(
Jump to the original submission)
Fri 07 Apr 2017 09:14:26 AM UTC, comment #8:
Your seem to be able to test on OSX. Maybe you could add some lines of Perl opening a composed filename and then check with diropen/readdir if that file exists. If not, the filesystem works with a 'translation layer'. In this case, skip the test by returning 77.
|
Wed 05 Apr 2017 08:17:44 PM UTC, comment #7:
Please disregard my previous comment. This tests tries to write files with names encoded in iso-8859-1. Not sure if it possible with Mac OS X, because on HFS+ file names are stored as NFD Unicode:
https://developer.apple.com/library/content/qa/qa1173/_index.html
Probably this test can be skipped on Mac OS X.
|
Wed 05 Apr 2017 04:23:20 PM UTC, comment #6:
FYI: Mac OS uses NFD in file names. Probably test can convert name to NFC using Unicode::Normalize.
|
Wed 15 Mar 2017 12:51:55 PM UTC, comment #5:
Thanks, your output looks like mine here on Linux (ignoring dates and '`). Except the result... that means, your filesystem changes the bytes in the filenames. A later check byte-by-byte, using diropen/readdir, fails.
We addressed a similar issue in the Wget2 test suite (e.g. HFS+ is case insensitive) where I check for case sensitivity of the filesystem and amend the checks regarding of the result.
I am not going to backport that (C code) to wget 1.x (Perl or Python). But if anyone makes up a patch, it'll be welcome.
The code is in (wget2) tests/libtest.c, wget_test_check_filesystem().
|
Mon 13 Mar 2017 08:54:16 PM UTC, comment #4:
I'm currently on journaled HFS
Content (IOContent): Apple_HFS
File System Personality: Journaled HFS+
Type (Bundle): hfs
Name (User Visible): Mac OS Extended (Journaled)
Journal: Journal size 40960 KB at offset 0xe8a000
Owners: Enabled
|
Mon 13 Mar 2017 08:36:40 PM UTC, comment #3:
I just pushed a commit to run Perl tests with -d (debug output).
Please git pull and 'make check' again. Than upload test-suite.log again.
What fileystem are you using ?
Maybe there is another layer that changes file names...
|
Mon 13 Mar 2017 04:36:41 PM UTC, comment #2:
attached logs and test-suite.log.
i was able to reproduce on git master (those are the bootstrapconfigmakecheck.txt and the test-suite.log from git master entries)
|
Mon 13 Mar 2017 09:41:52 AM UTC, comment #1:
The link doesn't exist any more. Maybe you can upload the log here ?
Please also check #50233
|
Mon 13 Mar 2017 12:19:10 AM UTC, original submission:
on a homebrew build (https://bot.brew.sh/job/Homebrew%20Core/18182/), make check fails for the iri-disabled test.
I tried to google what the issue was, but quickly got confused on the status (if it's not supposed to be addressed because of the upcoming python test suite, etc).
reproduced in non-ruby forked shell as well
|
(Note: upload size limit is set to 16384 kB, after insertion of the required
escape characters.)
Attach File(s):
Comment:
Depends on the following items: None found
Items that depend on this one: None found
Follow 6 latest changes.
Date |
Changed By |
Updated Field |
Previous Value |
=> |
Replaced By |
Wed 15 Mar 2017 12:51:55 PM UTC | rockdaboot | Status | None |  | Inspected |
Mon 13 Mar 2017 08:53:35 PM UTC | stepheng | Attached File | - |  | Added test-suite.log, #40001 |
Mon 13 Mar 2017 04:35:27 PM UTC | stepheng | Attached File | - |  | Added test-suite.log, #39993 |
Mon 13 Mar 2017 04:34:42 PM UTC | stepheng | Attached File | - |  | Added bootstrapconfigmakecheck.txt, #39992 |
Mon 13 Mar 2017 04:07:46 PM UTC | stepheng | Attached File | - |  | Added test-suite.log, #39990 |
| | Attached File | - |  | Added configmakecheck.txt, #39991 |