On 11/20/19 3:40 PM, Thomas Munro wrote: > On Fri, Nov 8, 2019 at 3:41 AM Andrew Dunstan > <andrew.dunstan@2ndquadrant.com> wrote: >> On 11/7/19 9:12 AM, Alvaro Herrera wrote: >>>> The patch says: >>>> >>>> + require Win32::API; >>>> + Win32::API->import; >>> Oh, you're right, it does. I wonder why, though: >>> >> On further inspection I think those lines are unnecessary. The remainder >> of the patch doesn't use this at all, AFAICT. > So does that mean we're back on, we can use a patch like Juan Jose's? > I'd love to get rid of these intermittent buildfarm failures, like > this one just now: > > https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=drongo&dt=2019-11-20%2010%3A00%3A10 > > Here you can see: > > could not read "C:/prog/bf/root/HEAD/pgsql.build/src/bin/pg_ctl/tmp_check/t_004_logrotate_primary_data/pgdata/current_logfiles": > Permission denied at > C:/prog/bf/root/HEAD/pgsql.build/src/test/perl/TestLib.pm line 397. > > That line is in the subroutine slurp_file, and says open(my $in, '<', > $filename). Using various clues from this thread, it seems like we > could, on Windows only, add code to TestLib.pm's INIT to rebind > *CORE::GLOBAL::open to a wrapper function that would just do > CreateFile(..., PLEASE_BE_MORE_LIKE_UNIX, ...).
Possibly. I will do some testing on drongo in the next week or so.
I think Perl's open() is a bad candidate for an overload, so I will update the previous patch that only touches slurp_file().
This version address the issues with the required libraries and uses functions that expose less of the Windows API.
