Hacker News Reader

HNHacker News

Top New Best Ask Show Jobs

Comment by d26900 | Hacker News Reader

Parent Full thread

d26900·

That's a cool library! Well done!

I made something similar to it being inspired by the C++ standard library. But it is for C89:

https://github.com/d26900/JamaisVu

I use tombstones for my hash table (unordered_map) though.

Another drawback of my lib: for each hash table entry, I allocate an extra byte for the tombstones.

View on HN

I've been making something kind of similar as well, although it's neither that pretty nor as minimal as I may have aimed for. And unlike the library in the post title, I only did container-style data structures, not other common stuff.

https://github.com/mwahlroos/libcoll-c

Btw I'm not the author, author posted on reddit : https://www.reddit.com/r/C_Programming/comments/lh9irz/commo...

I don't know much about hashmaps but the disadvantage with tombstones is if you add/remove a lot, then you have to rehash your map at some point right?

I really like tiny libs anyway, yours looks good, very readable :)

I am not an expert either, but I would say that you have to rehash it either way if it grows (or shrinks). Tombstones as I understand it, are a way to "mark" deleted entries off of the hash table. Having these deleted entries marked ensures that your hash table doesn't break its "search chain". This is important so that you still can find entries that are put at another index when one index ("bucket") is already occupied.

However, I also don't shrink my table, once it grows. It stays grown. (Another drawback of my lib I suppose).

Thank you for the feedback, I really appreciate it! :)

Shrinking is not that necessary I believe. C++ and Java data structures do not shrink automatically either.

Reply on news.ycombinator.com