Re: Maximum recursion depth exceeded

On Tue, Jul 7, 2009 at 3:16 PM, Dimitre Novatchev<dnovatchev@xxxxxxxxx> wrote:
> On Tue, Jul 7, 2009 at 1:19 AM, bryan
> rasmussen<rasmussen.bryan@xxxxxxxxx> wrote:
>> On Tue, Jul 7, 2009 at 9:55 AM, Michael Kay<mike@xxxxxxxxxxxx> wrote:
>>>> anyway, recursion problems are likely to be a problem in .NET
>>>> / C#, as far as I know currently the only .NET language that
>>>> can compile and optimize for tail recursion is F#..
>>>
>>> Well, I know of at least two others: XSLT and XQuery, as implemented by
>>> Saxon.
>>>
>> I don't think those would be considered as .NET languages, those are
>> languages implemented in .NET. Sorry if I was imprecise earlier.
>>
>
> 1. Quite some of Bryan's statements about how "System.Xml.Xsl" does
> work are far from correct. One of them is that tail recursion is not
> supported. Tail recursion *is* optimized by XslCompiledTransform in
> non-debugging mode.
>
I never said that Tail recursion is not supported in XSLT or
system.xml.xsl. I said that the only .NET language I am aware of where
it is supported is F# - I'm not sure if I would consider an XSLT
compiled via XslCompiledTransform  to be a '.NET language' (I believe
I used the example of C#) , I haven't really thought about it though -
would you say it is?

> 2. To the original OP: The message by Saxon is generally presented
> whenever Saxon runs out of memory -- this may or maynot be caused by
> recursion stack overflow!

>
> To summarize: Prefer using DVC recursion to tail-recursion as a
> practical way to avoid the recursion stack overflow problem.

you're right of course, specifically as regards the practicality - but
it is annoying when I know I have a function that is properly tail
recursive and it overflows anyway...(This would be a C# problem, not
XSLT in System.XML.XSL)


Cheers,
Bryan Rasmussen