[Gambas-user] variable declaration everywhere ???

[Jussi Lahtinen]

I suggest you try how long the declaration takes (I don't think it is
significant). But using global/public variables for loops is bad idea.
Just think what happens when you call function inside of some loop and they
both happen to use the same variable (or the function call something else
which do it).
This will create bug that can be really hard to debug.


Jussi

On Wed, Dec 12, 2018 at 11:36 AM Rolf-Werner Eilert <rwe-sse at osnanet.de>
wrote:

> Doesn't that slow down a bit?
>
> I know, on today's CPUs it may not be an instant drawback, but just
> consider what happens if such a function is called hundreds or thousand
> of times whithin a larger scope. Then redefining the variable will
> quickly grow exponentially.
>
> Just image TableView_Data, which rebuilds the whole table whenever the
> user moves something on the screen. Each time the Data event is called
> for every single cell of the table. In cases where I need such variables
> there, I define them globally and just re-use them.
>
> In my code, I define counting variables in the function they are used
> in. But object variables which are used very frequently I define once on
> module level and reuse them rather than defining them over and over
> again. It's this feeling that it takes a lot of pushing values forth and
> back, storing, deleting, garbage collection stuff... I just feel more
> secure knowing that a variable is redefined 20 times than 20,000 times
> during runtime.
>
> Regards
> Rolf
>
>
>
>
> Am 12.12.18 um 00:38 schrieb Jussi Lahtinen:
> > Every time you redefine it, it is redefined. And depending on
> > implementation the variable could be valid only inside of the loop. If
> > not, then it is what it is defined.
> >
> > For idx as integer = 0 to things.max
> >     dosomething(thing[idx])
> > next
> >
> > Above idx is integer.
> >
> > ...
> > For idx = 0 to others.max
> >     dosomthingelse(others[idx])
> > next
> >
> > Above idx is either undeclared or integer, depending on implementation.
> >
> > ...
> > For each idx as float in floatarray
> >     dowhatever(idx)
> >
> > Above idx is float.
> >
> >
> > Jussi
> >
> >
> > On Wed, Dec 12, 2018 at 1:31 AM Bruce <adamnt42 at gmail.com
> > <mailto:adamnt42 at gmail.com>> wrote:
> >
> >     Well, I haven't tried it yet but a few things pop into my mind.
> >
> >     Consider a method that has lots of use of "throw away"
> >     indexers/iterators,
> >     For idx as integer = 0 to things.max
> >         dosomething(thing[idx])
> >     next
> >     ...
> >     For idx = 0 to others.max
> >         dosomthingelse(others[idx])
> >     next
> >     ...
> >     For each idx as float in floatarray
> >         dowhatever(idx)
> >
> >     etc
> >     Is it going to redefine idx each time?
> >
> >     b
> >
> >     On 12/12/18 9:46 am, Benoît Minisini wrote:
> >      > Le 12/12/2018 à 00:11, Jussi Lahtinen a écrit :
> >      >> Why would it be harder to debug?
> >      >> David gave great example of it usage (pretty much only case
> where I
> >      >> use it) and I think it only makes code more readable and less
> scope
> >      >> for the variable.
> >      >>
> >      >>
> >      >> Jussi
> >      >>
> >      >
> >      > Because as soon as variables can be declared anywhere, people
> >     will start
> >      > to use that for writing ugly code. :-)
> >      >
> >
> >     ----[ Gambas mailing-list is hosted by https://www.hostsharing.net
> ]----
> >
> >
> >
> > ----[ Gambas mailing-list is hosted by https://www.hostsharing.net ]----
> >
>
>
> ----[ Gambas mailing-list is hosted by https://www.hostsharing.net ]----
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.gambas-basic.org/pipermail/user/attachments/20181212/dbbe886d/attachment.html>