[Patch] fhandler.cc (pust_readahead): end-condition off.

[Bas van Gompel]

Op Mon, 6 Dec 2004 10:16:19 -0500 schreef Christopher Faylor
in <20041206151619.GA11120@trixie.casa.cgf.cx>:
:  On Mon, Dec 06, 2004 at 02:45:10AM +0100, Bas van Gompel wrote:
[...]

: > 	* fhandler.cc (fhandler_base::puts_readahead): Fix end-condition.
:
:   This patch changes things so that len characters are always output if
:  len is != -1.  It has been a while since I worked on this code but it's
:  not clear that that is correct.

I found following clues:


C1:
`cvs annotate -r1.16 fhandler.cc`, line 75:
| 1.1          (cgf      17-Feb-00):   while ((((len == (size_t) -1) && *s) || len--) &&

This code seems to intend to not test ``*s'' when len != -1.


C2:
`cvs annotate fhandler.cc` (for current version), line 57
| 1.17         (corinna  09-May-00):   while ((*s || (len != (size_t) -1 && len--))
ChangeLog-2000 about above change:
| 	* fhandler.cc (fhandler_base::puts_readahead): Change
| 	while condition to disallow wild runs.

The wild runs would occur when *s became 0, len would then be
decremented to -2. This ``new'' code does not seem to want to decrement
len when ``*s'' is nonzero.


C3:
(fhandler_tty.cc (fhandler_pty_master::accept_input):)
...
|      rc = WriteFile (get_output_handle (), p, bytes_left, &written, NULL);
...
| 	  p += written;
| 	  bytes_left -= written;
...
| 	      puts_readahead (p, bytes_left);


I don't think there is a reason to not send all that wasn't written,
to the readahead-buffer.
[...]


I hope this helps.


L8r,

Buzz.
-- 
  ) |  | ---/ ---/  Yes, this | This message consists of true | I do not
--  |  |   /    /   really is |   and false bits entirely.    | mail for
  ) |  |  /    /    a 72 by 4 +-------------------------------+ any1 but
--  \--| /--- /---  .sigfile. |   |perl -pe "s.u(z)\1.as."    | me. 4^re