RE: return type demotion remarks

2021-11-28 Thread David Laight
From: Bernhard Reutner-Fischer
> Sent: 28 November 2021 19:37
> 
> On Sun, 28 Nov 2021 18:00:31 +
> David Laight  wrote:
> 
> > From: Bernhard Reutner-Fischer
> > > Sent: 27 November 2021 20:57
> > >
> > > Hi!
> > >
> > > I'm attaching a color log of some spots that may benefit of using
> > > narrower return types; See "could return".
> >
> > My mailer won't expand a .gz file .
> 
> yea, it was too big to send uncompressed, sorry for that.
> >
> > However making function arguments and return values smaller
> > than a normal cpu register (long is probably better than int!)
> > is likely to increase code size.
> 
> Arches will promote types as they see fit anyway.

The problem is that the values start needing to be masked to the
smaller type, not that they get promoted.

> I'm attaching something that i would _not_ apply since it was based off
> the user-visible fix-it not taking into account the shifts needed in
> the callers.
> Lessons learned in this v0.0:
> - iterate_on_dir callback should be void
> - find ACTF could use a narrower type
> - functors are a bit of a pain ;)
> cheers,

Try using outlook :-(
It really only wants to execute attachments, not display them.

David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, 
UK
Registration No: 1397386 (Wales)

___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox


RE: return type demotion remarks

2021-11-28 Thread David Laight
From: Bernhard Reutner-Fischer
> Sent: 27 November 2021 20:57
> 
> Hi!
> 
> I'm attaching a color log of some spots that may benefit of using
> narrower return types; See "could return".

My mailer won't expand a .gz file .

However making function arguments and return values smaller
than a normal cpu register (long is probably better than int!)
is likely to increase code size.

David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, 
UK
Registration No: 1397386 (Wales)
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox