The rest of it:
use Time::Piece;
use CGI::Carp (carpout);
{ local *CGI::Carp::stamp = sub {... };
open( my $log, ">>", "/path/to/error.log") or die $!;
carpout($log);
carp("foo happened");
# close($log)
}
carp("foo again but with module's timestamp");
On Sun, Mar 26, 2017 at 3:04 PM, X Dungeness <[email protected]> wrote:
> Shawn may have a different take but I think the "local" is
> misplaced and goes out of scope when the call's made.
>
> Here's a potential workaround:
>
> out of scope
>
> On Sun, Mar 26, 2017 at 2:02 PM, SSC_perl <[email protected]> wrote:
>>> On Mar 26, 2017, at 1:15 PM, Shawn H Corey <[email protected]> wrote:
>>>
>>> it would mean replacing the subroutine after the module was loaded.
>>
>> Thanks, Shawn, but I can’t get that to work. Reading perldoc Core
>> gives me the impression that I’d need to call the new sub, not the module.
>> If that’s true, I don’t see how it would work in this scenario. Here’s what
>> I tried:
>>
>> BEGIN {
>> use CGI::Carp qw(carpout fatalsToBrowser ); # noTimestamp
>> {
>> no warnings;
>> use Time::Piece;
>> local *CGI::Carp::stamp = sub {
>> my $frame = 0;
>> my ($id, $pack, $file, $dev, $dirs);
>> if (defined($CGI::Carp::PROGNAME)) {
>> $id = $CGI::Carp::PROGNAME;
>> } else {
>> do {
>> $id = $file;
>> ($pack, $file) = caller($frame++);
>> } until !$file;
>> }
>> if (! $CGI::Carp::FULL_PATH) {
>> ($dev, $dirs, $id) = File::Spec->splitpath($id);
>> }
>> return "$id: " if $CGI::Carp::NO_TIMESTAMP;
>> my $time = scalar(localtime);
>> my $dt = Time::Piece->strptime($time, '%a %b %d %H:%M:%S %Y');
>> my $datetime = $dt->strftime('%Y-%m-%d %H:%M');
>> return "[$datetime] $id: ";
>> }
>> }
>> open (my $log, '>>', ‘/path/to/error.log');
>> carpout($log);
>> close ($log);
>> }
>> --
>> To unsubscribe, e-mail: [email protected]
>> For additional commands, e-mail: [email protected]
>> http://learn.perl.org/
>>
>>
--
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
http://learn.perl.org/
