Re: Request for review: 7022370 Launcher ergonomics doesn't need per-architecture implementations

Hi David, I'm good with these changes, and I'm happy for ergo_zero.c to vanish. Thanks, Gary David Holmes wrote: > Hopefully all interested parties are addressed in the cc lists. > > webrev at: > > http://cr.openjdk.java.net/~dholmes/7022370/webrev/ > > The launcher ergonomics (ergo.c) currently

Re: OpenJDK Forum: Core Libraries Round Table

Dalibor Topic wrote: > Date/Time: Friday May 29th, 8 AM Pacific, 1600 GMT, 5 PM Germany Is that 1600 GMT or 1600 BST (ie 1500 GMT)? Cheers, Gary -- http://gbenson.net/

Re: Review Request: 6812511: Allow Interpreter-only builds

Ah, thank you. Cheers, Gary Martin Buchholz wrote: > [+build-dev] > > Not that it matters much, but build-dev is a better list for this > patch than core-libs-dev. > > Martin > > On Fri, May 8, 2009 at 09:14, Gary Benson wrote: > > Hi all, > > > >

Review Request: 6812511: Allow Interpreter-only builds

Hi all, I have a patch that's been sitting in the bug databases for a while: https://bugs.openjdk.java.net/show_bug.cgi?id=100011 http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6812511 It allows HotSpot to be built interpreter-only, which is required for the Zero assembler port. It touc

Re: [PATCH] 6788196: Array bounds checks in io_util.c rely on undefined behaviour

Martin Buchholz wrote: > I propose to commit this, if Gary and Alan give me thumbs up. Thumbs up, thank you :) Cheers, Gary -- http://gbenson.net/

Re: [PATCH] 6788196: Array bounds checks in io_util.c rely on undefined behaviour

Andrew Haley wrote: > There are some indentation SNAFUs here. Maybe due to tab width set > to 4? That's Martin's patch, which got inlined in the reply. My patch is the attachment. Cheers, Gary -- http://gbenson.net/

Re: [PATCH] 6788196: Array bounds checks in io_util.c rely on undefined behaviour

Alan Bateman wrote: > I checked the pre-OpenJDK history and this bounds check has been so > since 1.2 (10+ years old). It's kinda surprising this hasn't been > noticed with other ports. Did you run into this with an existing > test (JCK or regression/unit)? Ideally we should have more tests to > ca

Re: [PATCH] 6788196: Array bounds checks in io_util.c rely on undefined behaviour

Hi Martin, I like your method of avoiding the overflow, it's a nice idea. I've attached an updated version of my original patch, with that, and with an expanded comment too, to make sure the fix doesn't get reverted later on in the interests of readability or whatever. Can I ask that you file a s

Re: [PATCH] 6788196: Array bounds checks in io_util.c rely on undefined behaviour

Martin Buchholz wrote: > Does this actually change the behavior with recent gccs? I don't think anything changed recently, not on Intel or SPARC, but I develop on PowerPC, and GCC on 32-bit PowerPC seems to overflow to 1, -1 or 0... sometimes. But that's not the point; the behaviour is undefined,

Re: [PATCH] 6788196: Array bounds checks in io_util.c rely on undefined behaviour

Dalibor Topic wrote: > David Holmes - Sun Microsystems wrote: > > > In C, the result of an overflowing add of two signed integers is > > > undefined. > > > > Strewth! That's a surprise to me. I always thought that C defined > > integer arithmetic to always wrap. > > Only for unsigned operands (fro

Re: [PATCH] 6788196: Array bounds checks in io_util.c rely on undefined behaviour

Gary Benson wrote: > File file = new File( > System.getProperty("test.src", "."), "Test6779290.java"); Oops, that should be "Test6788196.java". Cheers, Gary -- http://gbenson.net/

[PATCH] 6788196: Array bounds checks in io_util.c rely on undefined behaviour

Hi all, In C, the result of an overflowing add of two signed integers is undefined. The array bounds checks in readBytes and writeBytes in jdk/src/share/native/java/io/io_util.c, however, rely on the assumption that the result of the overflowing add will be negative. The attached patch fixes. Ch