[Mono-list] JQuery

[Simon Lindgren]

sön 2011-08-28 klockan 21:40 +1000 skrev jupiter.hce at gmail.com:
> It was a simple test running javascript on client web browser to a web server 
> running on Apache & mono (with xsp and mon-mono) on CentOS 5.
> 
> The javascript called:
> 
> $.ajax({
>   url: myUrl,
>   data: dataString,
>   dataType: "json",
>   success: goodCallBack,
>   error: errorCallback
> });
> 
> If the dataString contained "&" for example dataString = "broken&data" or 
> if the dataString contained an empty string dataString = "", the web
> server would treat the dataString as a undefined object and the errorCallback 
> would be called.

What is the error, do you get a stacktrace or a generic HTTP 5xx from
apache?

> 
> The same dataString would be OK if it sends to a Windows web server
> running on ASP .Net.

Are you using ASP.Net WebForms, ASP.Net MVC, Castle MonoRail, or
something else?

> 
> If the dataString contained a space character, the Linux web server
> replaced the space character by "+" but the Windows web server still
> took it as a space character. 
> 

This might be because (for some reason) the url is not being decoded
correctly. See for example
http://msdn.microsoft.com/en-us/library/4fkewx0t.aspx which describes
the reason for url encoding briefly.

Since you are using a GET request, Url encoding is definitely necessary,
but if decoding is not done spaces will end up being replaced by +
characters. This should most definitely work though.

> If the dataString contained "=' character, the Linux web server would cut 
> the string "=" and any characters after the "=", the windows server was
> fine.
> 

If url encoding was the issue I would expect strings containing
ampersands rather than '=' to be the issue, and I would expect the
string to be truncated instead of causing an error.

> Also, if I open several client browsers to send dataString every
> second to the same web server, the Linux web server would stop the
> process after 20 minutes, but there was no problem on Windows server.
> 

In that case, I don't think this is only one issue.

-- 
Simon Lindgren