Vagrant Cascadian wrote 24 hours ago
(address . bug-guix@gnu.org)
87zfhs11il.fsf@wireframe
Guix seems to be a bit overly cautious about the filename lengths used
in the test suite:
make check-TESTS check-local
make[5]: Entering directory '/build/reproducible-path/guix-1.4.0+154709.ab1b557d8f3'
make[6]: Entering directory '/build/reproducible-path/guix-1.4.0+154709.ab1b557d8f3'
============================================================================
Testsuite summary for GNU Guix 1.4.0+154709.ab1b557d8f3
============================================================================
# TOTAL: 0
# PASS: 0
# SKIP: 0
# XFAIL: 0
# FAIL: 0
# XPASS: 0
# ERROR: 0
============================================================================
make[6]: Leaving directory
'/build/reproducible-path/guix-1.4.0+154709.ab1b557d8f3'
Cannot run tests because file name limits would be exceeded.
Look for 'length' in the 'config.log' file for details.
That is a base build path of only 55 characters. I managed to get it to
build successfully by shortening it down to use a version that was only
49 characters total... but that leaves very little wiggle room...
While that is an unusually long version number as part of the
directory(it a version roughly based on the output of "git
describe")... this seems likely to be a fairly arbitrary and unecessary
length limit, at least with any reasonably modern filesystem...
In another build performed where this check failed with 50 character
build paths, I checked config.log:
$ grep length config.log
configure:7282: checking the length of the installed socket file name
configure:7305: checking the length of the socket file name used in tests
configure:7316: checking the length of a typical hash bang line
configure:7328: checking the length of a hash bang line used in tests
ac_cv_guix_hash_bang_length=73
ac_cv_guix_socket_file_name_length=30
ac_cv_guix_test_hash_bang_length=127
ac_cv_guix_test_socket_file_name_length=90
File name limits of 90 characters? Many store items are much longer than
that!
Why is this check implemented, and do we really still need it? :)
live well,
vagrant
-----BEGIN PGP SIGNATURE-----
iHUEARYKAB0WIQRlgHNhO/zFx+LkXUXcUY/If5cWqgUCZ8/jEgAKCRDcUY/If5cW
qnETAP46KxVEUGGr0BMzwBSf47sp6h3D9xXEoZU4PJuXCpAmlwEAnp+9Jt+c4aux
gtzGjJtSaUJnsT3PJEIhhe7L8UneOAI=
=Aoy/
-----END PGP SIGNATURE-----