Skip Menu |
 

This queue is for tickets about the Test-Harness CPAN distribution.

Report information
The Basics
Id: 2904
Status: resolved
Priority: 0/
Queue: Test-Harness

People
Owner: Nobody in particular
Requestors: andy [...] petdance.com
Cc:
AdminCc:

Bug Information
Severity: (no value)
Broken in: (no value)
Fixed in: (no value)



Date: Sat, 5 Jul 2003 18:12:28 -0500
From: Andy Lester <andy [...] petdance.com>
To: bug-test-harness [...] rt.cpan.org
Subject: Patch for only-print-once-per-second
Download (untitled) / with headers
text/plain 3.3k
----- Forwarded message from Michael G Schwern <schwern@pobox.com> ----- Date: Sat, 5 Jul 2003 13:34:37 -0700 From: Michael G Schwern <schwern@pobox.com> To: Abigail <abigail@abigail.nl> Cc: Jarkko Hietaniemi <jhi@iki.fi>, perl5-porters@perl.org, petdance@cpan.org Subject: Re: maint @ 19975 [PATCH lib/Test/Harness.pm] On Fri, Jul 04, 2003 at 05:06:29PM +0200, Abigail wrote: Show quoted text
> The simple patch below speeds up testing dramatically in certain > conditions. It eliminates the printing of NNN/NNNN after each > test, saving about 35 bytes of output per test. When testing > a package like Regexp::Common, this saves about 5 Mb of output. > If you are doing a test when logged in remotely, this saves > a lot. I just ran a test, which took 25 minutes without the > patch. With the patch, 'make test' only took 65 seconds. > > Jarkko, is it too late to put this in 5.8.1?
This certainly cannot be put in, crudely striking out a longstanding feature at the last minute. But the problem remains. Even on a local connection, testing Regexp::Common takes a long time due to the printing. Nick's idea of only printing every second is a good one. Printing every N is less good because you can't tell if a test has hung or its just N-1 taking a long time. So here's a patch to make Test::Harness only print the 'ok' output every second, taking care to always print the first 'ok' and always print every 'NOK'. Printing once a second feels a bit too slow, but I'm loathe to get into attempting half-seconds. While I'm at it, it seems a fix for "local $Test::Harness::verbose" didn't get in. See http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2003-02/msg01230.html for the bug. --- lib/Test/Harness.pm 13 Dec 2002 22:11:51 -0000 1.41 +++ lib/Test/Harness.pm 5 Jul 2003 20:13:49 -0000 @@ -13,12 +13,12 @@ use vars qw($VERSION $Verbose $Switches $Have_Devel_Corestack $Curtest $Columns $verbose $switches $ML $Strap - @ISA @EXPORT @EXPORT_OK + @ISA @EXPORT @EXPORT_OK $Last_ML_Print ); # Backwards compatibility for exportable variable names. -*verbose = \$Verbose; -*switches = \$Switches; +*verbose = *Verbose; +*switches = *Switches; $Have_Devel_Corestack = 0; @@ -448,7 +448,7 @@ my $width = _leader_width(@tests); foreach my $tfile (@tests) { - + $Last_ML_Print = 0; # so each test prints at least once my($leader, $ml) = _mk_leader($tfile, $width); local $ML = $ml; print $leader; @@ -705,7 +705,7 @@ my $detail = $totals->{details}[-1]; if( $detail->{ok} ) { - _print_ml("ok $curr/$max"); + _print_ml_less("ok $curr/$max"); if( $detail->{type} eq 'skip' ) { $totals->{skip_reason} = $detail->{reason} @@ -740,6 +740,15 @@ print join '', $ML, @_ if $ML; } + +# For slow connections, we save lots of bandwidth by printing only once +# per second. +sub _print_ml_less { + if( $Last_ML_Print != time ) { + _print_ml(@_); + $Last_ML_Print = time; + } +} sub _bonusmsg { my($tot) = @_; -- I sit on the floor and pick my nose and think of dirty things Of deviant dwarfs who suck their toes and elves who drub their dings. -- Frito Bugger, "Bored Of The Rings" Show quoted text
----- End forwarded message ----- -- Andy Lester => andy@petdance.com => www.petdance.com => AIM:petdance
RT-Send-CC: schwern [...] pobox.com
Rolled out in 2.29


This service is sponsored and maintained by Best Practical Solutions and runs on Perl.org infrastructure.

Please report any issues with rt.cpan.org to rt-cpan-admin@bestpractical.com.