IPv4 parsing sometimes succeeds on invalid inputs #90544
Description
parseIPv4SSE doesn't check that dotmask is valid (it would add noticeable overhead). Easy to find counterexamples by trying random strings consisting of '1' and '.':
SELECT
arrayStringConcat(arrayMap(i -> if((rand((number, i)) % 3) = 0, '.', '1'), range(15))) AS s,
IPv4NumToString(IPv4StringToNumOrDefault(s)) AS ip
FROM numbers(1000)
WHERE ip != '0.0.0.0'
LIMIT 10
Query id: 1249091b-2c7c-40fc-979e-2bf71ce8c01a
┌─s───────────────┬─ip───────────┐
1. │ 1111111.1111111 │ 11.11.1.1 │
2. │ 111111.111.111. │ 1.1.11.1 │
3. │ 1111111111.1.11 │ 1.1.1.11 │
4. │ 11111111.111.11 │ 11.111.1.11 │
5. │ 111.1111.111.11 │ 1.1.1.1 │
6. │ 1111111.1111111 │ 11.11.1.1 │
7. │ 1111111.1111111 │ 11.11.1.1 │
8. │ 111111111111.11 │ 11.11.111.11 │
9. │ 1.111111111.1.1 │ 1.111.11.1 │
10. │ 111.1111111.1.1 │ 111.1.11.1 │
└─────────────────┴──────────────┘Sometimes the result for one row depends on the next row:
:) select IPv4NumToString(IPv4StringToNumOrDefault(x)) from system.one array join ['24', '5.123.234'] as x
SELECT IPv4NumToString(IPv4StringToNumOrDefault(x))
FROM system.one
ARRAY JOIN ['24', '5.123.234'] AS x
Query id: 38b0f4c1-e81c-45fc-b3e9-67519fb582d0
┌─IPv4NumToStr⋯Default(x))─┐
1. │ 245.1.3.234 │
2. │ 0.0.0.0 │
└──────────────────────────┘
2 rows in set. Elapsed: 0.002 sec.
:) select IPv4NumToString(IPv4StringToNumOrDefault(x)) from system.one array join ['24'] as x
SELECT IPv4NumToString(IPv4StringToNumOrDefault(x))
FROM system.one
ARRAY JOIN ['24'] AS x
Query id: cd961c40-c679-4d3d-b3a4-69aa4c041093
┌─IPv4NumToStr⋯Default(x))─┐
1. │ 0.0.0.0 │
└──────────────────────────┘
1 row in set. Elapsed: 0.002 sec. I guess we should make it a little slower and validate the input?
(EDIT: Ok, it turned out easier to fix than expected, didn't have to create an issue.)
Metadata
Metadata
Assignees
Labels
No labels