RE: [INTERNALS-WIN] RE: [PHP-DEV] Differences in VC6 and VC9 Windows builds and MSSQL Driver.

Hi Sriram,

Thanks for help. I tested it already. The plugin works with sun webserver
7.0 and it makes no difference if vc6 or vc9 is used in PHP. See my previous
mails on this list for a detailed description (subject containing NSAPI).
The NSAPI public API does not use any important/complex structured
symbols/datatypes/resources of CRT (the only one is "struct stat"), so there
are no resources exchanged between webserver and PHP that may look different
in VC6's or VC9's header file. The only one, "struct stat" is identical in
both CRTs (which I explicitly tested).

And in windows (which is different to UNIX), it is possible to have more
than one CRT in one single process (because linking in windows is *not* only
dependent on the symbol name, it uses symbol name *AND* dll name for
linking).

But I may come back to you in future when the next version is released.
Until now, I mostly talked with Chris Elving when I had problems with the
module.

Just one question NSAPI question in general: Is it planned to make the
internal subrequest API in SJSWS public in later servers releases, so the
"dirty dynamic linking hack" for virtual() can be ifdefed for older NSAPIs?
If not, this is an official "feature request".

Uwe

-----
Uwe Schindler
[email protected] - http://www.php.net
NSAPI SAPI developer
Bremen, Germany
> -----Original Message-----
> From: [email protected] [mailto:[email protected]]
> Sent: Monday, January 05, 2009 8:30 PM
> To: Uwe Schindler
> Cc: 'Pierre Joye'; [email protected]
> Subject: Re: [INTERNALS-WIN] RE: [PHP-DEV] Differences in VC6 and VC9
> Windows builds and MSSQL Driver.
> 
> Uwe
>  I work in  Sun Web Server engineering team and I can help you with your
> testing efforts. The next version of Web Server 7 - is in the process of
> being open sourced under BSD license and might be a good time to use VC9.
> 
> - Sriram
> 
> Uwe Schindler wrote:
> >>> I am not sure, if VC6 can be dropped easily. E.g. some SAPIs that
> >>>
> >> directly
> >>
> >>> map into servers may have problems if using the wrong CRT. Until now,
> I
> >>>
> >> had
> >>
> >>> no time to build up a Windows "Sun Java System Webserver" and test it
> >>>
> >> with
> >>
> >>> both CRTs. The code compiles fine on the snaps box, but I am not sure,
> >>>
> >> if
> >>
> >>> the DLL loads into the server.
> >>>
> >> VC6 will be droped after 5.3.0. VC6 can't be supported anymore, it is
> >> already a real pain to maintain as most of the modern tools or APIs do
> >> not work with it, the licenses can't be built anymore, etc.
> >>
> >
> > I think we should do some announcement to the SAPI maintainers to ask
> them
> > to check their modules with their servers (if the SAPI is available for
> > Windows). For NSAPI, I will do this some time in future (low priority),
> > because it is some work to install a windows server and add SJSWS 7.0 to
> it
> > and possibly debug it (if it does not work). SJSWS is not able to run on
> my
> > local Win XP Prof (I am not sure about this, maybe it works, I will give
> it
> > a try).
> >
> >
> >>> To the snaps build: The NSAPI module tests for "#define ZTS" and does
> >>> therefore not compile without thread safety. Why do the windows non-ts
> >>> builds on snaps.php compile? Is ZTS on windows always defined (it
> looks
> >>>
> >> like
> >>
> >>> so in the makefile). How to test on windows, if thread safety is
> >>>
> >> available?
> >>
> >>> But better would be to disable thread-unsafe builds for the NSAPI SAPI
> >>>
> >> from
> >>
> >>> the beginning.
> >>>
> >> NSAPI is disabled for the NTS builds. They were not explicitly
> >> disabled after I reinstalled the snaps box :).
> >>
> >
> > Thanks, I have not checked this the last days (since release of 5.2.8).
> >
> > Uwe
> >
> >
> >