Subject: Re: 0.9.3 blocker -- more close vs. exit testing needed
From: David Chart (linux@dchart.demon.co.uk)
Date: Sat Sep 01 2001 - 17:54:56 CDT
On 01 Sep 2001 10:55:53 -0700, Paul Rohr wrote:
> Hi folks,
>
> I just rebuilt and tried out the new implementation of the "don't close on
> exit" feature.
>
> While I think the general spirit of the current approach is admirable --
> thanks, Dom!
I agree.
> Two specific issues:
>
> 1. avoid it
> ------------
> I consider the availability of a working "do it like you did before" setting
> on all platform's preference dialogs to be a MUST for any 0.9.3 binary we
> release.
I agree with this, too.
> 2. test it
> -----------
> Since our current belief is that most people will prefer the new "don't exit
> on close" behavior, we need to make sure that it Just Works, too. Until we
> get to that point, I'd encourage all developers to use AbiWord in that mode.
> If we're not willing to eat our own dog food here, we shouldn't foist it
> off on our users.
I'm already doing this. I'll report back as I find things.
>
> For example, I just checked in a fix for one glitch that I'm shocked that
> nobody complained about:
I only saw it this morning, and I've been busy. Thanks for fixing it.
> Also, I'm still not thrilled with the net effect when you close an existing,
> untitled, empty document. What do *you* think should happen in this case?
> (I think we should definitely exit, but perhaps that's my "die, die" bias.)
> Currently, all that happens is the document number gets bumped up by one,
> which is very very weird.
We should exit in this case, just as we don't open a new window if
someone opens a new document.
> If both can't be done in an acceptable timeframe, then let's switch the
> default preference to hide the feature. That way, folks from the "keep it
> open" camp can get 0.9.3 binaries, edit their prefs files by hand to enable
> the new behavior, and do the necessary testing then.
I have no objection to setting the default the other way for 0.9.3, if
the feature isn't completely ready. We should, IMHO, be defaulting to
the new behaviour by 0.9.4, though.
-- David Chart
This archive was generated by hypermail 2b25 : Sat Sep 01 2001 - 16:56:03 CDT