Mercurial is a distributed source control management tool similar to Git and Bazaar. Mercurial is written in Python and is used by projects such as Mozilla for Firefox and Thunderbird.
Development versions of BLFS may not build or run some packages properly if LFS or dependencies have been updated since the most recent stable versions of the books.
Download (HTTP): https://www.mercurial-scm.org/release/mercurial-6.8.1.tar.gz
Download MD5 sum: f0c076d22c07bf65ffb02753fc516084
Download size: 7.9 MB
Estimated disk space required: 113 MB (with docs, add 1.5 GB for tests)
Estimated build time: 0.3 SBU (with docs; add 16 SBU for tests; both using parallelism=8)
docutils-0.21.2 (required to build the
documentation), git-2.46.2, GPGME-1.23.2 (with
Python bindings), OpenSSH-9.9p1 (runtime, to access
ssh://... repositories), Pygments-2.18.0, rustc-1.81.0 (see rust/README.rst
and rust/rhg/README.md
), Subversion-1.14.3
(with Python bindings), Bazaar, CVS, pyflakes, pyOpenSSL,
and re2
Build Mercurial by issuing the following command:
make build
To build the documentation (requires docutils-0.21.2), issue:
make doc
If you wish to run the tests, the rust tests must be removed as they are currently broken due to syntax errors. To do this, issue:
sed -i '138,142d' Makefile
To run the test suite, issue:
TESTFLAGS="-j<N>
--tmpdir tmp" make check
where <N>
is an
integer between one and the number of ( processor X threads ),
inclusive. Tests may fail because some error messages have changed
in Python or some deprecation warnings are printed that were not
present when the test was designed. Two tests are known to fail:
test-duplicateoptions.py and test-profile.t.
In order to investigate any apparently failing tests, you may use the run-tests.py script. To see the almost forty switches, some of them very useful, issue tests/run-tests.py --help. Running the following commands, you will execute only the tests that failed before:
pushd tests && rm -rf tmp && ./run-tests.py --tmpdir tmp test-gpg.t popd
Normally, the previous failures will be reproducible. However, if
you add the switch --debug
before
--tmpdir
, and run the tests again, some
failures may disappear, which is a problem with the test suite. If
this happens, there will be no more of these failures even if you
do not pass the --debug switch again.
An interesting switch is --time
, which
will generate a table of all the executed tests and their
respective start, end, user, system and real times once the tests
are complete. Note that these switches may be used with
make check by
including them in the TESTFLAGS
environment variable.
Install Mercurial by running the
following command (as root
):
make PREFIX=/usr install-bin
If you built the documentation, install it by running the following
command (as root
):
make PREFIX=/usr install-doc
After installation, two very quick and simple tests should run correctly. The first one needs some configuration:
cat >> ~/.hgrc << "EOF"
[ui]
username = <user_name> <user@mail>
EOF
where you must replace <user_name> and <your@mail> (mail is optional and can be omitted). With the user identity defined, run hg debuginstall and several lines will be displayed, the last one reading "no problems detected". Another quick and simple test is just hg, which should output basic commands that can be used with hg.
/etc/mercurial/hgrc
and
~/.hgrc
The great majority of extensions are disabled by default. Run hg help extensions if you need to enable any, e.g. when investigating test failures. This will output a list of enabled and disabled extensions, as well as more information such as how to enable or disable extensions using configuration files.
If you have installed make-ca-1.14
and want Mercurial to use the
certificates, as the root
user,
issue:
install -v -d -m755 /etc/mercurial &&
cat > /etc/mercurial/hgrc << "EOF"
[web]
cacerts = /etc/pki/tls/certs/ca-bundle.crt
EOF