Spell checking of POD documentation (was: [SCM] Debian package checker branch, master, updated. 2.5.12-13-g7b947e6)

[Niels Thykier <niels@thykier.net>]

On 2013-04-20 14:20, Jakub Wilk wrote:
> * Niels Thykier <niels@thykier.net>, 2013-04-20, 12:13:
>> -uses I<pure> logic to determine if dependencies are satisifies (i.e. it
>> +uses I<pure> logic to determine if dependencies are satisfies (i.e. it
> 
> s/satisfies/satisfied/
> 

I was thinking of adding a test to find misspelled words in our POD
documentation.  Apparently we^W I need it to keep our documentation
somewhat readable.

Attached is the script I used to find all of the mistakes in our
codebase that was corrected in commit 0af16e4 (using aspell/aspell-en as
the underlying checker).  It needed quite a bit of whitelisting to get
the test to pass - apparently it trips on quite a few argument names for
various subs.
  Do any of you have any useful tips on setting up and managing these
spell checking tests?

~Niels


#!/usr/bin/perl
use strict;
use warnings;
use Test::More;
use Test::Lintian;
eval 'use Test::Spelling';
plan skip_all => "Pod spell checking requires Test::Spelling" if $@;

my @GOOD_WORDS = qw(
  Allbery Barratt Braakman Brockschmidt Geissert Lichtenheld Niels Russ
  Thykier

  lintian Lintian Lintian's dpkg libapt debian Debian DEBIAN

  PTS QA qa uploader uploaders UPLOADER Uploaders changelog changelogs
  desc COND CURVALUE subdirectory subdirectories udeb deb dsc nlist olist
  KEYN BASEDIR METADATA OO TODO dir exitcode nohang substvar substvars
  listref metadata blockingly checksum checksums Nativeness
  src nativeness Indep debfiles diffstat gz env classpath conffiles objdump
  tasksel filename Pre pre hardlink hardlinking hardlinks PROC dirs PROFNAME
  CHECKNAMES COLLMAP ERRHANDLER LPKG unpacker worklist  BASEPATH stderr stdout
  stdin ascii html issuedtags subclasses showdescription printables overridable
  processables msg ORed SIGKILLs SIGTERM wildcard wildcards ar whitelist blacklist
  API amd armhf cpu linux whitelisted blacklisted shaX sha rstrip lstrip parsers
  customisation ALGO CLOC CMD DEBFILE DEST DSCFILE FOH NOCLOSE PARENTDIR PGP
  STARTLINE STR UTF bitmask cp debconf rw proccessable severities AND'ing
  superset YYYY dirname operm username Whitespaces whitespace Whitespace
  udebs multiword recognised eqv testsuite methodx multi multiarch relationA
  relationB Multi natively unordered
);

# md is md5 butchered by aspell
push(@GOOD_WORDS, 'md');

# 'soft'ly which was parsed as soft'ly.
push(@GOOD_WORDS, q{soft'ly});


# This is wrong in general, but it happens to be a package name that
# we use as an example.
push(@GOOD_WORDS, 'alot');

add_stopwords(@GOOD_WORDS);

chdir($ENV{'LINTIAN_ROOT'})
    or die("fatal error: could not chdir to $ENV{LINTIAN_ROOT}: $!");


my @CHECKS = glob('checks/*[!.]*[!c]');
my @DIRS = qw(collection doc/tutorial frontend lib private reporting t/scripts t/helpers);

all_pod_files_spelling_ok(@CHECKS, @DIRS, 't/runtests');