[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [cp-patches] TimeZone/VMTimeZone split
From: |
Michael Koch |
Subject: |
Re: [cp-patches] TimeZone/VMTimeZone split |
Date: |
Sun, 29 Aug 2004 19:44:00 +0200 |
User-agent: |
KMail/1.6.2 |
Am Sonntag, 29. August 2004 19:28 schrieb Mark Wielaard:
> Hi,
>
> This splits java.util.TimeZone into a platform independent and a
> platform dependent VMTimeZone class. It also incorparates the native
> code from libgcj (translated to JNI from CNI).
>
> The only (big) flaw is that it uses malloc() at this time (where
> libgcj uses _Jv_Malloc of course). We should factor that out somehow
> in the future. It is a little nicer then the libgcj version in that
> it gets rid of the printf() call which some embedded devices like to
> ommit whenever possible.
>
> Since the flawed, non-working target native misc macro isn't used now
> anymore (and I didn't see any versions of the macro for any non-posix
> platform anyway) I removed it. I have not come up yet with a way to
> integrate it into the target native layer for the moment. I would
> like to do that in a way so that libgcj can really share most of our
> native C/JNI code where possible (and still let them use CNI where
> possible).
>
> If nobody comes up with a much better approach I would like to check
> this version in now and make it "perfect" later when merging with
> libgcj again. It is already pretty big and much better then the
> (non-working) code we have now.
> Comments?
Generally looks good. Personally I would escape the arguments to AC_*
macros (e.g. AC_CHECK_FUNCS(foo bar baz) to AC_CHECK_FUNCS([foo bar
baz]). I think this is more safe with future versions of
autoconf/automake.
Michael