Bug#507241: lintian: --color not working any more
Package: lintian
Version: 2.1.0
Severity: minor
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
In 2.1.0 the friendly --color option doesn't work any more.
If I change the following:
#v+
- --- Output.pm.bak 2008-11-29 10:53:20.0 +0100
+++ Output.pm 2008-11-29 11:33:27.0 +0100
@@ -112,11 +112,11 @@
@{$EXPORT_TAGS{util}},
'string');
+my %default_colors = ( 'E' = 'red' , 'W' = 'yellow' , 'I' = 'cyan' );
+
# for the non-OO interface
our $GLOBAL = new Lintian::Output;
- -my %default_colors = ( 'E' = 'red' , 'W' = 'yellow' , 'I' = 'cyan' );
- -
sub new {
my ($class, %options) = @_;
my $self = { %options };
#v-
I get --colors=always back.
- --colors=auto still fails because $self-stdout gets changed from
\*STDOUT to \*::OUTPUT_PIPE somewhere and therefore _do_color returns false.
Unfortunately I have no idea where/how this happens ...
Cheers,
gregor
- -- System Information:
Debian Release: lenny/sid
APT prefers unstable
APT policy: (990, 'unstable'), (500, 'oldstable'), (500, 'experimental'),
(500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Kernel: Linux 2.6.27.200811231333
Locale: LANG=C, [EMAIL PROTECTED] (charmap=ISO-8859-15)
Shell: /bin/sh linked to /bin/bash
Versions of packages lintian depends on:
ii binutils2.18.1~cvs20080103-7 The GNU assembler, linker and bina
ii diffstat1.46-1 produces graph of changes introduc
ii dpkg-dev1.14.23 Debian package development tools
ii file4.26-1 Determines file type using magic
ii gettext 0.17-6 GNU Internationalization utilities
ii intltool-debian 0.35.0+20060710.1Help i18n of RFC822 compliant conf
ii libipc-run-perl 0.80-2 Perl module for running processes
ii libparse-debianchan 1.1.1-2 parse Debian changelogs and output
ii libtimedate-perl1.1600-9 Time and date functions for Perl
ii liburi-perl 1.35.dfsg.1-1Manipulates and accesses URI strin
ii man-db 2.5.2-3 on-line manual pager
ii perl [libdigest-sha 5.10.0-18Larry Wall's Practical Extraction
lintian recommends no packages.
Versions of packages lintian suggests:
pn binutils-multiarchnone (no description available)
ii libtext-template-perl 1.44-1.2 Text::Template perl module
ii man-db2.5.2-3on-line manual pager
- -- no debconf information
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (GNU/Linux)
iEYEARECAAYFAkkxG94ACgkQOzKYnQDzz+QA3QCgoVcjBtliv3VMA3ebPcxWilMp
pfEAoNUmJbwdBn8YyxgZhmWSJ4HCBTMM
=Brxl
-END PGP SIGNATURE-
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#507241: lintian: --color not working any more
On Sat, 2008-11-29 at 11:39 +0100, gregor herrmann wrote: In 2.1.0 the friendly --color option doesn't work any more. If I change the following: [define %default_colors before creating the $GLOBAL output object] I get --colors=always back. That looks obviously correct to me; thanks. - --colors=auto still fails because $self-stdout gets changed from \*STDOUT to \*::OUTPUT_PIPE somewhere and therefore _do_color returns false. Unfortunately I have no idea where/how this happens ... I'm guessing you're calling lintian with -i? If so, that's because the extra output is generated by piping lintian's output to lintian-info, which now causes the colour to be disabled when using auto. I'm looking at a patch that makes lintian able to generate the output itself, thus allowing the combination of -i and --color=auto to work again (and will post it to the list for review once I've done some more testing). Regards, Adam -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
[SCM] Debian package checker branch, master, updated. 2.1.0-6-gcb815e5
The following commit has been merged in the master branch:
commit cb815e536d7e33646c2cc6b23d14cee0a2d7e35c
Author: Adam D. Barratt [EMAIL PROTECTED]
Date: Sat Nov 29 15:11:11 2008 +
Initialise %default_colors before creating the global output object.
* lib/Lintian/Output.pm:
+ Make sure the default list of colours is initialised before
attempting to use it. Thanks, gregor herrmann. (Closes: #507241)
diff --git a/debian/changelog b/debian/changelog
index b4362e0..be61ffb 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -14,7 +14,10 @@ lintian (2.1.1) unstable; urgency=low
* lib/Lab.pm:
+ [FL] Fix breakage of populate_with_dist() I introduced in 2.1.0.
-
+ * lib/Lintian/Output.pm:
++ [ADB] Make sure the default list of colours is initialised before
+ attempting to use it. Thanks, gregor herrmann. (Closes: #507241)
+
-- Adam D. Barratt [EMAIL PROTECTED] Fri, 28 Nov 2008 21:47:33 +
lintian (2.1.0) unstable; urgency=low
diff --git a/lib/Lintian/Output.pm b/lib/Lintian/Output.pm
index 08a1e87..02beb03 100644
--- a/lib/Lintian/Output.pm
+++ b/lib/Lintian/Output.pm
@@ -113,10 +113,10 @@ our @EXPORT_OK = (@{$EXPORT_TAGS{messages}},
'string');
# for the non-OO interface
-our $GLOBAL = new Lintian::Output;
-
my %default_colors = ( 'E' = 'red' , 'W' = 'yellow' , 'I' = 'cyan' );
+our $GLOBAL = new Lintian::Output;
+
sub new {
my ($class, %options) = @_;
my $self = { %options };
--
Debian package checker
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#507241: lintian: --color not working any more
On Sat, 29 Nov 2008 14:01:42 +, Adam D. Barratt wrote:
[define %default_colors before creating the $GLOBAL output object]
That looks obviously correct to me; thanks.
Cool, thanks.
- --colors=auto still fails because $self-stdout gets changed from
\*STDOUT to \*::OUTPUT_PIPE somewhere and therefore _do_color returns false.
Unfortunately I have no idea where/how this happens ...
I'm guessing you're calling lintian with -i?
Indeed, sorry for not mentioning (and not testing without).
If so, that's because the
extra output is generated by piping lintian's output to lintian-info,
which now causes the colour to be disabled when using auto.
I see.
(And I was looking for OUTPUT_PIPE only in
/usr/share/lintian/lib/Lintian ...)
I'm looking at a patch that makes lintian able to generate the output
itself, thus allowing the combination of -i and --color=auto to work
again (and will post it to the list for review once I've done some more
testing).
Great, much appreciated.
Cheers,
gregor
--
.''`. Home: http://info.comodo.priv.at/{,blog/} / GPG Key ID: 0x00F3CFE4
: :' : Debian GNU/Linux user, admin, developer - http://www.debian.org/
`. `' Member of VIBE!AT, SPI Inc., fellow of FSFE | http://got.to/quote/
`-NP: Flying Pickets: Sign Your Name
signature.asc
Description: Digital signature
RFC: [PATCH] Make lintian --info --color=auto work again
Hi,
The attached patch replaces the pipe from the main lintian script to
lintian-info with an extra flag to Lintian::Output::print_tag()
indicating whether the full output should be displayed. This allows
lintian --info --color=auto to produce colourised output when being
output to a terminal.
Comments welcome.
Cheers,
Adam
diff --git a/debian/changelog b/debian/changelog
index be61ffb..de6964a 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -12,11 +12,25 @@ lintian (2.1.1) unstable; urgency=low
better match executions of the command rather than the use of its name
in pattern matching expressions. (Closes: #499847)
- * lib/Lab.pm:
-+ [FL] Fix breakage of populate_with_dist() I introduced in 2.1.0.
+ * frontend/lintian:
++ [ADB] Don't pipe output to lintian-info when --info is specified.
+ Instead, simply set a variable which can be used later by the output
+ code.
+
* lib/Lintian/Output.pm:
+ [ADB] Make sure the default list of colours is initialised before
attempting to use it. Thanks, gregor herrmann. (Closes: #507241)
++ [ADB] Optionally display the full description of a tag (i.e. when
+ the frontend was called with --info).
+
+ * lib/Lab.pm:
++ [FL] Fix breakage of populate_with_dist() I introduced in 2.1.0.
+ * lib/Read_taginfo.pm:
++ [ADB] Qualify the call to read_dpkg_control in read_tag_info(), so
+ that it can be called from Output::print_tag().
+ * lib/Tags.pm:
++ [ADB] When displaying a tag, indicate whether its full description
+ should be included.
-- Adam D. Barratt [EMAIL PROTECTED] Fri, 28 Nov 2008 21:47:33 +
diff --git a/frontend/lintian b/frontend/lintian
index fe9a823..3b1608a 100755
--- a/frontend/lintian
+++ b/frontend/lintian
@@ -29,7 +29,6 @@ use FileHandle;
# }}}
# {{{ Global Variables
-my $lintian_info_cmd = 'lintian-info'; #Command to run for ?
my $LINTIAN_VERSION = VERSION; #External Version number
my $BANNER = Lintian v$LINTIAN_VERSION; #Version Banner - text form
my $LAB_FORMAT = 8; #Lab format Version Number
@@ -473,10 +472,6 @@ if (defined $LINTIAN_ROOT) {
unless ($LINTIAN_ROOT =~ m,^/,) {
$LINTIAN_ROOT = $cwd/$LINTIAN_ROOT;
}
-# see if it has a frontend directory
-if (-d $LINTIAN_ROOT/frontend) {
-$lintian_info_cmd = $LINTIAN_ROOT/frontend/lintian-info;
-}
} else {
$LINTIAN_ROOT = '/usr/share/lintian';
}
@@ -684,6 +679,7 @@ map { $_-{'script'} = 'lintian'; Tags::add_tag($_) } @l_secs;
$Tags::show_experimental = $display_experimentaltags;
$Tags::show_overrides = $show_overrides;
+$Tags::show_description = $lintian_info;
%Tags::display_level = %display_level;
%Tags::display_source = %display_source;
%Tags::only_issue_tags = map { $_ = 1 } (split(/,/, $check_tags))
@@ -747,29 +743,6 @@ $LINTIAN_LAB = $LAB-{dir};
# }}}
-# {{{ Setup the lintian-info pipe
-
-# pipe output through lintian-info?
-# note: if any E:/W: lines are added above this point, this block needs to
-# be moved up
-if ($lintian_info) {
-open(OUTPUT_PIPE, '|-', $lintian_info_cmd) or fail(cannot open output pipe to $lintian_info_cmd: $!);
-$Lintian::Output::GLOBAL-stdout(\*OUTPUT_PIPE);
-}
-
-# Close the OUTPUT_PIPE in an END block so that we can ensure that
-# lintian-info exits before we do. Otherwise, we may get output from
-# lintian-info after lintian exits, which can confuse the shell output.
-END {
-if ($lintian_info) {
- my $status = $?;
- close OUTPUT_PIPE;
- $Lintian::Output::GLOBAL-stdout(\*STDOUT);
- $? = $status;
-}
-}
-# }}}
-
# {{{ Compile list of files to process
$schedule = new Lintian::Schedule(verbose = $verbose);
diff --git a/lib/Lintian/Output.pm b/lib/Lintian/Output.pm
index 02beb03..aaa904c 100644
--- a/lib/Lintian/Output.pm
+++ b/lib/Lintian/Output.pm
@@ -22,6 +22,9 @@ use strict;
use warnings;
use v5.8.0; # for PerlIO
+use lib $ENV{'LINTIAN_ROOT'}/lib;
+use Read_taginfo;
+use Text_utils;
=head1 NAME
@@ -240,16 +243,19 @@ can only be called as instance methods.
=over 4
-=item Cprint_tag($pkg_info, $tag_info, $extra)
+=item Cprint_tag($pkg_info, $tag_info, $extra, $show_description)
Print a tag. The first two arguments are hash reference with the information
about the package and the tag, $extra is the extra information for the tag
-(if any) as an array reference. Called from Tags::tag().
+(if any) as an array reference. $show_description indicates whether the
+tag's description should also be displayed. Called from Tags::tag().
=cut
+my %already_displayed;
+my %tag_list;
sub print_tag {
-my ($self, $pkg_info, $tag_info, $information) = @_;
+my ($self, $pkg_info, $tag_info, $information, $show_description) = @_;
my $extra = '';
$extra = @$information if @$information;
@@ -269,6 +275,16 @@ sub print_tag {
}
$self-_print('', $code: $pkg_info-{pkg}$type, $tag$extra);
+if ($show_description) {
+ if
Re: RFC: [PATCH] Make lintian --info --color=auto work again
On Sat, Nov 29, 2008 at 03:32:26PM +, Adam D. Barratt wrote: The attached patch replaces the pipe from the main lintian script to lintian-info with an extra flag to Lintian::Output::print_tag() indicating whether the full output should be displayed. This allows lintian --info --color=auto to produce colourised output when being output to a terminal. Comments welcome. While I don't doubt that your patch accomplishes it goal, I wonder whether it would be better to fix this correctly now. By this I mean: 1) The extended description should be available in the $tag_info 2) The list of already issued tags should probably be a object attribute instead of a module attribute. 3) I'm unsure whether the flag show_description shouldn't be an object attribute of Lintian::Output instead of a module attribute of Tags. Gruesse, -- Frank Lichtenheld [EMAIL PROTECTED] www: http://www.djpig.de/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: RFC: [PATCH] Make lintian --info --color=auto work again
On Sat, 2008-11-29 at 17:00 +0100, Frank Lichtenheld wrote: On Sat, Nov 29, 2008 at 03:32:26PM +, Adam D. Barratt wrote: The attached patch replaces the pipe from the main lintian script to lintian-info with an extra flag to Lintian::Output::print_tag() indicating whether the full output should be displayed. This allows lintian --info --color=auto to produce colourised output when being output to a terminal. Comments welcome. While I don't doubt that your patch accomplishes it goal, I wonder whether it would be better to fix this correctly now. By this I mean: That'll teach me to ask for comments. :-) Seriously, thanks for the feedback; constructive criticism is always appreciated. Adam -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#507273: lintian: please skip __init__.py from zero-byte-file-in-doc-directory check
Package: lintian Version: 2.0.0 Severity: minor Hello, __init__.py could be zero-sized, just to mark a directory as a python module, and sometimes such dir are installed in doc/ of the package, so please skip that file from the zero-byte-file-in-doc-directory check. Thanks in advance (and for your huge work), Sandro -- System Information: Debian Release: lenny/sid APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: i386 (i686) Kernel: Linux 2.6.26-1-686 (SMP w/1 CPU core) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/bash Versions of packages lintian depends on: ii binutils2.18.1~cvs20080103-7 The GNU assembler, linker and bina ii diffstat1.46-1 produces graph of changes introduc ii dpkg-dev1.14.22 Debian package development tools ii file4.26-1 Determines file type using magic ii gettext 0.17-4 GNU Internationalization utilities ii intltool-debian 0.35.0+20060710.1Help i18n of RFC822 compliant conf ii libparse-debianchan 1.1.1-2 parse Debian changelogs and output ii libtimedate-perl1.1600-9 Time and date functions for Perl ii liburi-perl 1.35.dfsg.1-1Manipulates and accesses URI strin ii man-db 2.5.2-3 on-line manual pager ii perl [libdigest-sha 5.10.0-17Larry Wall's Practical Extraction lintian recommends no packages. Versions of packages lintian suggests: pn binutils-multiarchnone (no description available) pn libtext-template-perl none (no description available) ii man-db2.5.2-3on-line manual pager -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#507278: lintian: complains about uninitialized values for severity and certainty
On Sat, 2008-11-29 at 18:26 +0100, Neil Williams wrote: lintian has started outputting a lot of perl warnings: Use of uninitialized value $severity in hash element at /usr/share/lintian/lib/Tags.pm line 293. Use of uninitialized value $certainty in hash element at /usr/share/lintian/lib/Tags.pm line 293. (repeated once for each test) In this particular case, I was running lintian after a build of the current version of ipsec-tools in unstable/testing (without changes). Apologies for the amount of information requested, but the warnings suggest that something is quite broken, and I can't reproduce them locally. The above sounds like lintian's not reading the tag files from /usr/share/lintian/checks/*.desc properly. The number of occurrences of Tag:, Severity: and Certainty: in each file should match. Would it be possible to get hold of a copy of the package(s) you're checking? I've just done apt-get source ipsec-tools; pdebuild on an amd64 unstable box and lintian 2.1.0 didn't produce any perl warnings when run against the .changes, .dsc or .deb. Finally, do you have LINTIAN_ROOT set in your environment? Regards, Adam -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
[SCM] Debian package checker branch, master, updated. 2.1.0-8-g66f927e
The following commit has been merged in the master branch:
commit 66f927e62c0a0f62309c83549d96631c7d59fe3c
Author: Adam D. Barratt [EMAIL PROTECTED]
Date: Sat Nov 29 18:26:49 2008 +
Don't output a start package message for a .changes files.
Don't output a Processing binary package foo.changes (version ) message.
Thanks Sandro Tosi for pointing the issue out.
.changes files aren't really packages, but we pretend they're binary
packages so that we can issue tags against them.
We already output a message indicating we're processing a .changes
file so the start package message is spurious and somewhat confusing.
Maybe we should have a changes type rather than faking them as binary
packages, but in the meantime this gets us back to the pre-::Output
output.
diff --git a/debian/changelog b/debian/changelog
index 920c05c..fea113e 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -21,6 +21,9 @@ lintian (2.1.1) unstable; urgency=low
* lib/Lab.pm:
+ [FL] Fix breakage of populate_with_dist() I introduced in 2.1.0.
+ * lib/Tags.pm:
++ [ADB] Don't output a Processing binary package foo.changes message.
+ Thanks Sandro Tosi for pointing the issue out.
-- Adam D. Barratt [EMAIL PROTECTED] Fri, 28 Nov 2008 21:47:33 +
diff --git a/lib/Tags.pm b/lib/Tags.pm
index 310a8a1..51f30c3 100644
--- a/lib/Tags.pm
+++ b/lib/Tags.pm
@@ -131,7 +131,9 @@ sub select_pkg {
$Lintian::Output::GLOBAL-print_end_pkg($info{$current});
}
$current = $file;
-$Lintian::Output::GLOBAL-print_start_pkg($info{$current});
+if ($file !~ /.changes$/) {
+ $Lintian::Output::GLOBAL-print_start_pkg($info{$current});
+}
return 1;
}
--
Debian package checker
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Processed (with 1 errors): Re: Bug#507278: lintian: complains about uninitialized values for severity and certainty
Processing commands for [EMAIL PROTECTED]: clone -1 Unknown command or malformed arguments to command. reassign -1 emdebian-tools Bug number -1 not found. (Is it archived?) retitle -1 emdebian.desc lintian checks should include certainty values Bug number -1 not found. (Is it archived?) severity 507278 wishlist Bug#507278: lintian: complains about uninitialized values for severity and certainty Severity set to `wishlist' from `minor' retitle 507278 please check for out-of-sync check files in /usr/share/lintian/checks Bug#507278: lintian: complains about uninitialized values for severity and certainty Changed Bug title to `please check for out-of-sync check files in /usr/share/lintian/checks' from `lintian: complains about uninitialized values for severity and certainty'. thanks Stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#507278: lintian: complains about uninitialized values for severity and certainty
clone -1 reassign -1 emdebian-tools retitle -1 emdebian.desc lintian checks should include certainty values severity 507278 wishlist retitle 507278 please check for out-of-sync check files in /usr/share/lintian/checks thanks On Sat, 29 Nov 2008 18:20:16 + Adam D. Barratt [EMAIL PROTECTED] wrote: On Sat, 2008-11-29 at 18:26 +0100, Neil Williams wrote: lintian has started outputting a lot of perl warnings: Use of uninitialized value $severity in hash element at /usr/share/lintian/lib/Tags.pm line 293. Use of uninitialized value $certainty in hash element at /usr/share/lintian/lib/Tags.pm line 293. (repeated once for each test) In this particular case, I was running lintian after a build of the current version of ipsec-tools in unstable/testing (without changes). Apologies for the amount of information requested, but the warnings suggest that something is quite broken, and I can't reproduce them locally. The above sounds like lintian's not reading the tag files from /usr/share/lintian/checks/*.desc properly. The number of occurrences of Tag:, Severity: and Certainty: in each file should match. Ah. In that case, the problem is /usr/share/lintian/checks/emdebian.desc I didn't know that the desc file had to be updated. Would it be possible to get hold of a copy of the package(s) you're checking? I've just done apt-get source ipsec-tools; pdebuild on an amd64 unstable box and lintian 2.1.0 didn't produce any perl warnings when run against the .changes, .dsc or .deb. It would be good if lintian wasn't quite so noisy about such errors, hence I've cloned this as wishlist for lintian as well as reassigning to emdebian-tools. -- Neil Williams = http://www.data-freedom.org/ http://www.linux.codehelp.co.uk/ http://e-mail.is-not-s.ms/ pgp0tkI5mLuPN.pgp Description: PGP signature
Bug#507273: lintian: please skip __init__.py from zero-byte-file-in-doc-directory check
On Sat, Nov 29, 2008 at 05:40:05PM +0100, Sandro Tosi wrote: __init__.py could be zero-sized, just to mark a directory as a python module, and sometimes such dir are installed in doc/ of the package, so please skip that file from the zero-byte-file-in-doc-directory check. We already skip zero-byte-file-in-doc-directory if the file is under examples/. But yeah, looking at the tag page, __init__.py and html/*.map might be good candidates for overriding it even outside of that. Gruesse, -- Frank Lichtenheld [EMAIL PROTECTED] www: http://www.djpig.de/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Processed: cloning 507278, reassign -1 to emdebian-tools ...
Processing commands for [EMAIL PROTECTED]: clone 507278 -1 Bug#507278: please check for out-of-sync check files in /usr/share/lintian/checks Bug 507278 cloned as bug 507285. reassign -1 emdebian-tools Bug#507285: please check for out-of-sync check files in /usr/share/lintian/checks Bug reassigned from package `lintian' to `emdebian-tools'. retitle -1 emdebian.desc lintian checks should include certainty values Bug#507285: please check for out-of-sync check files in /usr/share/lintian/checks Changed Bug title to `emdebian.desc lintian checks should include certainty values' from `please check for out-of-sync check files in /usr/share/lintian/checks'. End of message, stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#507278: more uninitialized values with lintian --color=auto
$ alias | grep lintian alias lintian='lintian --color=auto' $ lintian ../ipsec-tools_0.7.1-1.3_amd64.changes Use of uninitialized value $_ in split at /usr/share/perl/5.10/Term/ANSIColor.pm line 121. Use of uninitialized value in concatenation (.) or string at /usr/share/perl/5.10/Term/ANSIColor.pm line 187. W: ipsec-tools source: debhelper-but-no-misc-depends ipsec-tools Use of uninitialized value $_ in split at /usr/share/perl/5.10/Term/ANSIColor.pm line 121. Use of uninitialized value in concatenation (.) or string at /usr/share/perl/5.10/Term/ANSIColor.pm line 187. W: ipsec-tools source: debhelper-but-no-misc-depends racoon Use of uninitialized value $_ in split at /usr/share/perl/5.10/Term/ANSIColor.pm line 121. Use of uninitialized value in concatenation (.) or string at /usr/share/perl/5.10/Term/ANSIColor.pm line 187. W: ipsec-tools source: out-of-date-standards-version 3.7.3 (current is 3.8.0) Use of uninitialized value $_ in split at /usr/share/perl/5.10/Term/ANSIColor.pm line 121. Use of uninitialized value in concatenation (.) or string at /usr/share/perl/5.10/Term/ANSIColor.pm line 187. W: racoon: non-dev-pkg-with-shlib-symlink usr/lib/libracoon.so.0.0.0 usr/lib/libracoon.so Use of uninitialized value $_ in split at /usr/share/perl/5.10/Term/ANSIColor.pm line 121. Use of uninitialized value in concatenation (.) or string at /usr/share/perl/5.10/Term/ANSIColor.pm line 187. W: racoon: package-name-doesnt-match-sonames libracoon0 Use of uninitialized value $_ in split at /usr/share/perl/5.10/Term/ANSIColor.pm line 121. Use of uninitialized value in concatenation (.) or string at /usr/share/perl/5.10/Term/ANSIColor.pm line 187. W: ipsec-tools: package-name-doesnt-match-sonames libipsec0 The errors go away if I avoid the alias using /usr/bin/lintian. -- Neil Williams = http://www.data-freedom.org/ http://www.linux.codehelp.co.uk/ http://e-mail.is-not-s.ms/ pgpbH699ZTSI3.pgp Description: PGP signature
Bug#507278: more uninitialized values with lintian --color=auto
On Sat, 2008-11-29 at 20:22 +0100, Neil Williams wrote: $ alias | grep lintian alias lintian='lintian --color=auto' $ lintian ../ipsec-tools_0.7.1-1.3_amd64.changes Use of uninitialized value $_ in split at /usr/share/perl/5.10/Term/ANSIColor.pm line 121. Use of uninitialized value in concatenation (.) or string at /usr/share/perl/5.10/Term/ANSIColor.pm line 187. W: ipsec-tools source: debhelper-but-no-misc-depends ipsec-tools These are a side-effect of the bug reported in #507241, which is already fixed in the repository. Regards, Adam -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#505857: lintian: false positive debian-watch-file-should-mangle-version
Raphael Geissert [EMAIL PROTECTED] writes: 3.- Long term solution: Turn some portions of uscan into a perl module[1] and use it to extract the dversionmangle regex. If there are really so many cases out there where this bug occurs and people is desperate to have it fixed instead of overriding the warning until a proper fix is available, am willing to work on 2. Otherwise, I will be working on the pm part of 3 some time soon, which just like the checkbashisms as pm stuff will need to be integrated in lintian afterwards. 3 is definitely the right solution in my opinion. -- Russ Allbery ([EMAIL PROTECTED]) http://www.eyrie.org/~eagle/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#507241: lintian: --color not working any more
Adam D. Barratt [EMAIL PROTECTED] writes: I'm guessing you're calling lintian with -i? If so, that's because the extra output is generated by piping lintian's output to lintian-info, which now causes the colour to be disabled when using auto. I'm looking at a patch that makes lintian able to generate the output itself, thus allowing the combination of -i and --color=auto to work again (and will post it to the list for review once I've done some more testing). A short-term fix would be to have Lintian check whether stdout is a tty before spawning lintian-info, but your fix is the better one, so if you're close to having it working already, we should just go with that. -- Russ Allbery ([EMAIL PROTECTED]) http://www.eyrie.org/~eagle/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#498875: lintian: please detect 2.x licences and warn
tag 498875 - moreinfo thanks Attaching updated patch with a more accurate description of the situation and new, extra, references. As for excluding PHP from the check: since the check is binary-based it would be a PITA to maintain the list of binary packages, and using a regex to exclude packages won't help because it may cause false negatives. And of course, only old versions of PHP itself will trigger the warning (not even etch's version trigger it), so it should be just fine not to exclude anything. Cheers, -- Raphael Geissert - Debian Maintainer www.debian.org - get.debian.net php_licences.mbox Description: application/mbox signature.asc Description: This is a digitally signed message part.
Processed: Re: Bug#498875: lintian: please detect 2.x licences and warn
Processing commands for [EMAIL PROTECTED]: tag 498875 - moreinfo Bug#498875: lintian: please detect PHP 2.x licences and warn Tags were: moreinfo patch Tags removed: moreinfo thanks Stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: new set of changes
Raphael Geissert wrote: Frank Lichtenheld wrote: On Wed, Jul 30, 2008 at 08:45:08PM -0500, Raphael Geissert wrote: Hi all, Attached is a patch making the following changes: * correctly parse line continuations * avoid trying to perform any extra check on version 1 watch files * know about versionmangle when checking if there is any mangling because of $repack * add line information to some of the tags * attempt to detect when watch file tries to access a sf.net mirror directly (projectfoo.sf.net/some_file-(.+) and similar of course don't cause false positives) * debian-watch-file-specifies-wrong-upstream-version * debian-watch-file-specifies-old-upstream-version I would prefer if you could split that up in separate patches for the separate issues. Also if you could submit it in git patch format or offer a public git repository to pull from that would make applying them easier. Is a single mbox fine? Attached is one, but doesn't contain the testsuite entries. [...] ping Cheers, -- Raphael Geissert - Debian Maintainer www.debian.org - get.debian.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: RFC: [PATCH] Make lintian --info --color=auto work again
On Sat, 2008-11-29 at 17:00 +0100, Frank Lichtenheld wrote:
While I don't doubt that your patch accomplishes it goal, I wonder
whether it would be better to fix this correctly now. By this
I mean:
1) The extended description should be available in the $tag_info
2) The list of already issued tags should probably be a object
attribute instead of a module attribute.
3) I'm unsure whether the flag show_description shouldn't be an
object attribute of Lintian::Output instead of a module
attribute of Tags.
Thanks for the feedback. I've reimplemented the changes bearing your
suggestions in mind, and attached an updated patch.
Descriptions are now stored along with the rest of the tag data read by
the frontend. I considered doing it directly in Tags.pm but it seemed
cleaner not to have add_tag() modify the tag data directly. The
formatting work itself is all abstracted in Read_taginfo.pm so it can be
reused elsewhere.
The Output module has gained two attributes indicating which tags have
been issued by the module, and whether descriptions should be displayed.
Regards,
Adam
diff --git a/debian/changelog b/debian/changelog
index fea113e..1c1ac9d 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -12,15 +12,26 @@ lintian (2.1.1) unstable; urgency=low
better match executions of the command rather than the use of its name
in pattern matching expressions. (Closes: #499847)
+ * frontend/lintian:
++ [ADB] Don't call lintian-info when --info is specified. Set a new
+ flag on the Lintian::Output module instead.
++ [ADB] Add the description to the list of information stored about each
+ tag.
+
* lib/Lintian/Output.pm:
+ [ADB] Make sure the default list of colours is initialised before
attempting to use it. Thanks, gregor herrmann. (Closes: #507241)
++ [ADB] Maintain a list of tags which have been issued.
++ [ADB] Optionally include the description of a tag when displaying it.
* lib/Lintian/Output/XML.pm:
+ [ADB] Add a missing angle bracket to the closing tag tag, so that
the XML is well-formed.
* lib/Lab.pm:
+ [FL] Fix breakage of populate_with_dist() I introduced in 2.1.0.
+ * lib/Read_taginfo.pm:
++ [ADB] Move the work of formatting a tag's description to a separate
+ function so that it can be called from multiple locations.
* lib/Tags.pm:
+ [ADB] Don't output a Processing binary package foo.changes message.
Thanks Sandro Tosi for pointing the issue out.
diff --git a/frontend/lintian b/frontend/lintian
index fe9a823..66beaba 100755
--- a/frontend/lintian
+++ b/frontend/lintian
@@ -29,7 +29,6 @@ use FileHandle;
# }}}
# {{{ Global Variables
-my $lintian_info_cmd = 'lintian-info'; #Command to run for ?
my $LINTIAN_VERSION = VERSION; #External Version number
my $BANNER = Lintian v$LINTIAN_VERSION; #Version Banner - text form
my $LAB_FORMAT = 8; #Lab format Version Number
@@ -473,10 +472,6 @@ if (defined $LINTIAN_ROOT) {
unless ($LINTIAN_ROOT =~ m,^/,) {
$LINTIAN_ROOT = $cwd/$LINTIAN_ROOT;
}
-# see if it has a frontend directory
-if (-d $LINTIAN_ROOT/frontend) {
-$lintian_info_cmd = $LINTIAN_ROOT/frontend/lintian-info;
-}
} else {
$LINTIAN_ROOT = '/usr/share/lintian';
}
@@ -626,6 +621,7 @@ require Lab;
require Util;
require Read_pkglists;
+require Read_taginfo;
import Util;
@@ -663,6 +659,7 @@ $Lintian::Output::GLOBAL-verbose($verbose);
$Lintian::Output::GLOBAL-debug($debug);
$Lintian::Output::GLOBAL-quiet($quiet);
$Lintian::Output::GLOBAL-color($color);
+$Lintian::Output::GLOBAL-showdescription($lintian_info);
# Print Debug banner, now that we're finished determining
# the values and have Lintian::Output available
@@ -680,7 +677,8 @@ debug_msg(1,
my @l_secs = read_dpkg_control($LINTIAN_ROOT/checks/lintian.desc);
shift(@l_secs);
-map { $_-{'script'} = 'lintian'; Tags::add_tag($_) } @l_secs;
+map { $_-{'script'} = 'lintian'; $_-{'description'} =
+ format_tag_description($_, 3); Tags::add_tag($_) } @l_secs;
$Tags::show_experimental = $display_experimentaltags;
$Tags::show_overrides = $show_overrides;
@@ -747,29 +745,6 @@ $LINTIAN_LAB = $LAB-{dir};
# }}}
-# {{{ Setup the lintian-info pipe
-
-# pipe output through lintian-info?
-# note: if any E:/W: lines are added above this point, this block needs to
-# be moved up
-if ($lintian_info) {
-open(OUTPUT_PIPE, '|-', $lintian_info_cmd) or fail(cannot open output pipe to $lintian_info_cmd: $!);
-$Lintian::Output::GLOBAL-stdout(\*OUTPUT_PIPE);
-}
-
-# Close the OUTPUT_PIPE in an END block so that we can ensure that
-# lintian-info exits before we do. Otherwise, we may get output from
-# lintian-info after lintian exits, which can confuse the shell output.
-END {
-if ($lintian_info) {
- my $status = $?;
- close OUTPUT_PIPE;
- $Lintian::Output::GLOBAL-stdout(\*STDOUT);
- $? = $status;
-}
-}
-# }}}
-
# {{{ Compile list of files to process
$schedule
Bug#507331: lintian fails at being pendantic
Package: lintian Version: 2.0.0 Severity: minor If lintian is going to be pendantic, please be correct: spelling-error-in-changelog dont don\'t i didn\'t know thats how you were supposed to spell it, but we\'d all be smarter if that\'s how it is done, now wouldn\'t we? micah -- System Information: Debian Release: lenny/sid APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: i386 (i686) Kernel: Linux 2.6.26-1-vserver-686 (SMP w/2 CPU cores) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/bash Versions of packages lintian depends on: ii binutils2.18.1~cvs20080103-7 The GNU assembler, linker and bina ii diffstat1.46-1 produces graph of changes introduc ii dpkg-dev1.14.23 Debian package development tools ii file4.26-1 Determines file type using magic ii gettext 0.17-4 GNU Internationalization utilities ii intltool-debian 0.35.0+20060710.1Help i18n of RFC822 compliant conf ii libparse-debianchan 1.1.1-2 parse Debian changelogs and output ii libtimedate-perl1.1600-9 Time and date functions for Perl ii liburi-perl 1.35.dfsg.1-1Manipulates and accesses URI strin ii man-db 2.5.2-3 on-line manual pager ii perl [libdigest-sha 5.10.0-18Larry Wall's Practical Extraction lintian recommends no packages. Versions of packages lintian suggests: pn binutils-multiarchnone (no description available) ii libtext-template-perl 1.44-1.2 Text::Template perl module ii man-db2.5.2-3on-line manual pager -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
[SCM] Debian package checker branch, master, updated. 2.1.0-9-ga869088
The following commit has been merged in the master branch: commit a869088dd1e5fbfcd94bd1efb3bbfee4f080886e Author: Russ Allbery [EMAIL PROTECTED] Date: Sat Nov 29 20:51:49 2008 -0800 Don't needlessly escape apostrophes in spelling corrections * lib/Spelling.pm: + [RA] Don't add unnecessary backslashes in the list of spelling corrections. They're copied into the output. Thanks, Micah Anderson. (Closes: #507331) diff --git a/debian/changelog b/debian/changelog index fea113e..69df061 100644 --- a/debian/changelog +++ b/debian/changelog @@ -21,6 +21,10 @@ lintian (2.1.1) unstable; urgency=low * lib/Lab.pm: + [FL] Fix breakage of populate_with_dist() I introduced in 2.1.0. + * lib/Spelling.pm: ++ [RA] Don't add unnecessary backslashes in the list of spelling + corrections. They're copied into the output. Thanks, Micah + Anderson. (Closes: #507331) * lib/Tags.pm: + [ADB] Don't output a Processing binary package foo.changes message. Thanks Sandro Tosi for pointing the issue out. diff --git a/lib/Spelling.pm b/lib/Spelling.pm index 326ea96..2fc5776 100644 --- a/lib/Spelling.pm +++ b/lib/Spelling.pm @@ -109,7 +109,7 @@ our %CORRECTIONS = qw( cryptocraphic cryptographic deamon daemon debain Debian - debians Debian\'s + debians Debian's decompres decompress definate definite definately definitely @@ -127,7 +127,7 @@ our %CORRECTIONS = qw( documentaion documentation docuentation documentation documantation documentation - dont don\'t + dont don't easilly easily ecspecially especially edditable editable @@ -214,7 +214,7 @@ our %CORRECTIONS = qw( necesary necessary nescessary necessary noticable noticeable - o\'caml OCaml + o'caml OCaml optionnal optional orientatied orientated orientied oriented diff --git a/testset/binary/debian/control b/testset/binary/debian/control index 7aecae8..03cc317 100644 --- a/testset/binary/debian/control +++ b/testset/binary/debian/control @@ -19,7 +19,7 @@ Description: test handling of binary files is checked for picky spelling errors. . This package list [ subversion | gconf ] should not be flagged as a spelling - mistake. + mistake. The spelling correction for dont should be correct. . Homepage: http://lintian.debian.org/ diff --git a/testset/tags.binary b/testset/tags.binary index 33b28ff..b525026 100644 --- a/testset/tags.binary +++ b/testset/tags.binary @@ -116,6 +116,7 @@ W: binary: old-fsf-address-in-copyright-file W: binary: package-contains-hardlink usr/bar2 - usr/share/baz W: binary: package-contains-upstream-install-documentation usr/share/doc/binary/INSTALL W: binary: spelling-error-in-description debian Debian +W: binary: spelling-error-in-description dont don't W: binary: spelling-error-in-doc-base-abstract-field binary:10 speling spelling W: binary: spelling-error-in-doc-base-title-field binary:2 debian Debian W: binary: spelling-error-in-news-debian speling spelling -- Debian package checker -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#507331: lintian fails at being pendantic
Micah Anderson [EMAIL PROTECTED] writes: Package: lintian Version: 2.0.0 Severity: minor If lintian is going to be pendantic, please be correct: spelling-error-in-changelog dont don\'t i didn\'t know thats how you were supposed to spell it, but we\'d all be smarter if that\'s how it is done, now wouldn\'t we? Heh. :) Thanks, will be fixed in the next release. -- Russ Allbery ([EMAIL PROTECTED]) http://www.eyrie.org/~eagle/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#507331: setting package to lintian, tagging 507241, tagging 507331
# Automatically generated email from bts, devscripts version 2.10.35 # via tagpending # # lintian (2.1.1) unstable; urgency=low # # * lib/Lintian/Output.pm: #+ [ADB] Make sure the default list of colours is initialised before # attempting to use it. Thanks, gregor herrmann. (Closes: #507241) # * lib/Spelling.pm: #+ [RA] Don't add unnecessary backslashes in the list of spelling # corrections. They're copied into the output. Thanks, Micah # Anderson. (Closes: #507331) # package lintian tags 507241 + pending tags 507331 + pending -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Processed: setting package to lintian, tagging 507241, tagging 507331
Processing commands for [EMAIL PROTECTED]: # Automatically generated email from bts, devscripts version 2.10.35 # via tagpending # # lintian (2.1.1) unstable; urgency=low # # * lib/Lintian/Output.pm: #+ [ADB] Make sure the default list of colours is initialised before # attempting to use it. Thanks, gregor herrmann. (Closes: #507241) # * lib/Spelling.pm: #+ [RA] Don't add unnecessary backslashes in the list of spelling # corrections. They're copied into the output. Thanks, Micah # Anderson. (Closes: #507331) # package lintian Ignoring bugs not assigned to: lintian tags 507241 + pending Bug#507241: lintian: --color not working any more There were no tags set. Tags added: pending tags 507331 + pending Bug#507331: lintian fails at being pendantic There were no tags set. Tags added: pending End of message, stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
