Re: Regression when building static libraries with MSVC

On Mon, 15 Aug 2005 19:50:36 +0200, Bram Moolenaar
<[email protected]> wrote:

[...]

>> If I put the file back to the way it was, everything builds fine.

>This was changed half a year ago.  I can't remember why this was changed
>or who suggested it...

From looking at the archives, it was Lars Ivar Igesund. Here is his
explanation:

----
 You'll have to decide, but building static D libs won't work without it
 (unless I use the {onestep} option). What happens is that msvc's lib 
 action is preferred over D's (the one in dmd.py) because it is missing 
 the language prefix in the variable name. I'm not happy with other 
 tools messing up mine :}
----

Well, he just messed up my tool to make his work. :) Lars Ivar, did you
test your patch? Did it work on your system with MSVC?

Perhaps the prefixes should be there to prevent conflicts, but then
something else needs to be updated somewhere else to actually use the
prefixed variables.

-- 
Be seeing you.



-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf