Before, running a test would give you a build/ directory in the root
of your Mercurial source tree. The directory had a full copy of the
the source, so a grep in '**/*.py' would find files inside build/.
This adds a " (glob)" marker that works like a simpler version of
(re): "*" is converted to ".*", and "?" is converted to ".".
Both special characters can be escaped using "\", and the backslash
itself can be escaped as well.
Other glob-style syntax, like "**", "[chars]", or "[!chars]", isn't
supported.
Consider this test:
$ hg glog --template '{rev}:{node|short} "{desc}"\n'
@ 2:20c4f79fd7ac "3"
|
| o 1:38f24201dcab "2"
|/
o 0:2a18120dc1c9 "1"
Because each line beginning with "|" can be compiled as a regular
expression (equivalent to ".*|"), they will match any output.
Similarly:
$ echo foo
The blank output line can be compiled as a regular expression and will
also match any output.
With this patch, none of the above output lines will be matched as
regular expressions. A line must end in " (re)" in order to be matched
as one.
Lines are still matched literally first, so the following will pass:
$ echo 'foo (re)'
foo (re)
Regular expressions in the test suite are currently written assuming
that you need a trailing ".*" to avoid matching to the end.
Instead of matching regular expressions using "^pattern", this patch
makes matching more restrictive by matching "^pattern$".
Currently, the following unified test will pass:
$ echo foo
A blank output line (a line containing just two spaces) will match any
output.
The patch modifies the unified test runner to ignore empty strings
strings when do regular expression matching.
The options parsing logic replaces the options.blacklist initial value,
a list of filenames, with a dict of test names to filenames. When
runchildren rebuilds the command line to launch child processes the dict
is flattened and provided as a malformed argument when a filename is
expected.
Fix is to remove the blacklist option from child invocations since the
filtering is handled in the parent anyway.
Following discussions with Gilles Morris [1], it seems that it is preferable to
use several blacklist files in a blacklists/ directory. It is easier to add an
unversioned file for experiments than modifying a tracked file.
Also fall back to a simpler syntax, giving up ConfigParser, now that section
names are not needed anymore.
And allow --blacklist parameter to be a complete path, instead of only one
of the filenames contained in tests/blacklists/
[1] http://www.selenic.com/pipermail/mercurial-devel/2009-December/017317.html
Fixes bug introduced by 40ac669fd6c2 (issue1911: --tmpdir plus parallel
mode = fail), and also fixes the long-standing quirk that parallel mode
created multiple /tmp/hgtests.XXXXXX directories. Now there is only one
/tmp/hgtests.XXXXXX, with child0, child1, etc. under it.
In debug mode, we:
- don't try to diff expected vs actual output
- disallow --interactive
- disable timeouts (warn if user tried to supply one)
- don't try to parse hghave output, since we don't have the output
- don't try to save output to .err file
- instead of creating HGTMP inside tmpdir, now HGTMP is tmpdir
(thus, fail if tmpdir already exists)
- passing --tmpdir automatically turns on --keep-tmpdir
argument is a space-separated list of keywords that are searched for
in the name and body of each test. This makes it easy to run only
tests related to tags, hgweb, revert, etc. (eg -k "tag hgweb revert").
