| 1 | Notes on Perl on the Hurd |
| 2 | Last Updated: Fri, 29 Oct 1999 22:50:30 +0200 |
| 3 | Written by: Mark Kettenis <kettenis@gnu.org> |
| 4 | |
| 5 | If you want to use Perl on the Hurd, I recommend using the Debian |
| 6 | GNU/Hurd distribution (see http://www.debian.org), even if an |
| 7 | official, stable release has not yet been made. The old `gnu-0.2' |
| 8 | binary distribution will most certainly have additional problems. |
| 9 | |
| 10 | * Known Problems |
| 11 | |
| 12 | The Perl test suite may still report some errors on the Hurd. The |
| 13 | `lib/anydbm' and `pragma/warnings' tests will almost certainly fail. |
| 14 | Both failures are not really specific to the Hurd, as indicated by the |
| 15 | test suite output. |
| 16 | |
| 17 | The socket tests may fail if the network is not configured. You have |
| 18 | to make `/hurd/pfinet' the translator for `/servers/socket/2', giving |
| 19 | it the right arguments. Try `/hurd/pfinet --help' for more |
| 20 | information. |
| 21 | |
| 22 | Here are the statistics for Perl 5.005_62 on my system: |
| 23 | |
| 24 | Failed Test Status Wstat Total Fail Failed List of failed |
| 25 | ------------------------------------------------------------------------------- |
| 26 | lib/anydbm.t 12 1 8.33% 12 |
| 27 | pragma/warnings 333 1 0.30% 215 |
| 28 | 8 tests and 24 subtests skipped. |
| 29 | Failed 2/229 test scripts, 99.13% okay. 2/10850 subtests failed, 99.98% okay. |
| 30 | |
| 31 | There are quite a few systems out there that do worse! |
| 32 | |
| 33 | However, since I am running a very recent Hurd snapshot, in which a lot of |
| 34 | bugs that were exposed by the Perl test suite have been fixed, you may |
| 35 | encounter more failures. Likely candidates are: `op/stat', `lib/io_pipe', |
| 36 | `lib/io_sock', `lib/io_udp' and `lib/time'. |
| 37 | |
| 38 | In any way, if you're seeing failures beyond those mentioned in this |
| 39 | document, please consider upgrading to the latest Hurd before reporting |
| 40 | the failure as a bug. |