[tipc-discussion] Re: Submission of Print Buffer Improvements?

[Jon M. <jon...@er...>]

That sound good. You can check it into "unstable" when
you feel ready for it.

/jon

On Wed, 2005-02-23 at 09:07 -0500, Stephens, Allan wrote:
> Hi Jon:
> 
> I've made some changes to the Wind River version of dbg.c to enhance
> the 
> functionality and performance of the print buffer code.  Specifically:
> 
> 1) The print buffers now handle wrap-around they way you would
> expect. 
> (Currently when the end of the buffer is reached, the new text is
> added 
> to the start of the buffer and all previous text is lost.)
> 
> 2) It is no longer necessary to initialize the entire buffer to
> zeroes 
> when the buffer is initialized or reset.  (It now initializes just 2 
> bytes, instead of thousands.)
> 
> 3) The log buffer size is now a global variable rather than a #define 
> constant.  (Allows you to increase the size of the log buffer by 
> recompiling dbg.c only.  Also sets the groundwork in place in case we 
> want to make the log buffer size a user-configurable setting in the 
> future.)
> 
> 4) There are new routines for dumping a print buffer without grabbing 
> the print lock, and for copying the contents of a print buffer to 
> another print buffer.  (I don't think the Linux TIPC code needs
> these, 
> but the Wind River TIPC code does.)
> 
> I'd like to submit these changes to SourceForge, but I'm not sure 
> whether you'd like these added to the stable or unstable branch. 
> Personally I think the unstable branch makes more sense -- it will
> give 
> you (or others) a chance to take a look at the changes & try them out.
> 
> Regards, 
> Al
>