Re: [GT] patches for GT/Systems.pm and GT/Systems/Generic.pm

["Robert A. Schmied" <ras AT acm.org>]

Weigert, Thomas wrote:
> There is no GT::Systems::Module. The pod incorrectly mentions this,
> while the real line towards the end of Systems.pm should be 
>
> =item C<< GT::Systems->new($args) >>

ahha! or maybe aaarrg!

the word Module is code for <system_module_name>

so if you browse around GT/Systems you will find a number of
modules that have their own new method and so don't need to
inherit the one from GT::Systems.

in each module i looked at the return call to manage_object
sets what manage_object will assign to $key as "".

these appear in the current repository versions. i don't think
i made those changes, but may have discovered that the inheritable
new method from GT::Systems didn't have an argument in the $key
position and just gave it one. i think "" and '' are equivalent
but maybe not.

same observation holds for GT/Signals -- calls have "" in $key
position. don't see one passing a $key.

same observation holds for GT/Indicators, except where the
indicator is passing an actual key.

how most indicators don't make a call to manage_object?

ras


> Th.
>
>
> > -----Original Message-----
> > From: Robert A. Schmied [mailto:ras
AT
acm.org]
> > Sent: Saturday, March 15, 2008 12:20 AM
> > To: devel
AT
geniustrader.org
> > Subject: Re: [GT] patches for GT/Systems.pm and GT/Systems/Generic.pm
> >
> > Weigert, Thomas wrote:
> >
> > > If you pass only for arguments, the remaining two are undefined.
> > > Undefined is not the same as '' or 0, albeit often it works the same
> > > (and the latter gives less errors). It would be good to understand
>
> what
>
> > > should be assigned as key if no key is given.
> > >
> > > On this key and func parameters, I am not sure what they are used
>
> for...
>
> > > So I am nervous. In particular, if it removed warnings, that means
>
> that
>
> > > somewhere key is used in a calculation....
> >
> > thomas
> >
> > can you tell me who/where/what even makes a call to
>
> GT::Systems::Module-
>
> > > new.
> >
> > it happens in a graphic.pl, but i don't see any direct calls to
> > GT::Systems::Module->new (or ::Module->new) any place.
> >
> > maybe i'm mis-using grep and need to resort to etags
> >
> > ras
> >
> >
> >
> > > Th.
> > >
> > >
> > >
> > > > -----Original Message-----
> > > > From: Robert A. Schmied [mailto:ras
AT
acm.org]
> > > > Sent: Friday, March 14, 2008 11:25 PM
> > > > To: devel
AT
geniustrader.org
> > > > Subject: Re: [GT] patches for GT/Systems.pm and
>
> GT/Systems/Generic.pm
>
> > > > Weigert, Thomas wrote:
> > > >
> > > >
> > > > > I don't understand the patch to Systems.pm. What reason do you have
> > >
> > > to
> > >
> > >
> > > > > think that '' is a good parameter for key? Nowhere else is '' being
> > > > > passed as a parameter to $key. Please explain. Th.
> > > >
> > > > i have no log for that change beyond the original email -- all i can
> > >
> > > say
> > >
> > >
> > > > is i think (thought) it works(ed) better with a null value there
>
> than
>
> > > > having no parameter in that position.
> > > >
> > > > i'm thinking the change reduced or eliminated extraneous
>
> uninitialized
>
> > > > or similar warnings?
> > > >
> > > > there was nothing that looked like a key within the body of
> > > > GT::Systems::Module->new, and the orginial return had only
> > > > the initial four arguments.
> > > >
> > > > i don't understand how passing four arguments to something expecting
> > > > six works, nor does it make sense that passing 5 is any better!
> > > >
> > > >
> > > > ras