Hi Stephen,
I think there's merit in simply porting generics to the current Commons
Collections 3.2 codebase, but I believe we could get better bang for our
buck by discussing how we might reboot Commons Collections. I would
suggest a moderately radical approach:
01. Removing all deprecated code - let's take the opportunity to shrink
the size of the Commons Collections JAR while we can.
02. Removing (*not* deprecating) methods specifically related to type
safety pre-JSE5. I'm thinking of classes such as MapUtils here, whose
methods such as MapUtils#getDouble have probably had their day.
03. Checking our test case coverage, although I don't remember off the
top of my head if Cobertura is plugged into standard Jakarta Commons
components. Question: do we want to upgrade our test cases to JUnit 4 as
well?
04. Porting generics into the remaining codebase.
Once we've agreed upon a strategy (particularly on the subject of
deprecation, where I sense I'm going to upset people), I'm happy to
start submitting issues and patches via Bugzilla/Jira. To be honest, I'm
sick of developers choosing Map<String,String[]> over MultiValueMap and
airily saying "yeah well, it's not genericised".
I agree that JSE5 finally seems to be in the ascendancy, although
ironically it only seems to have happened now that JSE6 is out in the
wild and vendors are racing to catch up. Commons Collections would be a
good starting point for generics (and other JSE5 features?) to be
propagated across the Jakarta Commons projects.
Regards,
Steve.
---
Stephen Smith, MEng (Wales).
http://www.stephen-smith.co.uk/
Stephen Colebourne wrote:
OK Stephens (and anyone else!),
I'm not in a position to just grant commit access (I wish I could, but
thats not the Apache way). However, I am willing to review and apply
patches to get generified collections off the ground.
We already have a generics branch which is intended to be a 'reboot' of
collections rather than a port to generics.
The first decision that needs to be made is whether you want to work on
a port or a 'reboot'. There is merit in both, and I'm willing to go with
the group decision. So start discussing!!!
Stephen
PS. As a general point, I get the sense that Java 5 is now really in the
ascendant, and we could do with generified versions of other commons
components sooner rather than later in order to stay relevant.
Stephen Smith wrote:
Stephen C, if you're looking for volunteers you can count me in -
although 3
Stephens working on the same code might get a bit complicated. :)
On Fri Mar 9 8:28 , Stephen Kestle sent:
Hi, I'd like to help work on the generics stuff for collections, as
I've benefited in the past from it. Stephen (C), it looks like
you're quite busy and could use a hand, so let me know...
Cheers
Stephen
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
