Thread: Should we remove unused code?

Hello

Today I've read this post:

http://blog.2ndquadrant.com/code-coverage/

I think its great and that everyone in this mailing list should
familiarize themselves with it.

Reading the coverage report, I discovered that some code is never
executed. Sometimes its OK (error reporting in out-of-memory cases),
sometimes its not (not enough tests). But there are also cases I'm not
sure about.

For instance there are two flags - HASH_SEGMENT and HASH_FFACTOR that
are in fact never used (see attachment). I wonder whether we should
keep code like this or not.

What do you think?

--
Best regards,
Aleksander Alekseev
http://eax.me/

Aleksander Alekseev <a.alekseev@postgrespro.ru> writes:
> For instance there are two flags - HASH_SEGMENT and HASH_FFACTOR that
> are in fact never used (see attachment). I wonder whether we should
> keep code like this or not. 

> What do you think?

The fact that make check-world doesn't run any code that uses some
feature isn't a great argument for removing it.  Consider third-party
extensions, for starters.
        regards, tom lane

Hello, Tom.

Thanks for your reply.

> The fact that make check-world doesn't run any code that uses some
> feature isn't a great argument for removing it.  Consider third-party
> extensions, for starters.

What is current policy regarding such sort of things? Is everything
that is in .h files considered a public API and should be backward
compatible? In this case - for how long, forever or until next major
release? Or lets say there is some API that is not used neither
internally nor by any package available on PGXN. Can we break API then?

-- 
Best regards,
Aleksander Alekseev
http://eax.me/