Handle very old URLs

[miquelraynal]

On stack overflow I see stale links like:
http://lxr.free-electrons.com/source/drivers/tty/serial/8250/8250_early.c#L19
lxr.free-electrons.com is correctly swapped with elixir.bootlin.com, so the above link jumps to:
https://elixir.bootlin.com/source/drivers/tty/serial/8250/8250_early.c#L19
Instead of:
https://elixir.bootlin.com/linux/latest/source/drivers/tty/serial/8250/8250_early.c#L19
I have no idea if that is easily fixable, but if yes, might be worth creating the link which adds {,linux/latest/}

[fstachura]

Old links can also contain versions, example

I think it would also be good to redirect from links without versions to a selected old version, perhaps whatever was latest when Elixir was released

If we want to handle this in web.py, it will still be necessary to change Apache config a bit - currently, it seems that (even on production) a URL path that starts with source does not match web.py (also see Docker config and the default Apache 404 page on https://elixir.bootlin.com/source).

Handling this fully in Apache is probably possible, but will be hard, see: https://stackoverflow.com/questions/63314539/apache-rewrite-query-param-to-path-url

[tleb]

The lxr.free-electrons.com domain is unreachable now. This is a sysops question.

The big question: do we care?

There are 5.2K results on GitHub, and 137 entries on Google.

[miquelraynal]

In general I don't like when people break my internet without caring. I'd expect at least a viable result, even though the page that was initially targeted is not reached. But if this is too big, then nevermind.