[wpkg-users] [Bug 273] Per user install support

[bugzilla-daemon at bugzilla.wpkg.org]

http://bugzilla.wpkg.org/show_bug.cgi?id=273

--- Comment #16 from Keith Jones <k.e.jones at brighton.ac.uk>  ---
Hi Stefan,

(In reply to comment #12)
> The black magic of the host matching syntax is clearly described in the
> ChangeLog, including examples.

 Sorry. Maybe black magic was a bit strong :-)

 Then again (and with RTFM message in my head), I really wouldn't have made the
leap to work out using the %USERNAME% variable in that context so maybe I
should have said "guru magic"? I am still catching up with 1.3's changes so as
far as I'm concerned magic is magic :-)

> The documentation Wiki is filled by the community, so it lacks a bit of
> information.

 If I have time... I'm probably not your favoured candidate for clear and
precise explanations :-)

> ----
> The HKCU uninstall registry check would need to be done for native and WOW64
> registry keys, so one would need to use two registry checks.
> Conclusion: a "useruninstall" and "usergroup" check conditions would be handy.
> Both would be based on existing code, just the root to look for the information
> would be different to the currently implemented "uninstall" and "group" check
> conditions.

 Point taken and I think you're absolutely spot on. Now I've caught up on
things on a more geeky front it all makes much more sense :-)

 Yes. Those are attibutes worth putting in to steer when HKCU searches need to
be made (or not). It inspired thinking in my head so I've applied that to the
idea the "usergroup" and "user" attributes/check conditions as I've just posted
to the group. Please comment on what I've said to Rainer. I'm not sure if I've
described them well but it seems to me to be an opportunity to break down
checks and comparisons on a very simplisitic atomic scale. In the long run,
that is something worth investing in.

 Are my gut feelings about the implied refactoring of the legacy name matching
also worth considering? I really think you could synthesize code to cover the
legacy and concentrate on having a system the matched attributes in a flyweight
model concept. I can deeply understand not wanting to make dramatic and
untested changes to WPKG but it feels like the legacy name matching concept is
holding Rainer's highbrow attribute matching back.

 Please let me know I haven't offended you at least ;-)

> --
> Stefan

Have fun,

Keith

-- 
Configure bugmail: http://bugzilla.wpkg.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.