# HG changeset patch
# User Boris Feld <boris.feld@octobus.net>
# Date 2019-01-28 09:47:40
# Node ID d65ba1ff45592366a305fa27b38d09a38fbf479c
# Parent  ab6d1f82be32dbeef7981f3e10b84c5ea85698e0

perf: add some documentation to perfindex

It seems useful to document how the arguments can affect the benchmark.

diff --git a/contrib/perf.py b/contrib/perf.py
--- a/contrib/perf.py
+++ b/contrib/perf.py
@@ -1017,6 +1017,14 @@ def perfignore(ui, repo, **opts):
             (b'', b'rev', b'', b'revision to be looked up (default tip)'),
          ] + formatteropts)
 def perfindex(ui, repo, **opts):
+    """benchmark index creation time followed by a lookup
+
+    The default is to look `tip` up. Depending on the index implementation,
+    the revision looked up can matters. For example, an implementation
+    scanning the index will have a faster lookup time for `--rev tip` than for
+    `--rev 0`.
+
+    It is not currently possible to check for lookup of a missing node."""
     import mercurial.revlog
     opts = _byteskwargs(opts)
     timer, fm = gettimer(ui, opts)
diff --git a/tests/test-contrib-perf.t b/tests/test-contrib-perf.t
--- a/tests/test-contrib-perf.t
+++ b/tests/test-contrib-perf.t
@@ -93,7 +93,7 @@ perfstatus
                  find statistic about potential parameters for the
                  'perftracecopies'
    perfignore    benchmark operation related to computing ignore
-   perfindex     (no help text available)
+   perfindex     benchmark index creation time followed by a lookup
    perflinelogedits
                  (no help text available)
    perfloadmarkers