Script 'mail_helper' called by obssrc Hello community, here is the log from the commit of package socat for openSUSE:Factory checked in at 2026-10-23 20:21:19 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Comparing /work/SRC/openSUSE:Factory/socat (Old) and /work/SRC/openSUSE:Factory/.socat.new.2252680 (New) ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Package is "socat" Fri Oct 23 20:21:19 2026 rev:53 rq:1386768 version:1.8.1.3 Changes: -------- --- /work/SRC/openSUSE:Factory/socat/socat.changes 2026-08-06 16:18:55.461584423 +0200 +++ /work/SRC/openSUSE:Factory/.socat.new.2252680/socat.changes 2026-10-23 20:21:35.193722070 +0200 @@ -1,0 +2,5 @@ +Wed Oct 21 07:41:11 UTC 2026 - Jiri Slaby <[email protected]> + +- Add socat-chain-ssl-l-race.patch + +------------------------------------------------------------------- New: ---- socat-chain-ssl-l-race.patch ----------(New B)---------- New: - Add socat-chain-ssl-l-race.patch ----------(New E)---------- ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Other differences: ------------------ ++++++ socat.spec ++++++ --- /var/tmp/diff_new_pack.2NdEAB/_old 2026-10-23 20:21:35.799747431 +0200 +++ /var/tmp/diff_new_pack.2NdEAB/_new 2026-10-23 20:21:35.801747514 +0200 @@ -31,6 +31,8 @@ # TODO: as of 1.8.0.0, test.sh supports "--expect-fail <code>", this should be used # instead of ignoring all test failures Patch1: socat-ignore-tests-failure-boo1078346.patch +# PATCH-FIX-UPSTREAM fix race in socat-chain.sh with passive SSL-L (test 530) +Patch2: socat-chain-ssl-l-race.patch BuildRequires: iputils BuildRequires: net-tools BuildRequires: openssl-devel ++++++ socat-chain-ssl-l-race.patch ++++++ From: Jiri Slaby <[email protected]> Subject: fix race in socat-chain.sh References: build-fix Patch-mainline: no Fix race in socat-chain.sh where the connecting instance did not wait for the passive SSL-L instance to listen, and make test 530 (SOCAT_CHAIN_SSL_PTY) wait for the PTY link instead of calling waittcp4port with a file name; fixes sporadic failure of test 530 (seen on i586 with kernel 7.3). --- a/socat-chain.sh +++ b/socat-chain.sh @@ -223,7 +223,7 @@ ADDR1B="$ARG1A,bind=$LOCALHOST,$ARG1B" ADDR2A="TCP4-L:$PORT,reuseaddr,$FORK,bind=$LOCALHOST,range=$LOCALHOST/32" else - ADDR1B="TCP4:$LOCALHOST:$PORT,bind=$LOCALHOST" + ADDR1B="TCP4:$LOCALHOST:$PORT,bind=$LOCALHOST,retry=50,interval=0.1" ADDR2A="$ARG1A,reuseaddr,$FORK,bind=$LOCALHOST,range=$LOCALHOST/32,$ARG1B" fi ADDR2B="$ARG2" @@ -245,6 +245,8 @@ "$ADDR2A" \ "$ADDR2B" & pid2=$! +# Passive instance 2 might not yet listen when instance 1 connects (e.g., slow +# SSL initialization); the retry option on TCP4 in ADDR1B handles this sleep 0.1 #trap 'if ! kill -n 0 $pid1 2>/dev/null; then [ -z "$QUIET" ] && echo "$0: socat-$NAME1 exited with rc=$?" >&2; kill $pid2 2>/dev/null; exit 1; elif ! kill -n 0 $pid2 2>/dev/null; then [ -z "$QUIET" ] && echo "$0: socat-$NAME2 exited with rc=$?" >&2; kill $pid1 2>/dev/null; exit 1; fi' SIGCHLD --- a/test.sh +++ b/test.sh @@ -18644,7 +18644,7 @@ printf "test $F_n $TEST... " $N $CMD0 >/dev/null 2>"${te}0" & pid0=$! - waittcp4port $tp + waitfile $tp 1 100 # NetBSD-9 seems to need massive delay { echo "$da"; relsleep 100; } |$CMD1 >"${tf}1" 2>"${te}1" rc1=$?
