Re: [Mingw-users] Linking against MSVCR71.DLL
- Date: Sat, 02 Feb 2008 14:26:14 -0500
- From: wilson1442@xxxxxxxxxxxxx
- Subject: Re: [Mingw-users] Linking against MSVCR71.DLL
Brian Dessent wrote:
> Just for the record: There are very good reasons to only target
> MSVCRT.DLL. For one, it is the only one that is guaranteed to exist
> on Windows.
One so seldom sees "Windows" and "guaranteed" in the same sentence :) .
> In all other cases, you must make provisions to distribute the
> library along with your program.
It is indeed a true pain. I must actually install *two* copies of the
silly thing along side two referencing executables: Python and my own
EXE which I was foolish enough to locate in a separate directory. I
don't want to install it in the "standard place", probably for the same
reason Billy doesn't want anyone messing with his precious MSVCRT.DLL.
> Whether you are allowed to distribute this DLL with your software is
> unclear if you do not own a license to MS Visual Studio.
I'm piggybacking on Python's copy. I would have been perfectly
satisfied with the forbidden fruit, but the logistics of switching back
and forth would make my head explode.
> Furthermore, it may consitute a license violation of the GPL if you
> were to link GPL code to MSVCR71.DLL because it would be creating a
> combined work that is not GPL. The only reason that this is
> permissable when linking to MSVCRT.DLL is that the GPL contains an
> exception clause for "components of thee operating system" which
> MSVCRT is but the other versions are not.
So I may have to explode my head :) . I'll wait until Sir Richard
knocks on the door.
> Er, those are very old versions. mingw-runtime 3.10 is about a year
> and a half old (current is 3.14) and w32api 3.1 is about 3.5 years
> old (current is 3.11).
I tried the current version, but the cross-build procedure, which seems
wonderfully slick, is unfortunately borked. Something in autoconf, I
expect. I filed a report.
> The problem is libmoldname. As you might know, the MS runtime spells
> stat as _stat and so on. ...
I noticed Microsoft seem overly fond of '_'. In fact, some poking
around with nm seems to reveal that sometimes they prepend *two* of them
just to keep you on your toes. IThinkTheyPreferReallyLongIDs, and
throwing in an underscore is their way of subverting short,
easy-to-remember ones.
> But libmoldname ...
I saw "-lmoldname", and wondered what it was. I try to keep the mold in
my programs to a minimum.
> [fix] is simple enough, if you look in Makefile.in you will see:
> [...snip...]
Indeed it seems simple enough that even I can handle it.
Thank you very much Brian. My first experience with this list is a
wonderful contrast to a newsgroup I've had the misfortune to read
lately. If you have some time to waste, peruse
microsoft.public.platformsdk.msi; the contrast is quite stark.
Jim Wilson
Gainesville, FL
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
MinGW-users mailing list
MinGW-users@xxxxxxxxxxxxxxxxxxxxx
You may change your MinGW Account Options or unsubscribe at:
https://lists.sourceforge.net/lists/listinfo/mingw-users