[Wtr-development] Convio's version of Watir

Bret Pettichord bret at pettichord.com
Sat Sep 18 17:29:36 EDT 2010

On Sat, Sep 18, 2010 at 12:05 PM, Jarmo <jarmo.p at gmail.com> wrote:

> I didn't have the time to look these changes all through, but noticed
> that it's not possible to just merge all these changes into master.
> For example eval_in_spawned_process seemed to be unnecessarily complex
> and is now with a different name in master
> (http://jira.openqa.org/browse/WTR-449). Also, using sockets and
> tcpserver seems to be a overkill. So, in short, i think that all
> changes still needs to be reviewed

I agree that different changes raise different issues and that we need to
consider them each separately. For example, extensions like nextsibling and
prevsibling would require unit tests before I would be comfortable adding
them to master.

I should say that nearly all of this code was written by my collegue Hugh
McGowan, much of it while we were friends, but before I had been hired at
Convio. I am sharing his work with you with his permission.

But i have completely different question - why was there a convio
> branch created in the first place and not merged all useful
> changes/fixes into master taken into account the fact that it has been
> possible to do so? Now there's a 500+ lines of code to be reviewed...

As I just said in reply to Alan, I think this actually has happened. I've
been using this branch as a staging ground. Adding selected code from our
set of monkey patches and then from there merging them into master.

I've tried to use git properly to allow all these merges to be tracked, but
its possible that there are still things I need to do differently to make my
work easier to follow by others. Please let me know if you have suggestions.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rubyforge.org/pipermail/wtr-development/attachments/20100918/c116038b/attachment.html>

More information about the Wtr-development mailing list