Chris Quenelle wrote:
> It's likely that Sun Studio Next will have -xstrconst on by default.

What about the idea to add an option (e.g. "-xmergeconstdata" ; not
enabled by default) which extends the functionality of "-xstrconst" to
all constant data regardless of datatype (assuming content is the same,
alignment requirements are met and the storage flags are compatible) ?
AFAIK it's a ISO C violation but IMO this may cause a huge win for some
kind of applications (e.g. applications with multiple huge vector tables
which have lots of near-similar data)

> It's likely that the C compiler will transition to using the same
> syntax as the C++ compiler uses today.  (-features=[no%]conststrings)

Erm... is there a way to backport the new option to Studio 10 and Studio
11 to allow existing projects to use the matching options on demand
without the need for something like "autoconf" probes (somehow I doubt
anyone wants the old behaviour back... but it may be better to be
prepared to avoid "-m64"-style "fun...") ?

> Hopefully we will remember the -m64 lesson, and silently accept
> both options for at least one FCS release.  This change has already
> been approved by the compiler internal architecture team.  Now it's
> just a matter of implementing the change in the compiler driver.
> It's not known yet if this will be a "required feature" in the
> official project plan for Sun Studio Next.

What does that mean ?

----

Bye,
Roland

-- 
  __ .  . __
 (o.\ \/ /.o) roland.mainz at nrubsig.org
  \__\/\/__/  MPEG specialist, C&&JAVA&&Sun&&Unix programmer
  /O /==\ O\  TEL +49 641 7950090
 (;O/ \/ \O;)

Reply via email to