lp:haproxy

Branch merges

Propose for merging

No branches dependent on this one.

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

784. By Willy Tarreau <w@1wt.eu> on 2008-08-17

[TESTS] test-fsm: 22 regression tests for state machines

22 regression tests for state machines are managed by the new
file tests/test-fsm.cfg. Check it, they are all documented
inside. Most of the bugs introduced during the FSM extraction
have been found with these tests.

783. By Willy Tarreau <w@1wt.eu> on 2008-08-17

[BUG] process_request: HTTP body analysis must return zero if missing data

This missing return and timeout check caused an infinite loop too.

782. By Willy Tarreau <w@1wt.eu> on 2008-08-17

[BUG] process_cli/process_srv: don't call shutdown when already done

A few missing checks of BF_SHUTR and BF_SHUTW caused busy loops upon
some error paths.

781. By Willy Tarreau <w@1wt.eu> on 2008-08-17

[MEDIUM] merge inspect_exp and txn->exp into request buffer

Since we may have several analysers on a buffer, it's more
convenient to have the analyser timeout attached to the
buffer itself.

780. By Willy Tarreau <w@1wt.eu> on 2008-08-17

[BUILD] fix warning in proto_tcp.c with gcc >= 4

signedness issues.

779. By Willy Tarreau <w@1wt.eu> on 2008-08-17

[BUG] regparm is broken on gcc < 3

Gcc < 3 does not consider regparm declarations for function pointers.
This causes big trouble at least with pollers (and with any function
pointer after all). Disable CONFIG_HAP_USE_REGPARM for gcc < 3.

778. By Willy Tarreau <w@1wt.eu> on 2008-08-17

[OPTIM] process_cli/process_srv: reduce the number of tests

We can skip a number of tests by simply checking a few flags,
it saves a few CPU cycles in the fast path.

777. By Willy Tarreau <w@1wt.eu> on 2008-08-17

[MEDIUM] session: move the analysis bit field to the buffer

It makes more sense to store the list of analysers in the buffer
than in the session since they are precisely plugged onto one
buffer.

776. By Willy Tarreau <w@1wt.eu> on 2008-08-17

[MINOR] ensure the termination flags are set by process_xxx

When any processing remains on a buffer, it must be up to the
processing functions to set the termination flags, because they
are the only ones who know about higher levels.

775. By Willy Tarreau <w@1wt.eu> on 2008-08-17

[MEDIUM] centralize buffer timeout checks at the top of process_session

it's more efficient and easier to check all the timeouts at once and
always rely on the buffer flags than to check them everywhere.