<rdar://problem/7274595> Don't downgrade long-lived queries to polling queries if...
authorrlpm@apple.com <rlpm@apple.com@214c2c4a-bf3b-4dcf-9390-e4dd3010487d>
Tue, 20 Oct 2009 16:30:07 +0000 (16:30 +0000)
committerrlpm@apple.com <rlpm@apple.com@214c2c4a-bf3b-4dcf-9390-e4dd3010487d>
Tue, 20 Oct 2009 16:30:07 +0000 (16:30 +0000)
commiteca5a88831573825be6a888222066894be8d3502
tree7a73e2a9149a375d6768969bdb4f530a69cd00c9
parentbafe1f384a4893729f04d1b3ae0d74f085fef7ce
<rdar://problem/7274595> Don't downgrade long-lived queries to polling queries if SSLHandshake fails

Fix bug in previous commit that broke renewing LLQ leases: We were setting ThisQInterval to 0 in
the (currently completely expected) case where the server closes the connection¬†after we initially
setup the LLQ. ¬†Setting ThisQInterval to 0 makes the question inactive, never to be resurrected.

This fix is two-fold:
1) don't handle tcp-close as an error if we've received at least one response packet over this tcp connection
2) ensure we never set ThisQInterval to 0

git-svn-id: http://svn.macosforge.org/repository/mDNSResponder/trunk@6748 214c2c4a-bf3b-4dcf-9390-e4dd3010487d
mDNSCore/uDNS.c