A proof of concept of a semistable C++ vector container

▲ A proof of concept of a semistable C++ vector container(github.com)

30 points by joaquintides 6 days ago | 9 comments

▲ shin_lao 2 days ago | parent | next [-]

What is the core use case for this structure? Because it seems like a very heavy price to pay just to keep value stable, as opposed to make a copy of that value when you need it.

	▲	wavemode 2 days ago \| parent \| next [-]
		A stable vector generally improves append performance (because you never have to move the data), allows storing data that can't be moved, and allows for appending while iterating. The downside is slightly worse performance for random lookups (since the data is fragmented across multiple segments) and higher memory usage (depending on implementation). This "semistable" vector appears to only do the "allow appending while iterating" part, but still maintains a contiguous buffer by periodically moving its contents.
	▲	eps 2 days ago \| parent \| prev [-]
		It's basically a form of reference-counted data access as I understand it. If the code here operates with a bit of data from some container, the container will ensure that this bit will persist until all references to it are gone even if the bit is removed from the container. Depending on the datamodel this may be handy or even required. Consider some sort of hot-swappable code when both retired and new code versions running in parallel at some point. That sort of thing.

▲ Panzerschrek a day ago | parent | prev | next [-]

The utility of such container seems for me to be questionable. It's likely an error to modify vector while iterators to its elements exist. I see no legit reason to modify a vector with preservation of iterators.

▲ unnah 2 days ago | parent | prev | next [-]

Interesting that they chose not to implement any method to detect whether a given iterator has been invalidated, even though the implementation would be easy. Seems it would be a useful extension, especially since any serious usage of this vector type would already be relying on functionality not provided by the standard vector class.

▲ thechao 2 days ago | parent | prev | next [-]

Neat. Here's what I'd keep: just an epoch saying when the last valid element of the vector is. The iterator just needs a ptr to the vector and the vector's data. It's a constant time lookup, with somewhat heavier iterators. I guess this is something like MSFT's old debug iterators?

▲ omoikane 2 days ago | parent | prev | next [-]

> The same iterator object can't be used concurrently in different threads, even for nominally const operations such as dereferencing (internally, thread-unsafe epoch traversal is triggered)

If I understand correctly:

                         thread safety      random access  stable iterators
                         -------------      -------------  ----------------
    std::list            thread-compatible  no             yes
    std::vector          thread-compatible  yes            no
    std::deque           thread-compatible  yes            no
    semistable::vector   thread-unsafe      yes            yes

I think there are more times when I wanted concurrent reads and (random access OR stable iterators), than when I wanted both random access AND stable iterators but not concurrent reads. I wonder what's the intended application?

▲ quotemstr 2 days ago | parent | prev [-]

Adorable: they've reinvented Emacs markers

	▲	fooker 2 days ago \| parent [-]
		Why do you think there's a re in research?