Hi Henry
I agree, i think the bug is in IIS.
We havent been able to find anything about this bug on the microsoft site.
Also we havent reported it.
But we think therese buckleys chance of gettting microsoft to fix it.
They'll probably say its fixed in IIS 5.0.
We have modified the isapi_redirector.dll code for our own purposes.
I guess its up to you guys about how you want to approach the situatation.
Cheers
Sam
-----Original Message-----
From: GOMEZ Henri [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, July 10, 2001 5:55 PM
To: [EMAIL PROTECTED]
Subject: RE: Possible bug in isapi_redirector.dll
Hi Sam,
>-----Original Message-----
>From: sam [mailto:[EMAIL PROTECTED]]
>Sent: Tuesday, July 10, 2001 9:31 AM
>To: [EMAIL PROTECTED]
>Subject: Possible bug in isapi_redirector.dll
>
>
>Hello there everybody, We think we may have found a bug,
>possibly in IIS,
>which affects isapi_redirect.dll. We were wondering if anyone
>else has seen
>this. Perhaps this is a bug for MS, but it seems to affect
>tomcat so maybe
>someone here can shed some light.
>
>Our software versions are:
>IIS 4.0
>jk_isapi_plugin.c revision 1.5.2.2
>tomcat 3.2.2
>Windows NT 4.0 sp6
>
>
>
>The relevant code in is the file jk_isapi_plugin.c, line 960 in the
>function:
>
>static int get_server_value(LPEXTENSION_CONTROL_BLOCK lpEcb,
> char *name,
> char *buf,
> DWORD bufsz,
> char *def_val)
>
>
>When the function lpEcb->GetServerVariable(...) on line 965 is
>called to
>retrieve CERT_SERIALNUMBER, the value returned in bufsz is wrong. Our
>certificate serial number is 48 chars long, but the value
>contained in bufsz
>is 47, where it should be 49 (the length of the serial number
>+ 1 for string
>terminator).
>After the call to GetServerVariable() a '\0' is inserted into buf at
>bufsz-1. This causes our certificate serial number to be truncated by 2
>bytes.
>
>The microsoft doco says that the value returned should be the
>value length +
>1 for the string terminator. This is from:
>http://msdn.microsoft.com/library/en-us/iisref/html/psdk/asp/is
re8376.asp?fr
ame=true
The problem is clearly in IIS. It return 47 instead of 49....
Could you see if the bug is reported in IIS ????
