ASP.NET applications. The warmer allows you to warm the application on a
website basis which is not what you are looking for. Aside from using ngen,
or rolling your own code to call the method, I'm out of ideas on this one.
The O.W.C. Black Book with .NET
"rerdavies" <rerdavies@discussions.microsoft.com> wrote in message
news:5BCB5D60-A2B8-4573-990A-EDABA5AB1D29@microsoft.com...
> This would be for 2.0.
>
> A global configuration option sounds promising. A method for actually
> forcing the compile without calling the method would be even better, if
> the
> global configuration option does it on a background thread basis.
>
> If you could provide a hook or a starting point (method, search term,
> broad-strokes concept), I would be very grateful. I can run down the
> details.
> For some reason a search on ".net warming" returns lots of information on
> Global Warming of the planetary kind, even on site:microsoft.com. ;-)
>
>
>
> "Alvin Bruney" wrote:
>
>> What framework is this? 1.x or 2? For 2, .NET allows you to warm an
>> application. This is configurable and basically just calls each method -
>> similar to ngen. For 1.x, you will have to write a manual warmer
>> application. You can get clever by wrapping the warmer call inside a
>> static
>> method. When the thread of initialization passes thru sometime before the
>> application actually starts up, your methods should be nice and ready.
>>
>> --
>>
>> ________________________
>> Warm regards,
>> Alvin Bruney [MVP ASP.NET]
>>
>> [Shameless Author plug]
>> The O.W.C. Black Book with .NET
>>
www.lulu.com/owc, Amazon
>> Professional VSTO.NET - Wrox/Wiley 2006
>> -------------------------------------------------------
>>
>> "rerdavies" <rerdavies@discussions.microsoft.com> wrote in message
>> news:655EFF70-F31B-4FB2-ADA1-90ECB6E8568E@microsoft.com...
>> > NGEN.EXE will precompile MSIL to native code; actually calling a method
>> > will
>> > compile the MSIL to native. Is there another way to do this?
>> >
>> > The application: an audio programming language, where the cost of the
>> > initial JIT compile in the pretty-darned-close-to-realtime thread is
>> > not
>> > acceptable.
>> >
>> > The proof-of-concept code , as a temporary workaround, uses a flag to
>> > make
>> > called methods do nothing, just so I can get them compiled. But there
>> > has
>> > to
>> > be a better way to do this.
>> >
>> >
>> >
>>
>>
>>