[Mono-dev] Patch for some obscure cases in Dictionary (small)

Andreas Nahr ClassDevelopment at A-SoftTech.com
Fri Jun 13 11:16:33 EDT 2008


Thanks for testing. If nobody from the maintainers objects I'll commit.
I've been looking for a testcase, but the error is not easily reproducible
without unmanaged code.

Greetings
Andreas

> -----Ursprüngliche Nachricht-----
> Von: Juraj Skripsky [mailto:js at hotfeet.ch]
> Gesendet: Freitag, 13. Juni 2008 14:32
> An: Andreas Nahr
> Cc: 'mono-devel-list'
> Betreff: Re: [Mono-dev] Patch for some obscure cases in Dictionary
> (small)
> 
> Hi Andreas,
> 
> Your changes look good to me, it makes the code cleaner and leaner. I
> was the one who wrote (or rather re-wrote) the code you're changing,
> but
> I guess you need the okay-to-commit from one of the corlib
> maintainer(s).
> 
> - Juraj
> 
> 
> On Thu, 2008-06-12 at 09:50 +0200, Andreas Nahr wrote:
> > Re-sent without input file.
> > Please test the patch and tell me if it is ok to apply. Does not seem
> to
> > break anything for me, but did not test very extensively.
> > You can get the input file from
> > http://shootout.alioth.debian.org/download/knucleotide-input.txt
> >
> > Andreas
> >
> > > While trying to optimize the k-nucleotide benchmark from Computer
> > > Language Shootout (or whatever) I stumbled across a corner case in
> > > Dictionary.
> > > Situation is: We are behaving different from MS.Net and current
> Mono
> > > behavior is somewhat illogical. But the case itself is surely
> outside
> > > of the spec ;) We are changing the indexer itself in the indexer
> (base
> > > problem, concrete is a bit more complex).
> > >
> > > Anyways this could be patched (it's slightly faster, too).
> > >
> > > Happy Hacking
> > >
> > > P.S. Am attaching the benchmark application and the input for it (<
> at
> > > console)
> > >
> > > P.P.S. I AM aware that doing stuff this way is asking for trouble
> ;)
> > _______________________________________________
> > Mono-devel-list mailing list
> > Mono-devel-list at lists.ximian.com
> > http://lists.ximian.com/mailman/listinfo/mono-devel-list



More information about the Mono-devel-list mailing list