OVERVIEW

 The Regexp::Common::debian is collection of REs for various strings
 found in the Debian Porject <http://debian.org>.  It's no way intended
 to be a validation tool.

 R::C::d needs perl "v5.10.0" or later because:

 *   4 patterns (29%) make use of decent perls (one doesn't make it
     through with anything older);

 *   It's time to move, "v5.10.0" is six year old and three years
     no-support;

 *   "lenny" is four year old and a year no-support;

INSTALL

 The R::C::d builds with Module::Build.

     $ perl Build.PL
     $ perl Build
     $ perl Build test
     $ perl Build install

 Since we're about strings we need a lots of strings to test against
 (Test::More, unspecified version).  To access them easily (it's all
 about reuse, not implemented yet) I need an apropriate storage.
 Accidentally it's YAML::Tiny (unspecified version).

 "v0.2.1" Reading reports of cpantesters I've got to conclusion that
 YAML::Tiny isn't popular.  ("v0.2.13" Wandering through errors of
 "v0.2.12" I should say it totally is.) And avoiding installing (or
 unability to install (there could be reasons)) build requirements
 isn't that uncommon.  Although I experience a strong belief that some
 YAML reader happens to be installed anyway.  And still I can't find a
 way to specify that *%build_requires* one of but all known to me YAML
 reader.  So here is a dirty trick.  t::TestSuite attempts to require()
 one of known (to me, see below) YAML reader.  Then (upon initial "perl
 Build.PL") t::TestSuite is asked what it has found (if nothing then
 cosmetic "Compilation failed in require" message will be seen).  And
 one what has been found will be added to *%build_requires*; If nothing
 then fair YAML::Tiny will be added.  (I think it's fair because
 YAML::Tiny is pure-Perl, small footprint, and no dependencies.)

 (note) I'm talking about "known to me YAML readers" because I've found
 out that different YAML readers treat source differently.  So I
 attempt to keep t/*.yaml files semantically equal and sintactically
 correct.  Hopefully there're no differences among versions in wild.

 "v0.2.13" (Actually, this feature has been here for years.) Any
 supported YAML reader can be enforced with $ENV{RCD_YAML_ENGINE} magic
 (in spite of any build-time choice):

     RCD_YAML_ENGINE=syck ./Build test

 Readers are assigned by nicks.  Here they are:

 *   "syck" -- YAML::Syck.

 *   "xs" -- YAML::XS.

 *   "tiny" -- YAML::Tiny.

 *   "old" -- YAML.

 *   "data" -- Data::YAML::Reader 'does not support multi-line quoted
     scalars', 'YAML document header not found' -- unsupported, so far.

 "v0.2.2" Various (all, except t/preferences.t and t/sourceslist.t)
 test-units know a magic command '$ENV{RCD_ASK_DEBIAN}'.  Apply it this
 way (enabling all possible external inquiries):

     RCD_ASK_DEBIAN=all ./Build test

 or this (separate keys with any non-word):

     RCD_ASK_DEBIAN=binary,architecture ./Build test

 When applied a test-unit would ask Debian's commands or inspect Debian
 specific files for information the test-unit is interested in.  For
 obvious reasons that magic will fail on non-Debian system; So don't.
 Although if used correctly that could warn of strange ('not known
 before') compatibility problems.  Details:

 architecture of t/architecture.t
     This asks "dpkg-architecture -L" for list of known architectures
     (per Section 11.1 of debian-policy).  That wouldn't find
     architectures dropped (had that happen ever?) but omissions won't
     stay unnoted anymore.

 binary of t/archive.misc.t
     "v0.2.3" Inspects all records in /var/lib/apt/lists/*_Packages,
     extracts *Filename:* entries and matches all of them against
     "m/^$RE{debian}{archive}{binary}$/".  All (if any) failure will be
     reported at the end.

 changelog of t/changelog.t
     "v0.2.8" That will inspect /usr/share/doc/*/changelog.Debian
     files.  To do a complete scan it would take loads of time
     (really).  You should understand, that's not enough to just run
     through changelogs.  It has to be verified that none entry is
     skipped.  The only reliable (for sake of interface, and,
     trivially, presence) source of verification is
     "dpkg-parsechangelog".  And here's the fork-mare.  "perl" forks
     "shell", then "perl", then "perl" again.  There seems to be fork
     of "tail" too.  And that for each entry.  (Not to count "gunzip"
     to decompress the changelog.) "loadavg" climbs over 1.50..2.00
     You've got the picture.  Although that's where choice begins.

     "v0.2.12" That happens that "urgency=high", probably when it's
     that really high, is expressed in blocks (like this:
     "urgency=HIGH").  $RE{d}{changelog} is case-keeping, and then
     dpkg-parsechangelog(1) strikes back and lowercase.  From now on
     such manipulations won't fail a particular entry.

     *changelog*
         "v0.2.9" That defaults to "changelog=5".  See below.

     *changelog=package*
         Only one changelog will be checked.  The one that "eq"s.  The
         package name is picked from directory name.

     *changelog=a*
         Only those changelogs will be checked that "m/^a/".

     *changelog=5*
         "v0.2.9" That will check all changelogs, although it will look
         no more than requested number of entries deep.

              v0.2.9 ~15min ~1.2K changelogs
             v0.2.12 ~30min ~1.3K changelogs ~6.0K subchecks;
             v0.2.13 ~35min ~1.3K changelogs ~6.1K subchecks;

         And that has a perfect sense.  Do you know that "cron" once
         changed it's name to "Cron" (beware leading block)
         (cron_3.0pl1-46)? C'mon, it has happened 12 (tweleve) years
         ago! (And you know what? That default is pretty fair
         (liblog-log4perl-perl_1.16-1).  Probably it should look for
         time passed but entry number.)

     *changelog=-5*
         "v0.2.9" That's different.  It will check as many entries as
         possible (there are changelogs what $RE{d}{changelog} finds
         out more entries than dpkg-parsechangelog ("dpkg_1.2.13" vs
         "dpkg_0.93.79"), but if the offending record is more than that
         far from top then it's reported and otherwise ignored.

              v0.2.9  ~3h ~1.2K changelogs ~45K subchecks
             v0.2.12 ~5¼h ~1.3K changelogs ~63K subchecks
             v0.2.13 ~5½h ~1.3K changelogs ~59K subckecks

     *changelog=_5*
         "v0.2.12" That's a mix of *changelog=5* and *changelog=-5*
         (thanks to irda-utils_0.9.18-8.1 and mime-support_3.49-1).  It
         goes no more than configured entries deep and ignores (and
         reports) any errors.

     *changelog=0*
         (bug) "v0.2.9" That will check all changelogs, check all
         possible entries and BAIL_OUT off first failure.  Shortly --
         don't.  You're warned.  (Although, do it.  t/changelog.t will
         give up pretty soon.)

     To slightly sweeten all that, t/changelog.t attempts to filter
     duplicates.  And it BAIL_OUTs upon first failure.

 package of t/package.t
     "v0.2.10" Nothing special.  Output of "dpkg-query -f
     '${Package}\n' -W" is matched against m/^$RE{debian}{package}$/.
     Probably should parse *_Packagees.

 source of t/archive.source.t
     "v0.2.3" Inspects all records in /var/lib/apt/lists/*_Sources,
     extracts *Files:* entries, then collects trailing filenames.  They
     are matched against "m/^$RE{debian}{archive}{source_1_0}$/",
     "m/^$RE{debian}{archive}{patch_1_0}$/",
     "m/^$RE{debian}{archive}{source_3_0_native}$/",
     "m/^$RE{debian}{archive}{source_3_0_quilt}$/",
     "m/^$RE{debian}{archive}{patch_3_0_quilt}$/", and
     "m/^$RE{debian}{archive}{dsc}$/" (in fact "||").  If none matches
     then it will be reported at the end.
     "m/$RE{debian}{archive}{changes}/" is missing here because there
     is no source of such on no-build system.

 version of t/version.t
     "v0.2.10" Again nothing special.  Output of "dpkg-query -f
     '${Version}\n' -W" is matched against m/^$RE{debian}{version}$/.
     Probably should parse *_Packages too.

 If any test string fails I need to know what and how.  To provide that
 info I've picked Test::Differences (maybe there's other option I'm not
 aware of?) (I'm, Test::Deep).  "v0.60" of T::D closes
 [[email protected]] and [[email protected]].

AVAILABILITY

 Distribution -- <http://search.cpan.org/dist/Regexp-Common-debian/>

BUGS

 Please report here --
 <http://rt.cpan.org/Public/Dist/Display.html?Name=Regexp-Common-debian
 >

COPYRIGHT AND LICENSING

 *   Copyright 2008--2010, 2014 Eric Pozharski <[email protected]>

 *   AS-IS, NO-WARRANTY, HOPE-TO-BE-USEFUL

 *   GNU Lesser General Public License v3
     <http://www.gnu.org/licenses/lgpl.txt>

You are viewing proxied material from ftp.icm.edu.pl. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.