Column: Test and document behaviour for auto-incrementing IDs

[webmaster128]

I wrote this test to better understand how Column keys are supposed to work in case of gaps and removals. Turns out a key is never re-used. This limited the number of elements that can ever be created to the uint32 range, even if elements are consistently removed.

A typical use case for such a column would be logs. For all the things that happen, a new line is added. Some are cleared later on for cleanup purposes. I.e. there can be cases with significantly more row creations than number of elements over the lifetime of the container.

I don't think there is anything wrong with the above behaviour. It's nice and clean if well understood. But I am wondering if hardcoding a limit of 4 billion row creations is enough.

[uint]

The test looks good!

[uint]

I don't think there is anything wrong with the above behaviour. It's nice and clean if well understood. But I am wondering if hardcoding a limit of 4 billion row creations is enough.

Sure. I'm not very familiar with all the use cases and where the idea of columns came from, so I wasn't sure what I was solving when I implemented it. It's all up for change.

We can make the index generic (with a sane default), or something along those lines?

[webmaster128]

We can make the index generic (with a sane default), or something along those lines?

Yeah, I like. Maybe fix it to u32 in cw-storey which should be a great default.