FYI, the same applies to all major Linux distributions (Debian and
Ubuntu come to mind), so it's not just RedHat.

> The main argument I have against it is that it creates more work for 
> developers. RedHat library issues aside, what do we actually lose by not 
> including the version numbers in the file names?

1) Standardization.  At least from the sampling that I've looked at
(which is rather large), every JAR file I saw in a release archive had
its version in the name, including every single Apache project that I
looked at.  Given an industry standard, that should be the starting
point from which we base our decision.  The onus is on the
counter-point to prove a compelling reason to not put our version in
there.  From this debate, clearly those reasons are in dispute, which
is why we call for a vote.

2) Possible caching issues in the legacy (pre-6u10) java plug-in.  To
be clear: I haven't researched this - the point just occurred to me a
few minutes ago.  But I know that file names are often used as cache
keys, and as much as user agents *should* issue a HEAD request, they
often don't.

Reply via email to