Hi,
HAProxy 3.5-dev13 was released on 2026/11/18. It added 108 new commits
after version 3.5-dev12.
Among those, 65 were bug fixes, many of them coming from pending fixes
still under review, so they keep coming.
Several major bugs were fixed, with 3 or 4 CVEs coming (sorry, I don't
remember now to be honest, it's super hard to context-switch between all
these reports). One affects H2 and leaves exposed H2 servers at risk of
crashing. The other ones are only known to happen with a tarpit action
but can cause crashes as well with H2 and H3. Thus if you're using these
versions on 3.5-dev, you must upgrade to dev13.
For the rest of the issues, Christopher is preparing stable versions so
I'll defer it to his long and detailed changelog to explain what was fixed
in the various branches.
The flood of AI-assisted "findings" tends to convince the reader of the
importance of the findings and to urgently merge a fix (half of which
would only fix the symptom, not the root cause) and these generally need
to be reworked. And sometimes the fix causes more trouble than the
original bug it was supposed to address. I expect that in the next few
months we'll see some reverts. We're trying to remain extremely careful
about this but it's never 100% certain that a fix for a possible crash
won't have any visible side effect. As such, I'm really imploring you
all, if you ever seem to face a functional regression or even a visible
performance regression between successive versions of the same branch,
please try to narrow it to two consecutive versions or even better, to a
single commit, and report it, don't be shy! You may very well be the only
victim so don't count on others to report your issues, do it and we'll
address them. For now in a long series we've only seen a single revert
and I don't even think the bug was reported by AI, so I'm very satisfied
about how careful the team is but keep in mind that with this volume,
mistakes will eventually happen.
Now for the interesting stuff! Well, we're close to a release so there is
much less of it by now, of course (and that's reassuring):
- update to the libslz compression library. It can now preserve a small
context (4kB) between calls when dealing with the "text/event-stream"
content-type (SSE). This allows AI/API response traffic to be divided
by ~25 instead of being slightly inflated. I might perform a last
cosmetic update to the lib if I release it before haproxy (pretty
possible, week-ends are rainy and great for getting work done). Only
doc, measurements and a tag are left to be done, nothing that would
affect the code.
- a better fairness and respect of server weights in uneven server farms
with the random algorithm (suggestion from Alex).
- I thought I had already merged it a month ago but I found it waiting
for me in one of my local branches: "tune.h2.fe.enabled". This allows
to forcefully disable H2 on all frontends (alpn, binds, h1->h2 upgrade
etc). Do not use it, unless being instructed to do so! The goal is to
have a general emergency kill switch next time something really nasty
is reported on H2, since I don't see why this would stop given that the
protocol is particularly complex and has a few awkward architectural
principles. See this as a future safety measure. An "h2_enabled" sample
fetch was also added to report a boolean saying whether or not h2 is
supported+enabled or not (e.g. to decide whether or not to perform some
checks or redirects or to advertise alt-svc etc).
- some minor stuff around idle connection management on the SPOP mux.
- h3 will now reject malformed push frames
- "text/event-stream" wouldn't be matched to enable zero-delay in case
it was followed by a parameter (e.g. "; charset=utf8"). This is now
fixed.
- a "tr" converter, that does exactly the same as the unix "tr" utility,
it can replace one character class by another one, delete characters of
a class or outside a class. The main use case I'm having for this is to
count the number of '%' or '[' in query strings, indicating a likeliness
of attacks. Just use tr("-dc","%["),length and it should return the
number of these characters.
- a few more regression tests; we're now at 439 VTC files, totalizing
around 7500 control points.
- some minor cleanups, doc and CI updates.
We'll see how it goes over the following week. If everything goes well, we
could release in the middle of next week, Wednesday or Thursday. If we have
too many tricky fixes or problematic reports, we could emit a dev14 and
postpone the release by one week. Finding reports are still being reviewed
and addressed (even while I'm writing this). We will not think about trying
to get to the end of this in order to release, and the remaining ones are
not even specific to 3.5 so that would be stupid. What will matter is your
reports. If you find something that needs to be done differently (new config
stuff that's really not handy, etc), please report it so we can address it
before the release, otherwise the release will make it a feature.
Quite frankly, this version looks good and has received maybe 10 times
more review and code audits than any other one before. From our side it
can sometimes look depressing to see the number of reports in queue, but
we need to put into perspective that they are for theoretical bugs that
nobody has ever faced, and that their amount is limited anyway. So I'm
not really concerned about this.
Thus please test, don't worry that it's numbered 13, or use this as an
excuse for forcing us to emit a 14, and let's get prepared for a release
soon.
Please find the usual URLs below :
Site index : https://www.haproxy.org/
Documentation : https://docs.haproxy.org/
Wiki : https://github.com/haproxy/wiki/wiki
Discourse : https://discourse.haproxy.org/
Slack channel : https://slack.haproxy.org/
Issue tracker : https://github.com/haproxy/haproxy/issues
Q&A from devs : https://github.com/orgs/haproxy/discussions
Sources : https://www.haproxy.org/download/3.5/src/
Git repository : https://git.haproxy.org/git/haproxy.git/
Git Web browsing : https://git.haproxy.org/?p=haproxy.git
Changelog : https://www.haproxy.org/download/3.5/src/CHANGELOG
Dataplane API :
https://github.com/haproxytech/dataplaneapi/releases/latest
OpenTelemetry : https://github.com/haproxytech/haproxy-opentelemetry
Pending bugs : https://www.haproxy.org/l/pending-bugs
Reviewed bugs : https://www.haproxy.org/l/reviewed-bugs
Code reports : https://www.haproxy.org/l/code-reports
Latest builds : https://www.haproxy.org/l/dev-packages
Willy
---
Complete changelog :
Aleksandar Lazic (1):
DOC: config: explain the effect of the weights on random's draws
Amaury Denoyelle (13):
BUG/MINOR: quic: fix ACK_ECN parsing after optional ack ranges
BUG/MINOR: qmux: remove invalid BUG_ON() on empty record
BUG/MINOR: init: fix missing error on no listener on config check
BUG/MINOR: quic: remove CID on BE conn alloc failure
BUG/MINOR: h3: prevent missing request FIN on empty trailers
BUG/MEDIUM: mux_quic: fix Rx flow control increment on RESET_STREAM
BUG/MINOR: h3: skip content-length check on backend bodyless responses
BUG/MINOR: h3: reject CANCEL_PUSH on backend side
MINOR: h3: reject malformed push frames
BUG/MINOR: hq-interop: skip zero-copy with large buffers
BUG/MINOR: hq-interop: ignore uni streams in HTTP/0.9
BUG/MINOR: hq-interop: drop data after request
Revert "BUG/MINOR: h3: fix missing EOM after trailers without FIN"
Christopher Faulet (25):
BUG/MINOR: tcpcheck: fix out-of-bounds reads on short MySQL error packets
BUG/MEDIUM: mux-h2: put BE conn back in idle list after deferred shut
BUG/MEDIUM: mux-spop: do not enforce the timeout on connections with
streams
MINOR: mux-spop: ignore conn->owner when deciding if a connection is dead
BUG/MEDIUM: mux-spop: refresh the connection timeout before parking it
idle
MINOR: mux-spop: don't add a closing connection to an idle list on detach
BUG/MEDIUM: mux-fcgi: put BE conn back in idle list after deferred shut
BUG/MAJOR: cache: do not mix variants varying on different headers
BUG/MEDIUM: cache: fix "show cache" crash/skip with several caches
BUG/MEDIUM: cache: do not create a second applet when two cache-use rules
hit
BUG/MEDIUM: cache: forbid cache rules in defaults sections
BUG/MEDIUM: filters: order the implicit filters around the explicit ones
BUG/MINOR: cache: do not refuse to store responses with "Age: 0"
BUG/MINOR: cache: do not remove the Age header from the forwarded response
BUG/MINOR: cache: recognize null qvalues with decimals in Accept-Encoding
BUG/MINOR: cache: do not eat the request while waiting for a buffer
BUG/MINOR: cache: reduce the max-age lifetime by the response's Age
BUG/MINOR: cache: compare the Age to the heuristic lifetime when none is
given
BUG/MINOR: cache: do not ignore null lifetimes not written as "max-age=0"
REGTESTS: cache: cover the object lifetime with a single script
BUG/MEDIUM: mux-h1: Report EOS when a demux error is handled by the app
layer
BUG/MINOR: cache: skip already resolved filters when resolving caches
REGTESTS: http-messaging: Fix abortonclose issue with connect_tunnels
REGTESTS: http-messaging: Properly handle WINDOW_UPDATEs with
connect_tunnels
BUG/MEDIUM: stream: keep channel_auto_close() on the request error path
Frederic Lecaille (9):
BUG/MEDIUM: quic: do not keep 0-RTT packets which cannot be decrypted
BUG/MEDIUM: quic-be: keep the source port range of a closing backend
connection
BUG/MINOR: quic-be: do not register the connection before the quic_conn
is initialized
BUG/MINOR: quic: initialize the members read by quic_conn_release() first
BUG/MEDIUM: peers: bound stick-table update messages
BUG/MEDIUM: peers: bound stick-table switch messages
BUG/MINOR: peers: bound stick-table acknowledge messages
CLEANUP: peers: remove the unused intencode()
BUG/MINOR: quic/ssl: report SSL_CTX allocation failure
Gihyeon Ryu (6):
BUG/MINOR: sample: fix host_only and port_only on empty input
BUG/MINOR: sample: reject a \0 byte in raw string fetches
BUG/MINOR: http-ana: don't emit a network-path Location for a "/"
redirect prefix
BUG/MINOR: log: don't let an LF in a forwarded message create new events
BUG/MEDIUM: sample: parse json_query integers within the token length
BUG/MINOR: mjson: bound the exponent in mystrtod()
Olivier Houchard (20):
BUG/MEDIUM: fd: Mark a fd as dead on early close
BUG/MEDIUM: stick-tables: compute the shard of string keys on the stored
key
BUG/MINOR: debug: Do not allocate context when we don't need one
BUG/MINOR: http/hardening: free the new profile on strdup() failure
BUG/MEDIUM: stick-tables/cli: "set table" without key must not clear the
table
BUG/MEDIUM: peers: make the soft-stop "dont_stop" state per peers section
BUG/MEDIUM: htx: never merge the next DATA block into the referenced one
on defrag
BUG/MEDIUM: flt-decomp: don't expose EOM before the end of decompression
REGTESTS: decompression: test a large decompressed Content-Length response
BUG/MEDIUM: mux-h2: don't abort a tunnel attempt on interim responses
REGTESTS: http-messaging: check that 1xx don't abort H2 backend tunnels
REGTESTS: decompression: don't use yes(1) in http_large_cl
BUG/MEDIUM: dns: release the session slot of expired TCP queries
REGTESTS: http-messaging: send h1_big_chunk_size_line request at once
BUG/MINOR: checks: only trigger a DNS resolution with the resolvers lock
BUG/MINOR: dns: detect the server's shutdown on a partial TCP response
BUG/MINOR: map: split "add map/acl" payload keys on line breaks
MINOR: peers: move message length constants to peers-t.h
BUG/MEDIUM: stick-table: fix the peers overhead in proto_size
BUG/MEDIUM: peers: trim received string keys to fit in a buffer
William Lallemand (6):
CI: github: include WolfSSL version in job name
CI: github: test WolfSSL with and without debugging
CI: github: restrict WolfSSL sanitizer flags to debug jobs
BUG/MINOR: ssl: exclude wolfSSL from custom extension support
SCRIPTS: vtest: allow overriding the Git repository URL
CI: github: use haproxytech/VTest2 for regression tests
Willy Tarreau (28):
OPTIM: lb-chash: don't look for more servers when there is only one
OPTIM: lb-chash: limit the scan to few consecutive nodes
MEDIUM: backend: avoid picking the same server twice in random draws
IMPORT: slz: update slz to 1.4-dev5-24
MINOR: compression: pass COMP_CT_EVENT to the algorithms for server-sent
events
MINOR: compression: add tune.comp.res.ring to disable the rings of
responses
MEDIUM: compression/slz: use the event content type for server-sent events
DEV: sse-latency: add a parameter after the content-type
MINOR: http-ana: also match text/event-stream with trailing parameters
DEV: h2: add h2dec, a raw HTTP/2 frame decoder
DEV: h2: add mkhdr.c, a C port of mkhdr.sh
DEV: h2: mkhdr: add -n to repeat the emitted frames
DEV: h2: mkhdr: support stream ID ranges with -i
DEV: h2: mkhdr: support "--" as a delimiter between frame descriptions
MINOR: ssl: add alpn_remove_token() to drop a protocol from an ALPN string
MINOR: global: add tune.options2 to host new GTUNE2_* flags
MINOR: global: add tune.h2.fe.enabled to disable HTTP/2 on frontends
MINOR: sample: add h2_enabled to report whether H2 may be used on
frontends
REGTEST: ssl: test tune.h2.fe.enabled and h2_enabled on two instances
IMPORT: slz-prv: work around lobotomized compilers' fallthrough
diagnostics
MINOR: sample: add smp_prepare_out() to prepare an output buffer
OPTIM: sample: use smp_prepare_out() in the lower and upper converters
MINOR: sample: add a "tr" converter
BUG/MAJOR: buf: make the shift distance check less strict in b_move()
BUG/MAJOR: mux-h2: fail on impossible data transfers in rcv_buf()
BUG/MAJOR: h3: do not panic on HTX_FL_PARSING_ERROR after htx_xfer()
BUG/MAJOR: mux-quic: fail on impossible data transfers in rcv_buf()
Revert "MINOR: http: add <need_eot> to trailers parsing function"
---