Pkgsrc changes:
o Add commented-out HOMEPAGE using search.cpan.org
Upstream changes:
Version 2.45: ( Sat Dec 27 17:21:45 IST 2008 )
-----------------------------------------------
* Fixed bug #1230339:
- https://sourceforge.net/tracker/index.php?func=detail&aid=1230339&group_id=6926&atid=106926
- Can't store and retrive "\nEOT\n"
- Fixed by checking for the existence of the here-doc terminator in the
string, and if so appending more characters.
* Fixed bug #1565180
- https://sourceforge.net/tracker/index.php?func=detail&aid=1565180&group_id=6926&atid=106926
- -nocase breaks -default.
- If a default-values section is specified, then -nocase did not look it up
if it's not lowercase.
* Fixed bug #1831654
- https://sourceforge.net/tracker/index.php?func=detail&aid=1831654&group_id=6926&atid=106926
- multiline-values are tainted.
- Added a test as t/17untainted-multiline-values.t
Version 2.44: ( Thu Dec 25 11:44:07 IST 2008 )
-----------------------------------------------
* Eliminated an assignment of $VERSION into itself to "get rid of warnings".
Only caused warnings at Build.PL, and is unnecessary now that we
"use vars".
Version 2.43: ( Thu Dec 4 18:56:02 IST 2008 )
-----------------------------------------------
* Placed IniFiles.pm under lib/Config/IniFiles.pm.
* Added a Build.PL file for the distribution.
* Moved the ChangeLog to its own file - Changes.
* Made the README contain the useful information and nothing else
- previously it was a copy of the module's entire POD.
* Added a regression test for:
https://sourceforge.net/tracker/index.php?func=detail&aid=778593&group_id=6926&atid=106926
When copying tied-hashes representing sections with multiline values as
keys, one used to get a reference instead of a copy. This was fixed in SVN
but not in CPAN.
* Added a test for opening an empty file, inspired by:
https://sourceforge.net/tracker/index.php?func=detail&aid=927111&group_id=6926&atid=106926
* Added a fix for this bug:
https://sourceforge.net/tracker/index.php?func=detail&aid=767913&group_id=6926&atid=106926
With a test that I wrote myself: in list context, ->val() returned the list
containing (undef()) if there was no such key. Now returning the empty
list which evaluates to false.
* Added a test for bug
"[ 2030786 ] Brackets within values are treated as section headers"
It seems to be fixed on SVN.
Revision 2.40 2003/12/08 10:33:13 domq
Documentation tidyup
Revision 2.39 2003/12/06 07:54:21 wadg
[By Proxy for domq]
* Encapsulation of internal data structures even for use within the
class: e.g. ReadConfig() and the TIEHASH interface operate using accessor
methods on $self. Causes massive simplification of the code.
* TIEHASH interface made load-on-demand. Order of each() in TIEHASH
same as configuration file's order (i.e. $self->Sections() and
$self->Parameters()).
* push() and exists() methods
* -allowempty and -deltas parameters to new()
* support for loading config from a SCALAR reference. Unfortunately
this also requires a small patch to IO::Scalar, whose author has yet
to respond to my email
* ->{myparms} and ->{mysects} fields, supporting -delta=>1 option to
WriteConfig()
* _assert_invariants() method (used only in t/01basic.t and
t/02weird.t, maybe should be packaged otherwise)
Revision 2.39 2003/11/10 15:37:48 dom
* Encapsulation of internal data structures even for use within the
class: e.g. ReadConfig() and the TIEHASH interface operate using accessor
methods on $self. Causes massive simplification of the code.
* TIEHASH interface made load-on-demand. Order of each() in TIEHASH
same as configuration file's order (i.e. $self->Sections() and
$self->Parameters()).
* push() and exists() methods
* -allowempty and -deltas parameters to new()
* support for loading config from a SCALAR reference. Unfortunately
this also requires a small patch to IO::Scalar, whose author has yet
to respond to my email
* ->{myparms} and ->{mysects} fields, supporting -delta=>1 option to
WriteConfig()
* _assert_invariants() method (used only in t/01basic.t and
t/02weird.t, maybe should be packaged otherwise)
63dc0cf656