Hello all,
Sorry I missed all the recent excitement. And double-sorry for being so
crap over the past 6 months or so.
I've been working on a major project for a customer that has occupied me
in totality. I haven't been able to keep up with anything much on TT2
(or anything else - including sleep!) in that time.
So I'm extremely grateful to Adam for giving up on waiting for me to fix
the outstanding problems and going ahead with a release.
I've now merged his changes from 2.16 and 2.17 back into CVS and released
a shiny, new 2.18. This also includes all the extra bits from developer
releases 2.15a - 2.15c that I believe were missing from 2.16 and 2.17.
I've also fixed the tyop in Template-XML and released a new version of that,
too. And I fixed the problem in CVS that was stopping anyone from checking
the CVS source out to OSX (which in turn was stopping me from fixing the
original problem). The house of cards has been taped back together until
I have a tuit free to migrate the whole thing to svn.
I've uploaded the modules to CPAN and they should be appearing RSN. I'll
be getting back to normality with a few weeks when I'll be able to address
any outstanding issues with TT2 and get back to work on TT3.
Attached below are the summarised Changes for this release going back to 2.15:
Of particular interest (other than the fact that you can now install it on
your shiny powerbooks without errors) is that you can now use full, nested
expressions as arguments to subs/method calls. e.g.
[% add(a+5, b < 10 ? c : d + e*5) %]
You won't believe how hard it was to get that working! <ahem/>
I just hope it makes up for my recent crapness in some small way :-)
Cheers
A
#------------------------------------------------------------------------
# Version 2.18 - 9th February 2007
#------------------------------------------------------------------------
* Merged in Adam's changes in 2.16 and 2.17 back into the developer CVS
repository and added his name to the credits.
* Changed the parser grammar to accept expressions as arguments to a
subroutine, method or virtual method call. I'm embarrassed to admit
that it was a one line change that could (and should) have been made
long ago, if only I had realised just how trivial it was. Anyway,
you can now write nested expressions like this:
[% add(a+5, b < 10 ? c : d + e*5) %]
* Put the t/fileline.t test back in as this was fixed in 2.15a
* Added the Template::Toolkit documentation-only module.
#------------------------------------------------------------------------
# Version 2.17 - 8th Feb 2007
#------------------------------------------------------------------------
Another interim release from Adam Kennedy.
* Change in Makefile.PL to force an upgrade to File::HomeDir 0.64 on darwin.
This is due to problems caused by changes made to Perl on the new Intel
versions of Mac OS X.
* skip_all filelines.t on darwin
#------------------------------------------------------------------------
# Version 2.16 - 23rd Jan 2007
#------------------------------------------------------------------------
Interim release from Adam Kennedy.
* Skip fileline.t on Win32, as it has some hard-coded path seperator
assumptions. This will be fixed more comprehensively later.
* Handle spurious errors in Makefile.PL when a dev version of
ExtUtils::MakeMaker is installed.
* Don't say "nmake" on Win32 when $Config{make} is 'dmake'.
This corrects the message on Strawberry Perl.
#------------------------------------------------------------------------
# Version 2.15c - Not released ## DEVELOPER RELEASE ##
#------------------------------------------------------------------------
* Fixed a bug in Template::Parser which caused it to get confused about
block names if it encountered a syntax error inside a BLOCK. Thanks
to Bill Moseley for reporting the problem.
http://tt2.org/pipermail/templates/2006-July/008815.html
* Fixed a minor buglet in Template::Provider which came to light while
investigating the above problem. If a previously cached template is changed
on disk and then fails to compile, the provider now invalidates the cache
entry immediately. Without this fix, the provider would report the error
once, then reuse the cached good version of the template until $STAT_TTL
ticked over when it would try to load and compile the disk version again.
The problem was that error messages were only reported once every $STAT_TTL
second(s) and any requests for the same template in the interim time would
mysteriously work. This way errors get reported consistently and immediately
and no-one has to waste an afternoon trying to figure out where the errors
went!
#------------------------------------------------------------------------
# Version 2.15b - 30th May 2006 ## DEVELOPER RELEASE ##
#------------------------------------------------------------------------
* Changed the uri filter to escape all reserved characters as per
URI::Escape and RFC2396. This now includes &, @, /, ;, :, =, +, ?
and $ which were previously not escaped. Thanks to [EMAIL PROTECTED]
for reporting the problem.
http://rt.cpan.org/Ticket/Display.html?id=19593
* Also changed the uri filter to encode all wide characters as the
equivalent UTF escapes. Thanks to Jonathan Rockway for reporting
the problem.
http://rt.cpan.org/Ticket/Display.html?id=19354
* Fixed the redirect filter to not support relative paths. Thanks to
Paul Seamons for spotting the problem and providing a solution.
* Moved all the virtual methods out of Template::Stash and into
a new Template::VMethods module.
* Fixed the version number of Template::Stash which had rolled over
to 2.102 making it appear to predate the 2.86 stash in TT v2.14.
Thanks to Randal Schwartz for reporting the problem. Changed all
version numbers in other modules to be a hard-coded numbers instead
of grokking it automagically from the CVS revision.
* Changed the _recover() method of Template::Service to check if the
error thrown is a Template::Exception object rather than just a
reference. Thanks to David Wheeler for reporting the problem.
http://rt.cpan.org/Ticket/Display.html?id=17630
* Fixed the some tests in stash.t and stash-xs.t which were failing
under Perl 5.6.2 due to a slightly different error message being
generated. Thanks to Anton Berezin for reporting the problem.
* Fixed a bug in the Template::Provider _load() method to check that
$data is a hash ref before trying to mess with its innards. Thanks
to [EMAIL PROTECTED] for reporting the problem.
http://rt.cpan.org/Ticket/Display.html?id=18653
#------------------------------------------------------------------------
# Version 2.15a - 29th May 2006 ## DEVELOPER RELEASE ##
#------------------------------------------------------------------------
* Removed the latex filter from Template::Filters and related config
variables from Template::Config
* Changed the t/fileline.t test to remove the line number from what Perl
reports as "(eval $line)". It appears to get the $line wrong on
FreeBSD, although the correct line number is reported following that
so the tests still do the right thing. Thanks to Anton Berezin for
reporting the problem.
* Changed the t/compile3.t test to do something similar.
_______________________________________________
templates mailing list
[email protected]
http://lists.template-toolkit.org/mailman/listinfo/templates
