diff options
| author |  Alistair Veitch <aveitch@google.com> | 2016-01-20 09:17:11 -0800 |
|---|---|---|
| committer | Alistair Veitch <aveitch@google.com> | 2016-01-20 09:17:11 -0800 |
| commit | 8e5b21fd9faf0b1ba2f74e88fe7e59bc3d5ed299 (patch) | |
| tree | 56e8a033b203f2ffa79e68348bd525d12990e8bb /src/core | |
| parent | 4bbdb825677bde8e53cbab58ad4193c505dc9a50 (diff) | |
comment updates
Diffstat (limited to 'src/core')
| -rw-r--r-- | src/core/census/tag_set.c | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/src/core/census/tag_set.c b/src/core/census/tag_set.c index 718cfe3b25..3b5d345910 100644 --- a/src/core/census/tag_set.c +++ b/src/core/census/tag_set.c @@ -56,7 +56,8 @@ // are to aid in efficient parsing and the ability to directly return key // strings. This is more important than saving a single byte/tag on the wire. // * The wire encoding uses only single byte values. This eliminates the need -// to handle endian-ness conversions. +// to handle endian-ness conversions. It also means there is a hard upper +// limit of 255 for both CENSUS_MAX_TAG_KV_LEN and CENSUS_MAX_PROPAGATED_TAGS. // * Keep all tag information (keys/values/flags) in a single memory buffer, // that can be directly copied to the wire. This makes iteration by index // somewhat less efficient. If it becomes a problem, we could consider |