Re: Time to branch off 4.0 from 4.1?

[Hans-Bernhard B. <br...@ph...>]

On Mon, 21 Jun 2004, Ethan Merritt wrote:

> Did you do this manually, or were you able to automate the procedure
> somehow?   

Some automation was possible.  I essentially did an emacs
'tags-query-replace' operation to cover entire source directories, and a
rather tricky egrep search to spot things still left to do (search for a
function definition that has only one word before the first comma or
closing parenthesis of the parameter list).  I expect to have to use more
tools to hunt the K&R definitions missed by egrep; like gcc in ultra-picky
mode, or maybe splint.

> I ask with some trepidation since it seems likely that applying
> this to the cvs tree will break every patchset on SourceForge including a
> couple of large ones that I have been maintaining.  

As long as they leave the function definitions alone, most patchsets
should either be fine, or the conflict resolutions obvious.

> That's not an  argument against doing it, but I wonder if there is a way 
> to apply your ANSIfication to both a patched and unpatched CVS tree
> in parallel so that I can regenerate the patchset to apply against the
> ANSIfied version.

Not really.  The method is O(N) in the number of function definitions,
which means using it on two source trees would take twice as long, and the
majority of that would be wasted duplicate effort.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.