couple of observations:
   > 1. Small typo in the commit message "harcoded" → "hardcoded".
   > 2. While reviewing I noticed a related NULL-safety question. This isn't
   > caused by your patch, I was just digging to see whether it could turn into
   > a crash or segfault after the change.
   
   SetConfigOption() is called for hosts_file at an early startup stage,
   as of SelectConfigFiles().  init_host_context() and be_tls_init()
   require the GUCs to be loaded, meaning that a NULL value would not be
   an issue because it will be either the default of pg_hosts.conf or the
   custom value set in postgresql.conf.  HBA and ident files work the
   same way: we need them loaded before any authentication would kick in,
   including direct SSL requests.
   
   In short, I think that the patch should be OK.  Purely cosmetic, still
   OK.  Daniel?
   --
   Michael
   

4. Re: Use HostsFileName everywhere

   Daniel Gustafsson <daniel@yesql.se> — 2026-07-07T07:34:13Z

   > On 7 Jul 2026, at 08:36, Michael Paquier <michael@paquier.xyz> wrote:
   
   > In short, I think that the patch should be OK.  Purely cosmetic, still
   > OK.  Daniel?
   
   Agreed, I had this thread shortlisted for when back from vacation so I will
   take care of it shortly.
   
   --
   Daniel Gustafsson