[
https://issues.apache.org/jira/browse/TS-2181?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Bryan Call updated TS-2181:
---------------------------
Assignee: John Rushford
> Add an identifier ID to remap rules, and corresponding log tag and plugin APIs
> ------------------------------------------------------------------------------
>
> Key: TS-2181
> URL: https://issues.apache.org/jira/browse/TS-2181
> Project: Traffic Server
> Issue Type: New Feature
> Components: Logging, Remap API
> Reporter: Leif Hedstrom
> Assignee: John Rushford
> Fix For: sometime
>
>
> It would be useful to be able to have an identifier (not necessarily unique)
> in remap rules. Something like e.g.
> {code}
> map http://example.com http://a.example.com @id="Rule 1"
> {code}
> Where the ID can be any freeform string up to say 32 characters (fixed size
> for efficient APIs). In addition, we would also add a new log tag, RID or
> some such, which can be used in custom log formats.
> Finally, since not all remapping happens through remap.config, we'd want to
> add two new APIs as well (this is where the fixed size string comes to play,
> to avoid allocations):
> {code}
> const char* TSRemapIdentifierGet();
> {code}
> and
> {code}
> TSReturnCode TSRemapIdentifierSet(const char* str);
> {code}
> The string is guaranteed to be NULL terminated (for the Get()'er to be safe
> and easy to use).
> Question: Does it makes sense to instead use a (char*, int) to pass along ? I
> think not personally, and we have precedence in both (in marshaled buffers,
> this makes more sense).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
