view mercurial/help/urls.txt @ 26379:39d643252b9f

revlog: use existing file handle when reading during _addrevision

_addrevision() may need to read from revlogs as part of computing
deltas. Previously, we would flush existing file handles and open
a new, short-lived file handle to perform the reading.

If we have an existing file handle, it seems logical to reuse it
for reading instead of opening a new file handle. This patch
makes that the new behavior.

After this patch, revlog files are only reopened when adding
revisions if the revlog is switched from inline to non-inline.

On Linux when unbundling a bundle of the mozilla-central repo, this
patch has the following impact on system call counts:

Call     Before     After       Delta
write    827,639    673,390   -154,249
open     700,103    684,089    -16,014
read      74,489     74,489          0
fstat    493,924    461,896    -32,028
close    249,131    233,117    -16,014
stat     242,001    242,001          0
lstat     18,676     18,676          0
lseek     20,268     20,268          0
ioctl     14,652     13,173     -1,479
TOTAL  3,180,758  2,930,679   -250,079

It's worth noting that many of the open() calls fail due to missing
files. That's why there are many more open() calls than close().

Despite the significant system call reduction, this change does not
seem to have a significant performance impact on Linux.

On Windows 10 (not a VM, on a SSD), this patch appears to reduce
unbundle time for mozilla-central from ~960s to ~920s. This isn't
as significant as I was hoping. But a decrease it is nonetheless.
Still, Windows unbundle performance is still >2x slower than Linux.

Despite the lack of significant gains, fewer system calls is fewer
system calls. If nothing else, this will narrow the focus of potential
areas to optimize in the future.

Valid URLs are of the form::

  local/filesystem/path[#revision]
  file://local/filesystem/path[#revision]
  http://[user[:pass]@]host[:port]/[path][#revision]
  https://[user[:pass]@]host[:port]/[path][#revision]
  ssh://[user@]host[:port]/[path][#revision]

Paths in the local filesystem can either point to Mercurial
repositories or to bundle files (as created by :hg:`bundle` or
:hg:`incoming --bundle`). See also :hg:`help paths`.

An optional identifier after # indicates a particular branch, tag, or
changeset to use from the remote repository. See also :hg:`help
revisions`.

Some features, such as pushing to http:// and https:// URLs are only
possible if the feature is explicitly enabled on the remote Mercurial
server.

Note that the security of HTTPS URLs depends on proper configuration of
web.cacerts.

Some notes about using SSH with Mercurial:

- SSH requires an accessible shell account on the destination machine
  and a copy of hg in the remote path or specified with as remotecmd.
- path is relative to the remote user's home directory by default. Use
  an extra slash at the start of a path to specify an absolute path::

    ssh://example.com//tmp/repository

- Mercurial doesn't use its own compression via SSH; the right thing
  to do is to configure it in your ~/.ssh/config, e.g.::

    Host *.mylocalnetwork.example.com
      Compression no
    Host *
      Compression yes

  Alternatively specify "ssh -C" as your ssh command in your
  configuration file or with the --ssh command line option.

These URLs can all be stored in your configuration file with path
aliases under the [paths] section like so::

  [paths]
  alias1 = URL1
  alias2 = URL2
  ...

You can then use the alias for any command that uses a URL (for
example :hg:`pull alias1` will be treated as :hg:`pull URL1`).

Two path aliases are special because they are used as defaults when
you do not provide the URL to a command:

default:
  When you create a repository with hg clone, the clone command saves
  the location of the source repository as the new repository's
  'default' path. This is then used when you omit path from push- and
  pull-like commands (including incoming and outgoing).

default-push:
  The push command will look for a path named 'default-push', and
  prefer it over 'default' if both are defined.