tests/builders.scm hangs

[Andreas Enge]

Hello,

 

after downloading the lastest git version (a0f16bb), "./autoreconf -vi" and configuring with

./configure --prefix=/usr/local/guix-git --with-nix-prefix=/usr/local/nix-1.1

(by the way, would not "--with-nix" be a more standard choice than "--with-nix-prefix"?) on Debian testing, "make" passes, but "make check" poses problems. Some tests pass without the nix-worker daemon running:

PASS: tests/base32.scm

FAIL: tests/builders.scm

FAIL: tests/derivations.scm

PASS: tests/utils.scm

PASS: tests/build-utils.scm

FAIL: tests/packages.scm

PASS: tests/snix.scm

PASS: tests/union.scm

FAIL: tests/guix-build.sh

FAIL: tests/guix-download.sh

FAIL: tests/guix-package.sh

 

When the nix-worker daemon is running, builders.scm hangs.

"ps -ef|grep nix" shows the following list of processes:

root 13947 4755 0 20:03 pts/0 00:00:00 nix-worker --daemon

privat 14031 14026 0 20:06 pts/1 00:00:00 /bin/bash -c list='' list2='tests/base32.log tests/builders.log tests/derivations.log tests/utils.log tests/build-utils.log tests/packages.log tests/snix.log tests/union.log tests/guix-build.log tests/guix-download.log tests/guix-package.log'; for f in $list2; do \ test .log = $f && continue; \ if test check-TESTS = recheck; then \ test -f $f || continue; \ if test -r $f && read line < $f; then \ case $line in FAIL*|XPASS*) : ;; *) continue;; esac; \ fi; \ fi; \ if test -z "$list"; then list=$f; else list="$list $f"; fi; \ done; \ if test check-TESTS = recheck && test -n "$list"; then \ { am__dry=no; case $MAKEFLAGS in *\\[\ \?]*) echo 'am--echo: ; @echo "AM" OK' | make -f - 2>/dev/null | grep '^AM OK$' >/dev/null || am__dry=yes;; *) for am__flg in $MAKEFLAGS; do case $am__flg in *=*|--*) ;; *n*) am__dry=yes; break;; esac; done;; esac; test $am__dry = yes; } || rm -f $list || exit 1; \ fi; \ make test-suite.log TEST_LOGS="$list"

privat 14032 14031 0 20:06 pts/1 00:00:00 make test-suite.log TEST_LOGS=tests/base32.log tests/builders.log tests/derivations.log tests/utils.log tests/build-utils.log tests/packages.log tests/snix.log tests/union.log tests/guix-build.log tests/guix-download.log tests/guix-package.log

root 14061 13947 3 20:06 ? 00:00:00 nix-worker 14060

root 14062 14061 0 20:06 ? 00:00:00 /usr/bin/perl -w -I/usr/local/nix-1.1/lib/perl5/site_perl/5.14.2/x86_64-linux-gnu-thread-multi /usr/local/nix-1.1/libexec/nix/substituters/copy-from-other-stores.pl --query

root 14063 14061 0 20:06 ? 00:00:00 /usr/bin/perl -w -I/usr/local/nix-1.1/lib/perl5/site_perl/5.14.2/x86_64-linux-gnu-thread-multi /usr/local/nix-1.1/libexec/nix/substituters/download-using-manifests.pl --query

root 14064 14061 0 20:06 ? 00:00:00 guile --no-auto-compile -L /nix/store/7asnahbrlirf5dwyd16w0qpwciv3kj3z-module-import /nix/store/q41qjkxi1h8w6qjgv6j2vcv9rf9ka4pg-make-3.82.tar.bz2-guile-builder

 

I tried to execute the last few ones by hand.

/usr/local/nix-1.1/libexec/nix/substituters/copy-from-other-stores.pl --query

and

/usr/local/nix-1.1/libexec/nix/substituters/download-using-manifests.pl --query

do not return. So maybe there is a problem with my nix installation?

 

guile --no-auto-compile -L /nix/store/7asnahbrlirf5dwyd16w0qpwciv3kj3z-module-import /nix/store/q41qjkxi1h8w6qjgv6j2vcv9rf9ka4pg-make-3.82.tar.bz2-guile-builder

results in

starting download of `#f' from `http://ftpmirror.gnu.org/make/make-3.82.tar.bz2'...

following redirection to `http://ftp.igh.cnrs.fr/pub/gnu/make/make-3.82.tar.bz2'...

ERROR: Throw to key `bad-response' with args `("EOF while reading response body: ~a bytes of ~a" (17295 1242186))'.

starting download of `#f' from `ftp://ftp.cs.tu-berlin.de/pub/gnu/make/make-3.82.tar.bz2'...

ERROR: Wrong type (expecting string): #f

starting download of `#f' from `ftp://ftp.chg.ru/pub/gnu/make/make-3.82.tar.bz2'...

which is a very suspicious progress message; is the `#f' normal?

 

I can download the file http://ftp.igh.cnrs.fr/pub/gnu/make/make-3.82.tar.bz2

by hand without problem. The ftp access to ftp://ftp.chg.ru/pub/gnu/make/make-3.82.tar.bz2 seems to be blocked by my current isp.

 

Thank you very much for your help,

 

Andreas