harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Senaka Fernando" <senaka...@gmail.com>
Subject Re: Fixing HARMONY-1187 is a matter of decision
Date Sun, 27 Jul 2008 16:53:56 GMT
On Sun, Jul 27, 2008 at 9:45 PM, Nathan Beyer <ndbeyer@apache.org> wrote:

> Just so I understand, you're saying that an instance of URLConnection is
> assumed to be a JarURLConnection and in some cases it's not. I don't know
> that there's much to decide. If the API says URLConnection and there's no
> documentation/comment/assertion to indicate that is MUST be a
> JarURLConnection, then the code needs to be patched to support this
> condition. It's probably just some quickly written code that no one came
> back to.
> Do you have a suggested patch for fixing the class?

I'm working on it.

My suggestion is simple. If we come across an exception situation,
back-trace and use the default implementation. As the issue explains, RI
handles this, in some manner (they probably use a different approach) and
does not throw an exception, so if we are to throw an exception which might
sound logical and more explanatory we would still be differing from RI.
Therefore, unfortunately in order to not have a difference, we will need to
internally handle the exception and use the default behaviour, this might
sound a hack in some way.

What do you think?


  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message