Re: GHashTable and const
- From: Mark Mielke <mark mark mielke cc>
- To: BJörn Lindqvist <bjourne gmail com>
- Cc: gtk-devel-list gnome org
- Subject: Re: GHashTable and const
- Date: Fri, 04 Jul 2008 22:42:32 -0400
BJörn Lindqvist wrote:
Yes, i'm aware of that. Const variables just seem to grow like
mushrooms around const-accepting functions and those inevitably cause
trouble. Inside the function with a const parameter, that parameter
also only has to be passed to const parameter functions. That is
another case where warnings has to be fixed with irritating casts.
Why would a function taking a const parameter be calling a function with
the parameter that isn't defined to take a const?
It may seem like a mushroom to you - to me it seems like a mine field
*not* to use const. All compile time checking has this issue - and all
compile time checks provide value. Which is worse? To realize that your
supposedly const function is calling a non-const and having the compiler
tell you about it? Or to not realize that your supposedly read-only
function is doing modifications?
I prefer to use const, because I value what it provides. That it's
imperfect is not a reason to avoid it.
The main issue with glib, though, is history. Most of the programming
world is happy with using const, so I don't buy that the glib developers
are smarter in their choice not to. The choice was wrong. But, it's
water under the bridge. Life goes on. The question is now whether it's
worth it to change - and I suspect the answer is no.
Cheers,
mark
--
Mark Mielke <mark mielke cc>
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]