[Development] [SPAM] How bad QList really is
Matthew Woehlke
mwoehlke.floss at gmail.com
Wed Apr 29 15:41:02 CEST 2020
On 28/04/2020 19.41, Giuseppe D'Angelo via Development wrote:
> Il 28/04/20 21:45, Matthew Woehlke ha scritto:
>>> * QList gets adapted so that its internal array allocates 3 *
>>> sizeof(void*) per element, so that e.g. Q6StringList won't require a
>>> per-item allocation. This would then also avoid allocations for other
>>> datatypes, e.g. QStringView, QImage, maybe QVariant and QColor; but of
>>> course waste a ton of space for the ones which remain small (most of Qt
>>> implicitly shared datatypes).
>> Uh... can't it allocate sizeof(T) if T meets some criteria? IOW, I don't
>> see the second case penalizing smaller types unless the implementation
>> is poorly done.
>
> This way of working is the *key* of QList design. QList is always a
> vector of void* (so it's nice for Qt pimpled types).
...?
> This allows to type erase the entire management of the vector (it's
> always a vector of void*), reducing the amount of template code that
> needs to be instantiated.
Okay, I'm going to *assume* that what you're trying to say, poorly, is
that QList<T> for relocatable T, sizeof(T) ≤ sizeof(void*), is
implemented in terms of casting the bits of T into an intptr_t, which is
further mangled into a void*.
Just because that's how it's implemented *now* doesn't preclude a new,
ABI-breaking version from being implemented as void*[N] for N less than
some small value. It would increase the code overhead *somewhat*, but
not unmanageably so (all allowable instances could still be explicitly
instantiated), and that is probably worth it for the difference in space
efficiency.
--
Matthew
More information about the Development
mailing list