Re: Windows registry

[Phil Edwards <pedwards at disaster dot jaj dot com>]

> From law@upchuck.cygnus.com  Wed Jun 14 10:25:05 2000
           ^^^^^^^
Now /there/'s a man who's having a bad day...


Jeffrey A Law <law@cygnus.com>:
>   In message <20000614002542.A7971@cygnus.com>you write:
>   > It sounds to me like this should be a function of the host, not the
>   > target.
> That would be my initial guess too.

Ironic; my first patch operated on the host OS instead of the target OS.
I was /this/ close to submitting it and then thought it over.

I am not familiar enough with the 'doze host(s) to know how deeply the
registry key effects things.


>   > The registry key stuff records where gcc is installed when
>   > it is running, right?
> I don't know.  But it seems plausible, if that is the case then it is
> certainly a function of the host, not the build or target system.
>
> Phil -- can you take care of updating the patch so that it doesn't bother
> with the registry key stuff for the host (or if it is only doing it for the
> host, explain why since my initial reading of the code made me think it was
> operating on the target).

Your initial reading makes you think it's operating on the target because
/my/ patch examines the target.  I am no longer certain when the registry
key itself is used -- during building(host), or during normal operation
after installation(target).

Since we all are now uncertain, perhaps this is one of those things that
should be tabled until Mumit returns and can tell us when the key is used?
Those of us hacking around under Solaris and IRIX can just put up with
"checking if registry support requested... yes" for a while longer.  :-)


Phil