Case-Insensitive Response Compression May Cause Problems With Mixed-Case Data and Non-Conforming Clients

Mark Andrews marka at
Mon Feb 10 21:08:14 UTC 2014

In message <alpine.LSU.2.00.1402101617110.12813 at>, Tony Finch writes:
> Paul Vixie <paul at> wrote:
> > Tony Finch wrote:
> > > ...
> > >
> > > If you later create positive cache entries under these labels (by asking
> > > for SRV) you get the first answer back with canonical case (because it is
> > > copied from the authoritative answer) but subsequent answers from the
> > > cache use the same case as the negative cache entry.
> >
> > i recommend that this be solved with a 2181-bis that adds label
> > credibility to its section 5.4.1.
> Another thing I wonder about is what happens on authority servers with
> zones like below. But I gather Mark already has some possible solutions in
> mind.
> silly.example  A
> But making this much effort to preserve case seems contrary to RFC 4343.

The thing is people do actually complain when the case isn't that
which is entered.  When we went to multiple records per messages
for AXFR we got compliants.  Doing case sensitive compression on
AXFR fixed that set of complaints.

The last change preserves the case of domains names in the rdata
for those records which have compressable rdata fields like the
case is preserved in all the records which don't have compressable
rdata fields

Preserving the owner name is the next step.

The case of the question in the reply is still preserved (matches
the query) which allows for 0x20 to work.

This also makes the answer look sane regardless of the camel case
on the question when talking to a authority server.  For a cache
it depends on what the authority server does.

If I wanted "" as the owner name for a
record the additional work will allow it to be preserved.  There
are cases where the case matters for athetics.


> Tony.
> -- 
> f.anthony.n.finch  <dot at>
> Forties, Cromarty: East, veering southeast, 4 or 5, occasionally 6 at first.
> Rough, becoming slight or moderate. Showers, rain at first. Moderate or good,
> occasionally poor at first.
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: marka at

More information about the bind-workers mailing list