RE: Small ZipFile patch

2003-03-05 Thread Jeroen Frijters
Tom Tromey wrote: > Jeroen> Another interesting trick with the finalizer is creating > Jeroen> instances of classes that have a private constructor! The > Jeroen> attached runtime.j creates an instance of (a subclass of) > Jeroen> java.lang.Runtime. > > Interesting test case. > > With gij this pr

Re: Classpath AWT problems (or gtk problems...)

2003-03-05 Thread Dhek Bhun Kho
> serialization is more interesting to you then check out 'serialize' > and 'serialcompat' parts of japitools. We should check serialized > compatibility with 1.1 through current releases. Ok. Will do that. Thanks. Bhun. ___ Classpath mailing list

RE: New IO Native Provider Interface

2003-03-05 Thread Jeroen Frijters
Hi, Great job, this is almost exactly like I envisioned it. A few minor suggestions: - FileOutputStream should not request read access, so the modes should be "w" and "a" - write(byte[] buf, long offset, long length) should not use longs for offset and length - nativeReadByte shouldn't return a l

Re: C native layer (pre-announcement)

2003-03-05 Thread Prof. Etienne M. Gagnon
Hi Brian, I will attempt to answer some of your questions. Dr. Rupp will probably correct any error I make in my answers. I'm confused about when to worry about this. Does it matter for synchronized methods? Yes, as the system/library call you are doing might also be done by *another* method

Re: Small ZipFile patch

2003-03-05 Thread Tom Tromey
> "Jeroen" == Jeroen Frijters <[EMAIL PROTECTED]> writes: Jeroen> Sun's ClassLoader has a hack that prevents this from being exploitable: Jeroen> http://www.securingjava.com/chapter-five/chapter-five-8.html Sounds like we need this in ClassLoader and in any other class that performs security

Re: New IO Native Provider Interface

2003-03-05 Thread Aaron M. Renn
Jeroen Frijters ([EMAIL PROTECTED]) wrote: > Great job, this is almost exactly like I envisioned it. A few minor > suggestions: Thanks. > - FileOutputStream should not request read access, so the modes should > be "w" and "a" Good observation. Have you actually tested the JDK to see what it do

Re: java.io.File patch

2003-03-05 Thread Brian Jones
"Jeroen Frijters" <[EMAIL PROTECTED]> writes: > Hi, > > Attached is a patch for object serialization, to use the proper classes > when getting and setting fields in the object (and when calling > readObject/writeObject). Also some minor changes to make caching the > Field and Method objects possi

Re: Class loader optimization patch

2003-03-05 Thread Brian Jones
"Jeroen Frijters" <[EMAIL PROTECTED]> writes: > Hi, > > I made another attempt at reducing the number of ClassNotFoundExceptions > that are thrown when loading a class, this time I believe the patch can > be integrated without breaking anything. No API changes and existing > code should continue

Re: More finalize woes

2003-03-05 Thread Tom Tromey
> "Dalibor" == Dalibor Topic <[EMAIL PROTECTED]> writes: >> java.util.zip.[De|In]flater both have an empty finalize method "for >> compatibility". Is this really necessary? Dalibor> Yes, in order to have reflection on these methods Dalibor> working. A pretty good reason in my opinion. I'm no

More java.io

2003-03-05 Thread Aaron M. Renn
Ok, moving on in the gcj/classpath merger. 1. I'm thinking we should adopt the gcj LineNumberReader. It appears to be more efficient. However, skip() is definitely buggy (inserting regular skip()'s into your code leads to more lines than the file actually has). All Mauve tests pass, so I'll try

RE: New IO Native Provider Interface

2003-03-05 Thread Jeroen Frijters
Hi, Aaron M. Renn wrote: > Jeroen Frijters ([EMAIL PROTECTED]) wrote: > > - FileOutputStream should not request read access, so the > modes should > > be "w" and "a" > > Good observation. Have you actually tested the JDK to see what it > does in this situation? Is the file actually opened "wri

Re: New IO Native Provider Interface

2003-03-05 Thread Stephen Crawley
> Aaron M. Renn wrote: > > Jeroen Frijters ([EMAIL PROTECTED]) wrote: > > > - FileOutputStream should not request read access, so the=20 > > modes should > > > be "w" and "a" > >=20 > > Good observation. Have you actually tested the JDK to see what it > > does in this situation? Is the file actu

TestAWT

2003-03-05 Thread John Leuner
I was trying to debug a GTK peer crash with kissme and TestAWT, but while investigating this I noticed some strange behaviour for the Dialog window. Running this on the linux JDK from IBM and also on a Windows machine, I noticed that you have to push the OK button twice to close the Dialog window.

Re: TestAWT

2003-03-05 Thread Tom Tromey
> "John" == John Leuner <[EMAIL PROTECTED]> writes: John> Running this on the linux JDK from IBM and also on a Windows machine, I John> noticed that you have to push the OK button twice to close the Dialog John> window. John> Has anyone else seen this? The peers have all kinds of bugs. This