question around ngx_http_get_flushed_variable()
cnewton at netflix.com
Mon Apr 4 19:44:38 UTC 2022
I'm a little unclear as to when ngx_http_get_flushed_variable() should be
used rather than ngx_http_get_indexed_variable(). The dev guide is a little
sparse, so I'm hoping someone could help.
The relevant flags within the ngx_variable_value_t with their dev guide
descriptions (followed by my interpretation) are:
- valid — The value is valid
The function used to generate the value has indeed been able to create a
valid value. When this is an indexed variable, it is then stored within the
ngx_http_request_t.variables array so that it can be reused for this
request without needing to be re-generated
- not_found — The variable was not found and thus the data and len fields
are irrelevant; this can happen, for example, with variables like
$arg_foo when a corresponding argument was not passed in a request
This could not be generated, and re-attempting to do so won't succeed. As
such (for indexed variables), this is negatively cached within the same
- no_cacheable — Do not cache result
The difference between functions is that the ngx_http_get_indexed_variable()
returns a cached value and ngx_http_get_flushed_variable() flushes the
cache for non-cacheable variables.
That is, no_cacheable is only meaningful when
ngx_http_get_flushed_variable() is called rather than
ngx_http_get_indexed_variable(), when it causes the cached value to be
Are there any guidelines on when ngx_http_get_flushed_variable() should be
used instead of ngx_http_get_indexed_variable()?
It seems like it is 'practically never', but why for example does
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the nginx-devel