On Sunday 17 May 2009 07:50, Rob Landley wrote:
> On Saturday 16 May 2009 09:17:55 Denys Vlasenko wrote:
> > On Friday 15 May 2009 10:17, Rob Landley wrote:
> > > By the way, what the heck is smalluint?  Why isn't uint16_t usable?  c99
> > > was a full decade ago...
> >
> > Because bytes are more efficient on x86. On ARM, you'd rather use
> > full-sized words.
> 
> *shrug*  A bit fiddly for my tastes.  (Don't command line arguments get 
> promoted to a full sized words when passed on the stack or in a register?

make bloatcheck says smallint and smaluint reduce code size on x86.
When I globally change them to [unsigned] int, it says:

(add/remove: 0/0 grow/shrink: 268/51 up/down: 2221/-237)     Total: 1984 bytes
   text    data     bss     dec     hex filename
 816965     467    6960  824392   c9448 busybox_old
 818832     473    7076  826381   c9c0d busybox_unstripped


> Doesn't your patch promote the return type to a long long, which is darn 
> fiddly to generate _any_ code for on 32 bit platforms?)

Feel free to propose alternative patch.


> > > It's an improvement, but not a full fix.  Jumping to the left edge and
> > > overwriting data with a prompt is behavior other shells don't do for a
> > > _reason_.
> >
> > Please try this patch. It prints "ESC [ 6 n" before each prompt
> 
> Sounds good in theory, but...
> 
> patching file include/libbb.h
> Hunk #1 succeeded at 962 with fuzz 2.
> Hunk #2 FAILED at 975.
> 1 out of 2 hunks FAILED -- saving rejects to file include/libbb.h.rej
> patching file libbb/lineedit.c
> Hunk #3 FAILED at 1440.
> 1 out of 3 hunks FAILED -- saving rejects to file libbb/lineedit.c.rej
> patching file libbb/read_key.c
> Hunk #2 succeeded at 147 (offset -1 lines).
> 
> No idea what the patch is against...

It is against latest git tree. You can retrieve it with

git clone git://busybox.net/busybox.git

--
vda
_______________________________________________
busybox mailing list
[email protected]
http://lists.busybox.net/mailman/listinfo/busybox

Reply via email to