[VETO] old ws -- packages, define-behavior
ben at 666.com
Thu Nov 11 21:27:32 EST 2004
> As an old time UNIX user I disagree in part. IMHO the
> defaults should follow the conventions of the underlying
> toolkit. Pending-del has no place when running under Xt
> because no other Xt app has that behaviour. It is, however,
> quite reasonable when running under natively Windows.
[a] it seems too random to have a behavior such as this enabled in some
cases and not others
[b] the app toolkit is an implementation thing that has little relevance for
users and is not something we should conditionalize on.
[c] I'm positive the Motif style guide says pending-delete behavior should
[d] Netscape, e.g., is a Motif app with pending-del on.
> How about imposing a checkin delay instead? Ben's just
> dumped 70,000+ lines of patches onto us in the last two weeks
> and to start checking them a few days after the initial post
> with comments like:
> No one responded to my message of a week ago asking
> people to test this.
> is just plain annoying. We probably haven't responded
> because we busy reviewing and trying to work out what the
> heck he's done!
> Ben, I think the stuff that you contribute is wonderful and
> hope that you continue to contribute but if you're going to
> keep posting mega-patches then wait a month before
> committing. They may be old workspaces to you but we've only
> just seen them need time to analyse them and test them on
> non-Windows boxes. Isn't that the purpose of review-before-commit?
A month is an awfully long time. Please name one other project with such a
long delay. After a month I'll probably have forgotten some of what I did,
in any case, and bug-fixing will be much harder.
[In fact, how many other projects have review-then-commit at all? I
remember a statement awhile ago that the majority go the other way]
More information about the XEmacs-Beta