| From: | "Jonathan Gonzalez V(dot)" <jonathan(dot)abdiel(at)gmail(dot)com> |
|---|---|
| To: | Bryan Green <dbryan(dot)green(at)gmail(dot)com> |
| Cc: | pgsql-hackers(at)lists(dot)postgresql(dot)org |
| Subject: | Re: [PATCH] Avoid uninitialized-value error in poll_query_until timeout diagnostic |
| Date: | 2026-08-11 11:32:51 |
| Message-ID: | 87a4qs3na4.fsf@gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hello!!
Bryan Green <dbryan(dot)green(at)gmail(dot)com> writes:
> ...
> Cluster.pm runs under "use warnings FATAL => 'all'", so when $query is
> undefined this does not warn, it dies with "Use of uninitialized value
> $query in concatenation". The result is that a timeout in a connection only
> poll fails with an uninitialized-value error instead of printing the timeout
> diagnostic the code is trying to produce, hiding the actual failure.
>
> $ perl -e 'use warnings FATAL => "all"; my $q; my $s = qq(q: $q);'
> Use of uninitialized value $q in concatenation (.) or string at -e line 1.
This was a little bit tricky to reproduce, but I used the following
line:
perl -I src/test/perl -MPostgreSQL::Test::Cluster -MPostgreSQL::Test::Utils -MTest::More -e 'local $PostgreSQL::Test::Utils::timeout_default=0.1; my $n=PostgreSQL::Test::Cluster->new("poll_repro",install_path=>"$ENV{PWD}/build/tmp_install/usr/local/pgsql");ok(!$n->poll_query_until("postgres", undef, ""), "connection-only timeout"); done_testing();'
I wasn't able to reproduce this in a test, so this make sense to happen
only if there's any failure during one of the tests, but this snippet
proved the failure exists.
> The fix uses a fallback string when the query is undefined:
>
> my $msg_query = $query // '(undef - connection attempt only)';
>
> and interpolates $msg_query instead. Test-only, one line.
What about using the most common way in the code instead of using `//`
operator? Something like:
my $msg_query = '(undef - connection attempt only)' unless defined $query;
This is how is done everywhere else and is a bit more clear than the
`//` operator which looks pretty tricky even for Perl
Regards,
--
Jonathan Gonzalez V.
EDB
https://www.enterprisedb.com
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Rui Zhao | 2026-08-11 11:51:16 | Re: [PATCH] Add pg_get_table_ddl() to reconstruct CREATE TABLE statements |
| Previous Message | Jakub Wartak | 2026-08-11 11:31:26 | Re: [Bug Report + Patch] File descriptor leak when io_method=io_uring |