C0019 unbounded-read¶
Rule: A read to the end of a stream is bounded with take().
| Code | C0019 (stable, never reused) |
| Id | unbounded-read |
| Category | code |
| Source | spirit_level |
| Fix | by hand |
Level per tier¶
| Tier | Level | Reason |
|---|---|---|
| small | off | |
| small-workspace | off | |
| medium | off | |
| big | warn |
input from outside sets the memory a read to the end takes; candidates only, a place to look |
| ai | warn |
input from outside sets the memory a read to the end takes; candidates only, a place to look |
| embedded | off | |
| ui | off | |
| egui | off | |
| units-strict | off | |
| units-paranoid | off |
Parameters¶
| Name | Type | Default | Meaning |
|---|---|---|---|
methods |
list of strings | ["read_to_end", "read_to_string"] |
methods that read until the end |
bounds |
list of strings | ["take"] |
methods in the receiver chain that bound the read |
Why¶
read_to_end and read_to_string read until the other side stops. On a socket, a request body or a file from outside, the other side decides how much memory the process takes. Every input that grows from outside needs a limit.
The lint reports each call whose receiver chain has no take(...). It cannot tell a socket from a file the program wrote itself, so each finding is a candidate: a place to look at, never a failed run.
stream.read_to_end(&mut buf)?; // reported
stream.take(MAX_FRAME_BYTES).read_to_end(&mut buf)?; // passes
How to fix¶
Bound the read with take(LIMIT) and refuse input that reached the limit. Where the source is trusted (the program's own file), waive the line with that reason.
Configure¶
[lints.unbounded-read]
level = "deny"
methods = ["read_to_end", "read_to_string"]
bounds = ["take"]
reason = "why this project wants it at this level"
Allow one finding in the code with // spirit-level: allow(unbounded-read) <reason> (waivers); the code works there too, as in [lints.C0019].