pejobo
drop support for old JDKs
I started looking through the code and a lot of it is really messy because
of support for old JDKs. I suggest we assume a modern JDK, say JDK 1.5,
and clean up the code. Simpler code has fewer bugs.
Original issue reported on code.google.com by [email protected] on 6 May 2007 at 2:12
Original comment by [email protected] on 6 May 2007 at 2:12
- Added labels: Type-Enhancement
- Removed labels: Type-Defect
I suggest that dropping support for 1.4 is a mistake.
There are still a lot of systes out there that have not yet migrated from 1.4
to 1.5
For example, at my "day job" (Epocrates, Inc) which is running a cluster of
java
J2EE Tomcat servers they are still running 1.4.2 ... From an IT perspective
there
has been little compelling reason to upgrade to 1.5.
Original comment by [email protected] on 6 May 2007 at 2:09
How about agreeing that the jar should work under 1.4, but that the build
requires
1.5 . That way, I can use generic types which would make the code a lot more
understandable.
Original comment by [email protected] on 6 May 2007 at 6:50
Looking a little more, I don't think I can use 1.5 source to generate a 1.4
target,
though I don't understand why. I will look into this more later.
Original comment by [email protected] on 6 May 2007 at 8:39
I looked into it, and while Sun doesn't support compiling from 1.5 source into
1.4
bytecode, there are converters that will convert 1.5 bytecode into 1.4 bytecode:
http://retroweaver.sourceforge.net/
http://retrotranslator.sourceforge.net/
So I will go forward with 1.5 with the assumption that someone who cares about
1.4
will pick one of these tools and add it to the BeanShell build. I promise to
fix
anything I do that would break 1.4 compatibility, like using 1.5 libs that the
conversion tool doesn't support.
Original comment by [email protected] on 7 May 2007 at 4:46
I think this is acceptable.
The set of people who have not upgraded to Java 1.5 or greater, are likely
mostly
the same set of people who wont want to use this fork anyway, for the same
reasons.
( Not making a value judgement, just that the reasons for not upgrading the JDK
are
about the same reasons for not upgrading BSH )
Original comment by [email protected] on 8 May 2007 at 1:22
Hello,
I agree that there should be upgrades in the programming build environment to help make development
easier. I think that things like upgrading the asm tool and cleaning internal
structures for uniformity are a
universal need. I do think that if you added a tool like those metioned above
to your main build process, you
could save a ton of time in development. It would anit-productive you spend a
month going in one direction
for a feature that could have gone another way and still benefit a wider
community.
You may not even export the translated version out to the community but it would at least be known to you
what are some expected issues. Like "this block using generics and nio proved
to not be backward
compatible". Just maybe to not have to many forks of beanshell and take
advantage of collective innovations.
Yeah it is those who run servers or enterprise environments who would benefit
most as there upgrade
schedule is slower for reasons like support contracts and cross environments
compatability. Desktop
application development / execution tends to be closer to the edge. Good day.
Yemi Bedu
Original comment by [email protected] on 17 Apr 2008 at 11:25
i vote for dropping pre-1.5 support as well
Original comment by [email protected] on 20 Aug 2008 at 5:39
We use this in some embedded environments where java 1.5 is not available. We
can keep using current beanshell, but it would be nce to benefit from
beanshell2 fixes and enhancements here as well.
Original comment by [email protected] on 29 Mar 2012 at 9:00