Mark == Mark Wielaard [EMAIL PROTECTED] writes:
Mark I see you have recently closed that PR. But while merging classpath with
Mark libgcj I noticed the InflaterInputStream.java is still different in both
Mark classpath trees (strangely enough libgcj doesn't use an override, but
Mark has an
Tom == Tom Tromey [EMAIL PROTECTED] writes:
Tom I'm reverting the patch for now.
Tom I'll debug it later and write a new mauve test as well.
Tom Sorry about that.
I looked into this a little the other day.
The Classpath InflaterInputStream will break libgcj's zip reading.
The libgcj
Hi Tom,
On Fri, 2006-03-10 at 14:07 -0700, Tom Tromey wrote:
2006-03-10 Tom Tromey [EMAIL PROTECTED]
* java/util/zip/InflaterInputStream.java (read): Replace with libgcj
implementation.
* java/util/zip/GZIPInputStream.java (readHeader): Use DEFLATED,
not '8'.
Mark == Mark Wielaard [EMAIL PROTECTED] writes:
Mark This breaks loading resources from zip files (at least with jamvm).
Sorry. I'll fix this soon.
Meanwhile you could just revert the patch.
Tom
Mark == Mark Wielaard [EMAIL PROTECTED] writes:
Mark This breaks loading resources from zip files (at least with jamvm).
I'm reverting the patch for now.
I'll debug it later and write a new mauve test as well.
Sorry about that.
Tom
Index: ChangeLog
from Tom Tromey [EMAIL PROTECTED]
I'm checking this in.
While merging more of java.util.zip in libgcj (we now only have 2
classes that differ) I found one method where I preferred libgcj's
implementation. The replacement removes an InternalError, which I
think is a definite improvement, the rest seems more cosmetic (but