[darcs-users] Re: Patches are immutable

Mark Stosberg mark at summersault.com
Wed Nov 17 18:20:21 UTC 2004


On 2004-11-17, Juliusz Chroboczek <jch at pps.jussieu.fr> wrote:
>> History is superfluous,
>
> Interesting opinion.
>
> What do you do when you receive a bug report that says ``I'm getting a
> coredump with hello-world current of 13 October 2004.  Here's a stack
> dump.''?  (Of course, the lines in the stack dump make no sense with
> your current version of the code.)

 From my experience receiving and reporting bugs against CVS-hosted
projects, the standard answer is: "Try again with the latest version and
let us know if it's still there". 

The related historical case is a bug in a released version, which should
correspond to a tag in your tree.

> What do you do when a user submits a very interesting patch, but it
> doesn't apply cleanly to your current version?

Well, It'd say that with darcs concept of a "mutual ancestor", you are
a better place than with CVS. 

When I have recieved patches I liked that didn't applied cleanly, I
manually fix them, because it's probably faster than writing and
debuggin the code myself. :)

Like above, you have the options above of getting a patch against
the latest version, or patching against a tagged release. 

> What do you do when (real example I was involved in) the vice-CTO of
> ACME Inc. makes claims that ``Free Software only recently integrated
> that kind of functionality, more than n months after ACME Inc.'' ?

Check the timestamps on your patches? :)

    
    Mark





More information about the darcs-users mailing list