[fpc-pascal] Multi-scope helpers draft

[Sven Barth]

Am Mo., 26. Nov. 2018, 03:12 hat Ryan Joseph <ryan at thealchemistguild.com>
geschrieben:

>
>
> > On Nov 25, 2018, at 11:03 PM, Sven Barth via fpc-pascal <
> fpc-pascal at lists.freepascal.org> wrote:
> >
> > It's a good thing that you saw the errors in your first design and
> rectified those in the third commit. ;) That commit however contains
> unnecessary noise (new line changes? space changes?), so when reworking the
> commits for a patch please try to get rid of these.
> >
>
> Yeah I don’t know what Git thinks changed so I just ignored it. Maybe my
> text editor (Sublime Text) changed the indention from tabs to spaces or
> something.
>
> It thinks these 2 lines are different for some reason (copied from git):
>
> -        if not (ocf_check_non_overloadable in ocf) and not
> isunaryoperatoroverloadable(t.nodetype,inlinenumber,ld) then
> +        if not (ocf_check_non_overloadable in ocf) and not
> isunaryoperatoroverloadable(t.nodetype,inlinenumber,ld) then
>

You could use an editor that shows non printable characters and check what
got changed there.


> > One thing that bothers me is the "lastonly" parameter. Why did you add
> that? In the two locations you added them you'd now have a problem if
> multiple helpers are in scope, but the last one does not contain the
> requested symbol. In my opinion that parameter is not needed at all.
>
> I was trying to reduce the exposure of my changes to the rest of the code
> base but I’ll remove it if you think it’s safe.
>

That's were running the testsuite comes in. Both for existing tests to
discover regressions and new tests to ensure that everything works as
expected.
In this case you'd write a test for the situation I described, run that an
notice that it won't even compile. Thus you'd learn that the "lastonly"
parameter made things worse.

A bit more information for testing FPC can be found here if you don't have
seen that already: http://wiki.freepascal.org/Testing_FPC

Regards,
Sven

>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freepascal.org/pipermail/fpc-pascal/attachments/20181126/065dc220/attachment.html>