Subject | Re: [IBO] Enhancement Suggestion: Client Libraries |
---|---|
Author | Geoff Worboys |
Post date | 2005-07-15T08:20:27Z |
G>> I have never liked trying to use the order of the uses items.
G>> It is so easily broken and far from explicit because most of
G>> the time the order can be ignored.
doing something a little bit different or tricky. But if there
is not a good reason it should be avoided - all IMHO of course,
perhaps because I have managed to get through many years of
Pascal programming without bothering about the order. :-)
G>> Its funny actually, I was going to "pick on you" for trying to
G>> assign a value to a constant (IB_GDS32) but checked first and
G>> discovered it is a "var" - not what I expected in IB_CONSTANTS,
G>> I did not even think about trying to assign to it. Oh how easy
G>> it is to miss the obvious!
the alternative IB_session units were all about.
I had effectively ignored those from the beginning because I
have not really needed them and because I never really saw it
as a viable solution to the problem - again my prejudice
against ordered uses clauses. It had not actually dawned on
me that it must be editing what I thought was a "constant".
Well duh! - dont I feel silly.
Something to this effect is what I have in the C++ library that
I use, so I did think about it while I was looking at the IBO
changes I described - because something along those lines would
obviously be a better solution. I decided it was too much work
for something I did not need (at the moment).
--
Geoff Worboys
Telesis Computing
G>> It is so easily broken and far from explicit because most of
G>> the time the order can be ignored.
> well it is explicit, and it cannot be ignored. Many tools thatSure, there are times when there is just no choice - some code
> need to be loaded before everything else in an application
> require to be put at the beginning of the uses clause.
> It's not really trying to use the order, just the first item.
> WRT to the "never liked" part, I'm not going to argue of
> course. :-)
doing something a little bit different or tricky. But if there
is not a good reason it should be avoided - all IMHO of course,
perhaps because I have managed to get through many years of
Pascal programming without bothering about the order. :-)
G>> Its funny actually, I was going to "pick on you" for trying to
G>> assign a value to a constant (IB_GDS32) but checked first and
G>> discovered it is a "var" - not what I expected in IB_CONSTANTS,
G>> I did not even think about trying to assign to it. Oh how easy
G>> it is to miss the obvious!
> that's what people have been doing for years.Of course... the light finally dawns. That is effectively what
the alternative IB_session units were all about.
I had effectively ignored those from the beginning because I
have not really needed them and because I never really saw it
as a viable solution to the problem - again my prejudice
against ordered uses clauses. It had not actually dawned on
me that it must be editing what I thought was a "constant".
Well duh! - dont I feel silly.
> Agreed. You might know it or not (probably not) but I onceNo I did not know (or do not remember) that.
> did quite a bit of work to allow IBO to load multiple client
> libraries at once, one for each session component.
Something to this effect is what I have in the C++ library that
I use, so I did think about it while I was looking at the IBO
changes I described - because something along those lines would
obviously be a better solution. I decided it was too much work
for something I did not need (at the moment).
--
Geoff Worboys
Telesis Computing