Version 1.5000: Thu Jan 11 2007
- The Cache and Grep implementations now detect when changes have occurred to
the file being parsed, and automatically invalidate their cache values and
revert to the Perl implementation. NOTE: This works well for appending to
the mailbox, but undefined behavior occurs if modifications are made to the
mailbox at locations earlier than the current position. (Thanks to Armin
Obersteiner <armin@xos.net> for the feature suggestion.)
- Changed the reset test so that it doesn't create output with inconsistent
line endings in the case of a dos mailbox.
- Fixed a bug where occasionally the reading of the next email will go into an
infinite loop when reading from standard input. (Thanks to
<hpeberhard@users.sourceforge.net> for the bug report and sample input.)
- read_next_email now returns undef on end of file. (Thanks to Lucas Nussbaum
<lucas@lucas-nussbaum.net> for first reporting the bug. Thanks to Stephen
Gran <sgran@debian.org>, Frank Lichtenheld <djpig@debian.org>, Steinar
H. Gunderson <sgunderson@bigfoot.com>, Christian Hammers <ch@debian.org>,
gregor herrmann <gregor+debian@comodo.priv.at>, and Joey Hess
<joeyh@debian.org> for their efforts to fix or work around the bug. Finally,
thanks to Tassilo von Parseval <Tassilo.von.Parseval@rwth-aachen.de> for
(mis)using the module in an intuitive way that prompted the solution.)
- Simplified code by removing end_of_file attribute of parser objects.
Based on patch provided by Martin Wilke via PR 34390.
Changelog:
- Fixed a bug where emails with a line near the end that start with "From "
would cause the Grep implementation to go into an infinite loop. (Thanks to
Volker Kuhlmann <VolkerKuhlmann@gmx.de> for the bug report.)
- Fixed some minor coding style issues.
Version 1.4004: Tue Jul 11 2006
- Fixed a bug where, when emails are incomplete, the Perl parser would cache
incorrect information, causing the Cache implementation to go into an
infinite loop when it tried to use the invalid information.
- Improved the behavior for multi-part emails that lack a valid ending
boundaries. Instead of treating the rest of the mailbox as part of the
email, the Perl and Grep parsers now find the end of the email using (1) a
Content-Length header if it is present, or (2) the next valid "^From " line
after the email's header. (Many thanks to Volker Kuhlmann
<VolkerKuhlmann@gmx.de>, Eduard Bloch <edi@gmx.de>, and Joey Hess
<joeyh@debian.org> for their efforts to track down the cause of the bug.
Special thanks to Volker for suggesting the right behavior. :)
- Simplified the code some
Version 1.4003: Sun May 21 2006
- Fixed a bug where multi-part emails having boundaries containing characters
like " " and "+" would cause the remainder of the mailbox to be treated as
part of the email with the boundary. (Thanks to Volker Kuhlmann
<VolkerKuhlmann@gmx.de> for first reporting the bug, and thanks to Joey Hess
<joeyh@debian.org> for the bugfix.)
- Fixed a previously unrevealed fault in the mailbox for the separators test.
(The last message's separator wasn't used correctly.)
Version 1.4002: Thu Feb 9 2006
- Dropped tzip support. The program seems poorly supported and buggy. (I can't
get it to run right on Mac, for example.)
- An invalid cache is detected and overwritten. This can occur if one changes
architectures, or if the cache is otherwise corrupted. (Thanks to Volker
Kuhlmann <VolkerKuhlmann@gmx.de> for the feature suggestion.)
- Fixed a major memory consumption bug in the Perl implementation, where the
read buffer would grow exponentially. This caused the module to use memory
proportional to the size of the mailbox, rather than the size of the largest
email. (Thanks to David Cantrell <david@cantrell.org.uk> for the bug
report.)
Version 1.4001: Tue Aug 2 2005
- Fixed a bug where emails involving time zones of length more than 3
characters (e.g. "WETDST") would not be processed correctly. (Thanks to
Hans F. Nordhaug <Hans.F.Nordhaug@hiMolde.no> for the bug report.)
- Fixed some undefined value warnings for some test cases.
- Fixed a bug where mailboxes having emails with rfc822 attachments would not
be parsed correctly
Version 1.4000: Thu Jul 7 2005
- Fixed a long-standing bug in the parsing of mail messages containing mail
attachments. (Thanks to Brian May <bam@debian.org> for the bug report.)
- Dropped X-From-Line support for two reasons: (1) it seems to have
disappeared from newer versions of Gnus, and (2) this module is for mbox
format. (Thanks to Brian May <bam@debian.org> for prompting this.)
- Changed the mail parsing so that a blank line *must* separate mail messages,
as per mail(5) (http://www.qmail.org/man/man5/mbox.html). Lack of a blank
line will cause the second email to be considered to be part of the
preceding email.
Version 1.3001: Mon Jun 6 2005
- Changed the testing code to use a more aggressive technique for clearing any
existing cache, even if it is not readable.
- Fixed a problem with the grep implementation where locale settings (LC_ALL,
LC_COLLATE, LANG, LC_CTYPE, LC_MESSAGES) would cause it to fail. (Thanks to
Joey Hess <joey@kitenet.net> for the bug report.)
Version 1.3000: Mon Mar 14 2005
- Merged the internal caches used by the different mailbox parser
implementations. This allows sharing of caching information. NOTE: Tighter
integration of the classes means that you can no longer instantiate any of
the implementations directly; you must use Mail::Mbox::MessageParser only.
- Fixed goofy version dependency for Benchmark::Timer.
- Fixed improper identification of mailboxes whose first email has a body
containing a large number of foreign characters. (Thanks to Nigel Horne
<njh@bandsman.co.uk> for the bug report and sample data.)
- Fixed a spurious warning in a test case. (Thanks to Nigel Horne
<njh@bandsman.co.uk> for the bug report.)
- Fixed a test code bug that would cause some failed tests to incorrectly
pass.
- Fixed a bug where emails with attachments would cause the mailbox parser to
enter an infinite loop. (Thanks once again to Joey Hess <joey@kitenet.net>
for the excellent bug report.)
- Fixed a bug where the cache would not be saved to disk when the file was
finished being read.
- Fixed a bug in the test cases where some differences in test output would
not be detected.
- Fixed a bug in Mbox::Mail::MessageParser::Grep that would cause it to
improperly identify separate emails in a mailbox if the "From " line looked
like "From klopp Mon Jan 5 08:50:15 +0100 2004". (Thanks to Frederic Klopp
<klopp@math.univ-paris13.fr>" for the bug report and sample mailbox.)
Version 1.2130: Tue Dec 21 2004
- Fixed version numbers, which were incompatible with some modules. (Thanks to
Tassilo von Parseval <tassilo.von.parseval@rwth-aachen.de> for the bug
report)
Version 1.21.2:
- Fixed a dependency version error for Benchmark::Timer
- Switched to Test::More for better error reporting
- Improved test failure reporting
- Fixed a bug where messages embedded in other messages as RFC 822 attachments
would be treated as separate emails. (Thanks to Will Fiveash
<william.fiveash@sun.com> for the bug report.)
- Moved all configuration data to M::M::MP::Config
Version 1.21.1: Thu Sep 16 2004
- Fixed broken dependency specification in Makefile.PL
- Removed Module::Install extension dependency code from Makefile.PL. (Not
needed with fixed extensions.)
developer is officially maintaining the package.
The rationale for changing this from "tech-pkg" to "pkgsrc-users" is
that it implies that any user can try to maintain the package (by
submitting patches to the mailing list). Since the folks most likely
to care about the package are the folks that want to use it or are
already using it, this would leverage the energy of users who aren't
developers.
module directory has changed (eg. "darwin-2level" vs.
"darwin-thread-multi-2level").
binary packages of perl modules need to be distinguishable between
being built against threaded perl and unthreaded perl, so bump the
PKGREVISION of all perl module packages and introduce
BUILDLINK_RECOMMENDED for perl as perl>=5.8.5nb5 so the correct
dependencies are registered and the binary packages are distinct.
addresses PR pkg/28619 from H. Todd Fujinaka.
Changes since 1.14:
- More robust use of GNU grep to find emails on DOS systems
- Fixed an uninitialized value warning in Cache.pm
- Made everything work with DOS-style line endings. Added endline() function
to return the detected line ending
- Now keeps reading the first paragraph until a maximum number of bytes have
been read.
Mail::Mbox::MessageParser is a feature-poor but very fast mbox
parser. It uses the best of three strategies for parsing a mailbox:
either using cached folder information, GNU grep, or highly optimized
Perl.
XXX: might need some adaptation for OSs where diff and bzip2 are
not in the base system.
