On 2020-05-03, Lucas Nussbaum wrote: > During a rebuild of all packages in sid, your package failed to build > on amd64. > > Relevant part (hopefully): >> make[5]: Entering directory '/<<PKGBUILDDIR>>/tests' ... >> FAIL: dist.scm >> ============================================ >> Guile-SSH 0.11.3: tests/test-suite.log >> ============================================ >> >> # TOTAL: 11 >> # PASS: 10 >> # SKIP: 0 >> # XFAIL: 0 >> # FAIL: 1 >> # XPASS: 0 >> # ERROR: 0 >> >> .. contents:: :depth: 2 >> >> FAIL: dist >> ========== >> >> %%%% Starting test dist (Writing full log to "dist.log") >> >> Some deprecated features have been used. Set the environment >> variable GUILE_WARN_DEPRECATED to "detailed" and rerun the >> program to get more information. Set it to "no" to suppress >> this message. >> %%%% Starting test dist (Writing full log to "dist.log") >> tests/dist.scm:216: FAIL with-ssh >> # of expected passes 18 >> # of unexpected failures 1 ... >> Test begin: >> test-name: "with-ssh" >> source-file: "tests/dist.scm" >> source-line: 216 >> source-form: (test-assert "with-ssh" (begin (set-log-userdata! "with-ssh") >> (run-client-test (lambda (server) (server-listen server) (server-set! server >> (quote log-verbosity) (quote functions)) (let ((session (server-accept >> server))) (server-handle-key-exchange session) (start-session-loop session >> (lambda (msg type) (format-log/scm (quote nolog) "server" "msg: ~a; type: >> ~a" msg type) (case (car type) ((request-channel-open) (let ((c >> (message-channel-request-open-reply-accept msg))) (format-log/scm (quote >> nolog) "server" "channel 0: ~a" c) (write-line "Enter `,help' for help." c) >> (format-log/scm (quote nolog) "server" "channel 1: ~a" c) (usleep 100) (poll >> c (lambda args (let ((result (read-line c))) (format-log/scm (quote nolog) >> "server" "sexp: ~a" result) (or (string=? result "(begin (+ 21 21))") (error >> "Wrong result 1" result))) (let ((result (read-line c))) (format-log/scm >> (quote nolog) "server" "sexp: ~a" result) (or (string=? result "(newline)") >> (error "Wrong result 2" result))) (write-line "scheme@(guile-user)> $1 = >> 42\n" c) (sleep 5) (close c) (while #t (sleep 60)))))) (else >> (message-reply-success msg))))))) (lambda () (call-with-connected-session >> (lambda (session) (authenticate-server session) (format-log/scm (quote >> nolog) "client" "session: ~a" session) (unless (equal? (userauth-none! >> session) (quote success)) (error "Could not authenticate with a server" >> session)) (let ((n (make-node session #:start-repl-server? #f))) (= >> (with-ssh n (+ 21 21)) 42)))))))) >> result) (or (string=? result "(newline)") (error "Wrong result 2" result))) >> (write-line "scheme@(guile-user)> $1 = 42\n" c) (sleep 5) (close c) (while >> #t (sleep 60)))))) (else (message-reply-success msg))))))) (lambda () >> (call-with-connected-session (lambda (session) (authenticate-server session) >> (format-log/scm (quote nolog) "client" "session: ~a" session) (unless >> (equal? (userauth-none! session) (quote success)) (error "Could not >> authenticate with a server" session)) (let ((n (make-node session >> #:start-repl-server? #f))) (= (with-ssh n (+ 21 21)) 42)))))))) >> Test end: >> result-kind: fail >> actual-value: #f >> actual-error: (guile-ssh-error "read_from_channel_port" "Error reading >> from the channel" #<unknown channel (freed) 564c40af72c0> #f) >> Group end: dist >> # of expected passes 18 >> # of unexpected failures 1 >> FAIL dist.scm (exit status: 1) ... > The full build log is available from: > http://qa-logs.debian.net/2020/05/01/guile-ssh_0.11.3-2_unstable.log > > A list of current common problems and possible solutions is available at > http://wiki.debian.org/qa.debian.org/FTBFS . You're welcome to contribute!
I can't reproduce this issue on amd64 or arm64, with versions 0.11.3-2 (sid/bullseye) or 0.12.0-1 (experimental). Both versions also built fine on the buildd's, originally... Reading the above mentioned wiki page, I tried building without network to ensure it wasn't somehow trying to ssh over the network, but the tests passed. How is network disabled in your tests? According to https://tests.reproducible-builds.org/debian/guile-ssh, amd64 and arm64 have failures, but they appear to be different tests entirely. > About the archive rebuild: The rebuild was done on EC2 VM instances from > Amazon Web Services, using a clean, minimal and up-to-date chroot. All my tests were also on updated, clean, minimal chroots... is there anything unusual with your test setup not mentioned on the FAQ? > Every failed build was retried once to eliminate random failures. Did the retried build fail in the same way vs. just testing that it failed twice? Thanks for all your QA work! live well, vagrant
signature.asc
Description: PGP signature
