Hi,
On running compiler-tester, cygwin/bash command line expands the
argument file wildcard (test-*.cs etc.) to a full-length argument
list. Thus compiler-tester never runs fine on cygwin.
I created a patch to avoid this problem by adding extra "--pattern:"
prefix onto the wildcard argument.
It
Hi there.
I just got kinda confusing error string with mcs:
ServerLink.cs(35,24): error CS0118: `ServerLink.System.Net.Dns' is a
`type' but a `type' was expected
the line has an error, but this error string isn't very helpful, neither
is logical. The bad line was:
IPAddress addr = new Dns.
The patch looks ok to me.
El lun, 13-03-2006 a las 09:19 -0800, Vladimir Krasnov escribió:
> Hello,
>
> Action property has quoted value in ProcessMessage method on server
> side, but it should be without quotes. Look attached sample that
> reproduces the problem.
>
> Please approve the attached
This patch is not correct. You can't add a protected property that's
doesn't exist in the MS api. Moreover, I don't think this is the right
way of updating the request. Why not set the initial content type to the
message, process it, and then update the request from the resulting
message?
El lun,
Did you put the directive on the ascx, to have it inherit from your
class MyControl?
If not the created control descend from Control and can't be cast to a
MyControl.
Remember that the source for a new class is created from the xml/html
in the ascx/aspx file in ASP.NET 1.x. That class is compiled
Hi all,
I have such a piece of code:
class MyControl : UserControl
{
.
}
class MyPage : Page
{
private void Page_Load(...)
{
MyControl control = ( MyControl ) LoadControl( "mycontrol.ascx" );
}
}
At line MyControl control = ( MyControl ) LoadControl( "mycontrol.a
Hi,
If they are invoked from the runtime or somewhere else (either managed or
unmanaged code) I think the best solution is to disable the warnings and put
a comment that describes that it is invoked externally (more intuitive with
the name of the method that invokes it) so that it will not be acc
This version of the patch is cleaner, and finally shows some
results:
[1] no common benchmark gets worse, and
[2] code heavily using accessors gets faster.
As an example of [2], I wrote a simple numeric benchmark, a
sort of particle simulator, which you can find here:
http://primates.ximian.com/~
Working on treeprop, I noticed that sometimes the SciMark2 score
got worse, but only on amd64.
Looking at the generated code, I found that the problem was that
a variable heavily used in deep nesting BBs (level 3) was left on
the stack, while variables used at level 1 were put in registers.
So I
fedotenko wrote:
Im a bit new to this. Wher do i file a bug?
Oops, sorry for that. It's at http://bugzilla.ximian.com/ .
Or in case you couldn't understand the bugzilla interface, posting
the stylesheet and the xml source is still ok.
Atsushi Eno
___
10 matches
Mail list logo