String Library Functions

strlen/strcpy/strcmp and friends, and their sharp edges.

StatusCollected

Six pieces on <string.h> — the functions I'd used for years without ever reading the fine print on what happens when things go wrong.

Function	Does	Sharp edge
`strlen(s)`	counts bytes up to (not including) `\0`	no bounds check — reads forever if `\0` is missing
`strcpy(dst, src)`	copies including the terminator	no length limit — classic buffer overflow if `dst` is too small
`strncpy(dst, src, n)`	copies at most n bytes	doesn't null-terminate `dst` if `src` is n bytes or longer — has to be done manually
`strcmp(a, b)`	lexicographic compare	returns 0 for equal — easy to misread as a boolean "true means equal"
`strcat(dst, src)`	appends src onto dst	same overflow risk as strcpy, plus has to walk dst first to find where to start

strncpy is the one that actually surprised me — it looks like the "safe" bounded version of strcpy, but if the source is exactly n bytes or longer it silently leaves the destination without a terminator at all, which just moves the bug somewhere less obvious instead of fixing it.

Notes from readers

Comments — via GitHub