Mindless requirements are always with us. I once worked on a project
where the word was passed on from on high that there would be no "magic"
numbers in the code all such number would be defined in the header file
for that translation unit with a meaningful name. Well for some numbers
there was no meaningful name, they were just numbers. I wish I had
thought of this, but another programmer on the project produced a header
file and code that included this...
somefile.h
#define one 1
#define two 2
#define three 3
somefile.c
switch ( myNumber )
{
case one
.
case two
.
case three
.
}
Doug Franklin wrote:
On Wed, 13 Jul 2005 21:32:53 -0400, Mishka wrote:
http://source.winehq.org/source/dlls/winsock/socket.c#L2542
That source code is a perfect example of what's wrong with WINE in my
opinion ... the comments tell you nothing about why it does what it
does and tell you a lot about what it's doing, which you could figure
out easily enough without any commentary whatsoever. I have to believe
that somewhere, someone has documentation that explains the division of
labor between the wine-preloader processes and the WINE implementations
of the Win32 API, and the wineserver process, and the rationale for
dividing things that way. I just haven't found it yet. And I suspect
it will just p*** me off when I find it. :-)
TTYL, DougF KG4LMZ
--
When you're worried or in doubt,
Run in circles, (scream and shout).
