Sorry about the return address - this braindead UI keeps tripping me up.
Thanks for the STLport info. I can use that to keep the build-meister
happy.
Which version of the library are you blaming Dinkumware for?
If my understanding is correct, Plauger started updating their library code
after every committe meeting dating from the July 1993 meeting in Munich.
MSFT has been *SHIPPING* (and programmers using) versions of the Dinkumware
library since 1996. The C++ standard was [technically] frozen in 1998. I
believe Dinkumware has been shipping a conforming version since early 1998
but, to tell you the truth, I'm not sure which version of the library MSFT
is still shipping. I suspect they didn't perform a wholesale replacement
just to maintain conformance - might have broken some existing code. (I
can sympathize. I'm responsible for implementing a set of protocols with
"ietf-draft" in the spec titles.)
David N
Bertoni/Cambridge To:
[email protected]
/IBM cc:
<david_n_bertoni@ Subject: Re: YAUR
(yet-another-user-request)
us.ibm.com>
07/23/2002 06:03
PM
Hi Dick,
Can you make sure your replies go to the list, rather than directly to me?
That way, the discussion stays public.
I believe you can use STLport. I know others have used it successfully.
Unlike you, I can indeed blame Dinkumware for their library. It's their
code, so why didn't they make sure it built cleanly before they shipped it?
Didn't they know people like setting the warning level to 4?
Dave
----- Forwarded by David N Bertoni/Cambridge/IBM on 07/23/2002 03:00 PM
-----
Dick.
[EMAIL PROTECTED] To:
[EMAIL PROTECTED]
cc:
07/23/2002 02:49 Subject: Re: YAUR
(yet-another-user-request)
PM
Oh poo. ;)
You know of anyone that's used STLport instead of the MSFT(1) STL?
(1) I can't bring myself to blame Plauger & Friends for the decisions that
led to MSFT's VC++ 6.0 version of the STL.
David N
Bertoni/Cambridge To: [EMAIL PROTECTED]
apache.org
/IBM cc:
<david_n_bertoni@ Subject: Re: YAUR (yet-
another-user-request)
us.ibm.com>
07/23/2002 01:42
PM
This was actually requested a long time ago, and was rejected. I think the
reasons are still valid:
1. It affects only Win32 so it's extra noise in the header files for the
other platforms. I hate the idea of changing sources to work around
Dinkumware's noisy library.
2. It affects a lot of files, so it's just more busy work. We already
spent tons of time just keeping up with Xerces' changes.
3. Users can modify their copy of the VC++ header files to do the same
thing, which not only obviates the need for Xalan to do this, but cleans
up compiles for other projects built using the same header files.
4. VC++ version 7 has an implementation of the Dinkumware library which
has been cleaned up, so the pragmas are not needed. We'll likely move
to VC++ 7 as our primary Windows compiler fairly soon.
Dave
Dick.
[EMAIL PROTECTED] To:
[email protected]
cc: (bcc: David N
Bertoni/Cambridge/IBM)
07/23/2002 01:25 Subject: YAUR
(yet-another-user-request)
PM
As you modify source files that #include stl headers, maybe you could wrap
the #includes in pragmas like:
<eg>
#pragma warning(push,3)
#include <someheader>
#pragma warning(pop)
</eg>
That would remove 3-4k warnings from the Win32 build.
Shouldn't be a problem for any compilers - they're supposed to ignore
#pragmas that are not recognized.
