Downloads graph at bottom of page overwhelmed by alpha versions of crate

[tspiteri]

Current Behavior

For example for my fixed crate I have seven 2.0.0-alpha versions, (six of which are yanked). The graph at the bottom of the page just shows these, so the graphs for the stable 1.* versions are not visible.

It seems that currently the graph just shows the first five versions using semver ordering.

Expected Behavior

I didn't expect alpha/prerelease/beta versions, most of which are yanked, to completely overwhelm the graph so that no stable versions are visible. Some options that would solve this, though I don't know if any of them is even preferable to the current semver order:

• sort by semver, but do not include yanked versions
• sort by semver, but include only the highest for any x.x.x version number, so that for example from 2.0.0-alpha.1, 2.0.0-alpha.2, 2.0.0-beta.1, 2.0.0-rc.1, only one line for 2.0.0-rc.1 would be shown
• sort by release date

Steps To Reproduce

No response

Environment

No response

Anything else?

No response

[Turbo87]

I've converted this to a discussion because the feature is working as designed, which is to "show the latest X versions". I agree that this is less useful than it could be in this case, but I'd rather treat this as an improvement than an actual bug.

[cuviper]

There's a similar issue with the graph on indexmap even with just regular releases:

You can see that the 2.* uptake is decent, but there's still a whole lot of "other" that I expect are 1.* users. In this case, I think I'd rather see "latest, plus the next X most popular", rather than just the latest X. Maybe there could be a dropdown like the stacked/unstacked choice to select different sorting modes?

[cuviper]

I like that!

I just noticed in your previous "downloads" graph, the total stack is higher than in the semver view, which shouldn't happen -- I wonder if perhaps some listed versions were double-counted in "other". Your new 7-day graph doesn't have that problem.

[eth3lbert]

Oops, thanks for pointing it out. It must have been something wrong before. I just updated the graph.

[cuviper]

Great, with that update, I like both versions of the download graphs. If we have to pick just one, I guess 90 makes more sense in general, but the 7 feels more current. They'll probably be about the same for a lot of crates though.

[eth3lbert]

Yep, 7 days or any period far less than 90 days is more like recent trends.

[EdJoPaTo]

On the fixed crate the prerelease versions are still polluting the semver graph making it basically useless. What about a toggle to only show non-prerelease, non-yanked semver versions?