[gnutls-devel] GnuTLS | Datum.c cleanup (!1002)

[Development of GNU's TLS library]

>> Nevertheless, in the general change, are you confident with this version that other uses do not rely on the old behavior?
> How confident can you be ? If someone relies on undocumented side-effects of a function, then ¯\\*(ツ)*/¯.

I'd say that the function itself is the documentation for internal APIs, so I see it as changing a "documented" behavior. If there is code which relied on the behavior of always initializing the output, we may be adding new memory safety issues.

I'd need your help to understand the balance between the risk of introducing a new issue vs the value of the change. Is your objection on the behavior of the function the fact that a `set()` function should not modify its output on failure, would a rename or explicit documentation about its behavior address it?

Nevertheless, I find all the other changes in this patch set, as very useful.

-- 
Reply to this email directly or view it on GitLab: https://gitlab.com/gnutls/gnutls/merge_requests/1002#note_175214754
You're receiving this email because of your account on gitlab.com.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.gnupg.org/pipermail/gnutls-devel/attachments/20190528/67460e42/attachment.html>