Re: [gtk-list] Re: C++ and gtk.

[Guillaume Laurent <glaurent worldnet fr>]

pavel <pavel@klebanov.ne.mediaone.net> writes:

> Fair enough. However, it seems illogical (even symantically incorrect) to
> me that gtk_table_attach() takes a GtkAttachOption and not an int (just
> like it used to be until recently).
> Afterall, EXPAND|SHRINK is an int and not a GtkAttachOption.

No, but it can be used to init a GtkAttachOption.

> Besides, while I don't know the particulars of compiler behavior, it
> seems quite feasible to me that casting an int to a GtkAttachOption
> may result in a 1 or 2 or a 4 (i.e. a GtkAttachOption) and not, say,
> 5 as desired.

No. Quoting Stroustrup's TC++PL 3rd ed (#4.8) : "The range of an
enumeration holds all the enumeration's enumerator values rouded up to
the nearest larger binary power minus 1".

So while

GtkAttachOption a = 5;

is incorrect

GtkAttachOption a = GtkAttachOption(EXPAND|SHRINK);

is correct and guaranteed to give the expected result.

> Any any case, what was the rationale behind changing an int to a
> GtkOption?

I'd assume code readability.

-- 
					Guillaume.
					http://www.worldnet.fr/~glaurent