HostDBHash::set_host(): null termination (and lifetime) of string parameter ambiguous.

[ywkaras]

The prototype,

trafficserver/iocore/hostdb/HostDB.cc

Line 172 in bfd5265

HostDBHash::set_host(const char *name, int len)

, implies the string is not null terminated. But it seems to assume the string is null terminated. in line 180.

We also seem to be sending a pointer to a string into deep call stacks and multiple source files, with no clear mechanism to guarantee that it doesn't go stale.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. Marking it stale to flag it for further consideration by the community.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. Marking it stale to flag it for further consideration by the community.

[bryancall]

This cleanup has been addressed. The HostDB restructure in #8953 (first released in 10.0.0) replaced the old HostDBHash::set_host(const char *name, int len) with set_host(TextView name) and made the host_name member a length-aware swoc::TextView. The hash code now uses host_name.data() and host_name.size() instead of assuming null termination, and the hash carries its own backing store (hash_host_name_store[MAXDNAME + 1]), so it no longer relies on a borrowed pointer staying valid. Both the null-termination ambiguity and the string-lifetime concern raised here are resolved, so I am closing this.