name changes for the upcoming 4.0

Poul-Henning Kamp phk at
Sun Jan 15 18:15:16 CET 2012

In message <4BD31D35-EA22-4858-9F81-B357646BBD11 at>, =?iso-8859-1?
Q?Ask_Bj=F8rn_Hansen?= writes:

>Since you were bike-shedding, I'd add my 2 cents that I'd find
>"invalidate" much easier to understand than "ban".   Invalidate
>clearly acts on "stuff from the past", where "ban" seems like it'd
>operate on data in the future.  (Even if that's how it works, the
>effect of a ban is to invalidate old data).

The problem is that a "ban" does operate in the future:  It bans
(matching) cached objects from being served.

"invalidate" implies immediate action, and the entire point is to
avoid just that.

Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

More information about the varnish-dev mailing list