Date:	Thu, 5 Feb 1998 11:54:35 +0200 (IST)
From:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To:	Robert Hoehne <robert DOT hoehne AT gmx DOT net>
cc:	andrewc AT rosemail DOT rose DOT hp DOT com, dj AT delorie DOT com, djgpp-workers AT delorie DOT com
Subject:	Re: iostream concern
In-Reply-To:	<34D8F566.DCEB7410@gmx.net>
Message-ID:	<Pine.SUN.3.91.980205115407.28596G-100000@is>
MIME-Version:	1.0

On Thu, 5 Feb 1998, Robert Hoehne wrote:

> > I think he is referring to the one that comes with djdev.  I would suggest
> > trying to move towards usung the new one, since it does not
> I would prefer this too (thatswy I will include it in gpp280b.zip).

Let me remind you the Murfy's Law of Software Releases.  One of its
corollaries says that a need for a feature which was removed from a
distribution as unneeded will arise within a few hours of the release
that had that feature removed.  Let me further remind you that we have
seen this bite us several times in the past.

More seriously, I think we need to wait for some time and gather user
responses before we decide to make g++ supercede DJ's gxx.exe.  For
this reason, I think we should better call it differently.  I don't
have any particular sentiments for gpp.exe, but the point of my
message was that it should not overwrite gxx.exe from djdev.

> At first only a small change:
> 
> [gxx]
> LIBRARY_PATH=%/>;LIBRARY_PATH%

Why is this needed?  Does g++ barf on backslashes?  If so, it should
be changed to support them.  Does it require the variable to end with
a semi-colon, or be all lower-case?  If so, why?

I think we had enough trouble with the above issues to allow one more
case where the exact format of the environment variable makes or
breaks things.

> Yes, this is confusing to me too
> 
> libstdc++.a   -> libstdcxx.a
> libg++.a      -> libgpp.a
> g++.exe       -> gpp.exe
> c++filt.exe   -> cxxfilt.exe
> 
> I think we should come to one technique. Changing ++ to xx or pp, where
> I prefer to change ++ to xx.

I think prior practice is also important.  Since names like libstdcxx
and libgpp were picked up a long time ago, I say let's stick with them.

> Yes, that's a good idea, but then we should aware about reports
> from error messages, when the user unzip gpp280b.zip, about the
> invalid g++.exe (even when all the other is extracted OK, which
> I tested).

The error message is printed by PKUNZIP, right?  InfoZip does the
right thing (extracts it as g__.exe) and never complains.

- Raw text -