[v3] more sophisticated versioning checks, enable by default

[Phil Edwards]

On Thu, Feb 21, 2002 at 03:39:50PM -0800, Benjamin Kosnik wrote:
> I think this makes this --enable flag consistent with other
> libstdc++-v3 configure flags, allows some systems to immediately start
> using symbol versioning without disrupting other systems, and at the
> same time allows a configuration option so that Loren can purposefully
> build a broken libstdc++ configuration on BSD to work through
> configuration issues. Does this seem fair, and a good way to proceed?

Yes.

> Also, this enable flag should probably be documented on configopts.html.

I was waiting until it stabilized.  :-)

This comment change:

>   dnl       Where DEFAULT is either `yes' or `no'.  If ommitted, it
> ! dnl       defaults to `no'.  Passing `yes' tries to choose a default style
> ! dnl       based on linker characteristics.
> --- 2042,2049 ----
>   dnl       Where DEFAULT is either `yes' or `no'.  If ommitted, it
> ! dnl       defaults to `yes'.  Passing `yes' tries to choose a default style
> ! dnl       based on linker characteristics. Passing 'no' disables versioning.

is actually wrong.  The "if ommitted" refers to the "DEFAULT" flag passed
when calling GLIBCPP_ENABLE_SYMVERS from configure.in, not what gets
passed by the user to --enable-symvers.  (If nothing is passed by the user,
a value of "yes" is dropped in by autoconf.)

"If ommitted, defaults to no" refers to

>   define([GLIBCPP_ENABLE_SYMVERS_DEFAULT], ifelse($1, yes, yes, no))dnl

i.e., what happens in we were to write

    GLIBCPP_ENABLE_SYMVERS

by itself.

I'll fix up that comment when I write something for configopts.html.


Phil

-- 
If ye love wealth greater than liberty, the tranquility of servitude greater
than the animating contest for freedom, go home and leave us in peace.  We seek
not your counsel, nor your arms.  Crouch down and lick the hand that feeds you;
and may posterity forget that ye were our countrymen.            - Samuel Adams