You'll never believe it, but I actually just put a release on the web site
(http://inline.perl.org).
It has one single patch:
For interpreted ILSMs, validate() is called before load() even when build()
is not called. If your name is not Mark Fowler, this probably won't mean
much to you.
This is just the start. I have a short list of patches to make. Here's a ToDo
file that's inside the distro:
# 0.44 To Do List
done:
- Mark Fowler fix: >
This fix involves calling validate() even when build() is not
called. It affects only the interpreted ILSMs
High priority:
- Inline::MakeMaker replacement: >
This rewrite is completed for the core Inline. Should be easy to
bring over.
- Mitchell Charity patch: >
Mitchell wrote a drop in replacement for the Parse::RecDescent
one. We need to decide how to enable this. I previously tried a
PARSER option. But now it seems that a more powerful glue option
would be better.
- use File::Spec: >
It is top priority to turn all file path operations under the
control of File::Spec, so that Inline can be ported to other
platforms.
Medium Priority:
- Redo the test suite to be faster. Eliminate duplicate tests.
- Apply user patches
- Regression tests: >
These can be run outside of make test. They can use YAML, etc.
Lower Priority:
- Inline/Inline::Devel/Inline::Build split:
This needs to happen soon. Maybe in 0.45.
- Documentation review
- Add an Inline typemap to replace the Ext::Utils one
- Allow *.h typemap *.c from cwd() to work without full pathnames
- Remove the 'config' complexity
Cheers, Brian
