On Thu, 6 Mar 2003, Bruce Adams [TEPG Sunbury] wrote:
> >-----Original Message-----
> >From: Max Bowsher [mailto:[EMAIL PROTECTED]
> >Sent: 06 March 2003 14:43
> >To: [EMAIL PROTECTED]
> >Subject: Re: Exception: STATUS_PRIVILEGED_INSTRUCTION occurs
> >before main is executed.
> >
> >> On Thu, 6 Mar 2003, Bruce Adams [TEPG Sunbury] wrote:
> >>> I think it would be a great idea if the post setup phase did some
> >>> basic configure style checking (assuming it doesn't already). i.e.
> >>> Checking if the compiler gcc works.... No
> >>>
> >>> It would have saved me a couple of months of heartache.
> >>> I seems very strange that nobody else spotted it though.
> >>> How exactly could an installation become so corrupt that nothing
> >>> short of a complete download (from a different mirror site)
> >>> fixes it without the underlying package being in error?
> >
> >Somehow, it managed it, since there are lots of us who haven't
> >done a clean install for ages.
> >
> >Igor Pechtchanski wrote:
> >> I'm reasonably sure that if you create such a post-install script, it
> >> will be at least thoughfully considered by the gcc maintainer...
> >
> >Well, I'm not the gcc maintainer, but I'd be surprised if such
> >a script was accepted. Here's why:
> >
> >1) There is no (good) way for a postinstall script to talk to the user.
> >
> Is that a limitation of the current postinstall process? I'm not sure
> what your definition of 'good' entails.
> A report that doesn't explain the problem is better than no report at all.
> For a first cut I'd be happy with
>
> tk_messageBox -title "Installation Error" -message "Problem with C++
> compilation"
>
> Though something more like the below would be preferable for the general
> case:
>
> Gcc post install
> IF post-install-error
> Add error to list
>
> Post post install :-)
> IF console-based-install
> print list of errors found to console
> ELSE // gui-based-install
> print list of errors found to message box instead of "installation ok"
Ah, there's the use of the script that depends on all others that John
Morrison was thinking about... :-D
Alternatively, setup could simply read something like
/var/log/setup.log.errors and print the contents (if any) as part of the
"installation complete" mbox...
> >2) If the package was broken immediately after installation,
> >then wouldn't the maintainer have noticed before releasing it?
> >
> One would hope so. But if they did it was not enough to spot my dodgey
> installation problem. Anyway, does it ever hurt to double check?
> We wouldn't have to run the complete gcc test-suite (though as an option
> it might be nice). Just a smoke-test of some reasonably simple code.
> It wouldn't be such a bad thing if all packages had one, especially
> experimental ones.
>
> Regards,
> Bruce A.
Yep, the above could be a generic mechanism for a postinstall script to
notify setup that something went wrong...
Igor
--
http://cs.nyu.edu/~pechtcha/
|\ _,,,---,,_ [EMAIL PROTECTED]
ZZZzz /,`.-'`' -. ;-;;,_ [EMAIL PROTECTED]
|,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski
'---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow!
Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
-- /usr/games/fortune
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
