25 Nov
2014
25 Nov
'14
7:29 a.m.
On 24/11/14 23:03, Arran Cudbard-Bell wrote:
Almost no one used the rlm_cache module to cache the results of xlat expansions however,
I was.
and the behaviour had some odd side effects when working with attribute references, so it seems more sensible to leave it disabled by default.
If anyone has any objections to the above behaviour changes, feel free to post them to the list.
The return code change seems inappropriate for me in a 3.0.x release. 3.1 or 4.x would make more sense.
Testing of current v3.0.x branch would be appreciated.
Sorry, but a straight-out maintenance release I could test. One with behaviour changes, I have no time for.