Hi Konstantinos,
On Thu, Apr 28, 2005 at 12:03:48AM +0300, Konstantinos Margaritis wrote:
> Please consider letting the newer version of l-c into testing.
> It fixes a number of bugs (though nothing critical) but mainly it
> introduces extensive logging in /var/log/localization-config.log.
> This at least makes it worthwhile to have, IMHO.
A couple of the changes here look suspect to me:
--- /tmp/files0TieQ/localization-config-0.111/conffiles.d/common/editconfig.pl Fri Feb 4 12:16:56 2005
+++ /tmp/fileonIWiw/localization-config-0.113/conffiles.d/common/editconfig.pl Tue Apr 12 21:26:22 2005
@@ -24,8 +26,8 @@
}
if ($updated) {
- print "Updating $filename\n" if $debug;
- open(FILE, ">$filename.new") || die "Unable to write $filename.new";
+ log_msg("$0: Updating $filename");
+ open(FILE, ">$filename.new") || log_msg("$0: Unable to write $filename.new");
print FILE @lines;
close(FILE);
rename "$filename.new", $filename;
Were the changes from fatal errors to log messages intentional?
--- /tmp/files0TieQ/localization-config-0.111/conffiles.d/gdm.postinst Thu Feb 3 02:49:50 2005
+++ /tmp/fileonIWiw/localization-config-0.113/conffiles.d/gdm.postinst Tue Apr 12 21:15:36 2005
@@ -9,16 +9,12 @@
use warnings;
use Getopt::Long;
-my $debug;
-
-# Get command line options
-GetOptions("debug" => \$debug);
-
# Get lang entry
-my $lang = $ARGV[0] or die "No language given";
+my $lang = $ARGV[0] or log_msg("$0: No language given");
# We use methods in common.pl
require '/usr/lib/localization-config/common/common.pl';
+require '/usr/lib/localization-config/common/log.pl';
# The path where the scripts are kept
my $LIB = '/usr/lib/localization-config';
Has this been tested? Perl doesn't let you call a function prior to the
'require' statement that defines it... And again, the question of errors
vs. logs...
conffiles.d/woody/ispell also seems to have an encoding error in the
spelling of 'bokmål'.
I'm not willing to approve this package as it is right now; if you can
address the above issues (the die vs. log_msg issue is particularly
pervasive, and seems likely to have very bad unintended side effects), it
should be ok to let in.
Thanks,
--
Steve Langasek
postmodern programmer
Attachment:
signature.asc
Description: Digital signature