Michael Grunder

Looks like it's not [actually too bad](https://github.com/redis/hiredis/search?q=sds+filename%3A*.h+-filename%3Asds.*&unscoped_q=sds+filename%3A*.h+-filename%3Asds.*). That said, `redisContext->obuf` is an SDS string but we [export it as a char*](https://github.com/redis/hiredis/blob/ea063b7cc895131c6f2955351e8ea5b9d0a48048/hiredis.h#L240). We might be able to avoid breaking changes by...

Hi @sg893052 it's hard to say without code I can use to replicate the crash. That said, v0.14.0 is very old and there have been several possible crashes fixed since...

Hi @sg893052 I'm not sure I totally follow. Obviously corrupting the `sdshdr` could cause a crash, but the question I would have is _how_ is the header becoming corrupted? Have...

The issue with checking for corruption inside of SDS is that it doesn't actually solve your problem and could even make it worse. What you really want to do is...

Closing old issue

There's good stuff in here that we should probably get merged before v1.0.0 I went ahead and rebased your PR against the new pluggable-allocation branch to make eventual merging easier....

Hi @tudor, sorry for the delay. The logic looks good at first glance. We have to be pretty careful about backward compatibility now that we've released v1.0.0, but I will...

Apologies, I will test and merge this weekend.

Hi @YegorZdanovich I need to play around with this a bit more to completely think through the right place to catch `EAGAIN`. The reason it's a bit tricky is that...

Hi, @mtuleika-appcast can you take a look at #1106.