tests: add a way to fix revision number usages automatically in tests

Summary:
Some tests use too many revision numbers (ex. test-rebase-scenario-global.t).
It's time consuming to fix it manually. Write a script to help with that.

This will be used by the next diff.

Reviewed By: DurhamG

Differential Revision: D22174956

fbshipit-source-id: a873ad326ecbd9cdfa2df58839b0ef21626e1506

eden/scm/edenscm/mercurial/revset.py

@@ -197,6 +197,7 @@ def stringset(repo, subset, x, order):
     if x.startswith("-") or x == str(i):
         # 'x' was used as a revision number. Maybe warn and log it.
         _warnrevnum(repo.ui, x)
+        scmutil.trackrevnumfortests(repo, [x])
 
     x = i
     if x in subset or x == node.nullrev and isinstance(subset, fullreposet):

eden/scm/edenscm/mercurial/scmutil.py

@@ -1384,3 +1384,58 @@ def contextnodesupportingwdir(ctx):
         )
 
     return ctx.node()
+
+
+def trackrevnumfortests(repo, specs):
+    """Attempt to collect information to replace revision number with revset
+    expressions in tests.
+
+    This works with the TESTFILE and TESTLINE environment variable set by
+    run-tests.py.
+
+    Information will be written to $TESTDIR/.testrevnum.
+    """
+    if not util.istest():
+        return
+
+    trackrevnum = encoding.environ.get("TRACKREVNUM")
+    testline = encoding.environ.get("TESTLINE")
+    testfile = encoding.environ.get("TESTFILE")
+    testdir = encoding.environ.get("TESTDIR")
+    if not trackrevnum or not testline or not testfile or not testdir:
+        return
+
+    for spec in specs:
+        # 'spec' should be in sys.argv
+        if not any(spec in a for a in pycompat.sysargv):
+            continue
+        # Consider 'spec' as a revision number.
+        rev = int(spec)
+        if rev < -1:
+            continue
+        ctx = repo[rev]
+        if not ctx:
+            return
+
+        # Check candidate revset expressions.
+        candidates = []
+        if rev == -1:
+            candidates.append("null")
+        desc = ctx.description()
+        if desc:
+            candidates.append("desc(%s)" % desc.split()[0])
+            candidates.append("max(desc(%s))" % desc.split()[0])
+        candidates.append("%s" % ctx.hex())
+
+        repo = repo.unfiltered()
+        for candidate in candidates:
+            try:
+                nodes = list(repo.nodes(candidate))
+            except Exception:
+                continue
+            if nodes == [ctx.node()]:
+                with open(testdir + "/.testrevnum", "ab") as f:
+                    f.write(
+                        "fix(%r, %s, %r, %r)\n" % (testfile, testline, spec, candidate)
+                    )
+                break

eden/scm/tests/fix-revnum.py

@@ -0,0 +1,125 @@
+#!/usr/bin/env python3
+# Copyright (c) Facebook, Inc. and its affiliates.
+#
+# This software may be used and distributed according to the terms of the
+# GNU General Public License version 2.
+
+"""
+Rewrite rev numbers in tests reported by scmutil.trackrevnumfortests
+
+Usage:
+- Delete `.testrevnum`.
+- Set `TRACKREVNUM`.
+- Run tests to fix. Rev number usages are written in `.testrevnum`.
+- Run this script.
+"""
+
+import re
+import sys
+
+
+contents = {}  # {path: lines}
+fixed = set()
+
+commithashre = re.compile(r"\A[0-9a-f]{6,40}\Z")
+
+quote = repr
+
+progress = 0
+
+
+def fix(path, linenum, numstr, spec):
+    """Attempt to replace numstr to spec at the given file and line"""
+    # Do not fix the same thing multiple times.
+    key = (path, linenum, numstr)
+    if key in fixed:
+        return
+    else:
+        fixed.add(key)
+
+    if path not in contents:
+        with open(path, "rb") as f:
+            contents[path] = f.read().splitlines(True)
+    lines = contents[path]
+    line = lines[linenum].decode("utf-8")
+    newline = processline(line, numstr, spec)
+    lines[linenum] = newline.encode("utf-8")
+
+    global progress
+    progress += 1
+    sys.stderr.write("\r%d fixes" % progress)
+    sys.stderr.flush()
+
+
+def processline(line, numstr, spec):
+    """Replace numstr with spec in line, with some care about escaping"""
+    # Do not replace '1' in "hg commit -m 'public 1'".
+    # Do not replace '1' in "hg clone repo1 repo2".
+    # Do not replace '1' in "hg debugbuilddag '..<2.*1/2:m<2+3:c<m+3:a<2.:b<m+2:d<2.:e<m+1:f'"
+    # Do not rewrite non-hg commands, like "initrepo repo1".
+    if (
+        any(s in line for s in ["hg commit", "hg clone", "debugbuilddag"])
+        or "hg" not in line
+    ):
+        return line
+
+    alnumstr = "abcdefghijklmnopqrstuvwxyz0123456789()"
+    if numstr.startswith("-"):
+        # A negative rev number.
+        alnumstr += "-"
+    alnumstr = set(alnumstr)
+
+    newline = ""
+    buf = ""
+    singlequoted = False
+    doublequoted = False
+
+    for ch in line:
+        if ch in alnumstr:
+            buf += ch
+        else:
+            # A separator. Append 'buf'.
+            # Do not rewrite a commit hash, even if it has numbers in it.
+            # Do not replace '1' in 'HGPLAIN=1'.
+            # Do not replace '1' in '-R repo1' or '--cwd repo1'.
+            # Do not rewrite bookmark-like name (ex. foo@1).
+            # Do not rewrite redirections like `2>&1`.
+            # Do not rewrite recursively (ex. 2 -> desc(a2) -> desc(adesc(2))).
+            if (
+                not commithashre.match(buf)
+                and not any(
+                    newline.endswith(c)
+                    for c in ("=", "&", ">", "<", "@", "-R ", "--cwd ")
+                )
+                and ch not in {">", "<"}
+                and spec not in buf
+                and "desc(" not in buf
+            ):
+                if singlequoted or doublequoted or "(" not in spec:
+                    buf = buf.replace(numstr, spec)
+                else:
+                    buf = buf.replace(numstr, quote(spec))
+            newline += buf
+            buf = ""
+
+            if ch == "'" and not newline.endswith("\\"):
+                singlequoted = not singlequoted
+            elif ch == '"' and not newline.endswith("\\"):
+                doublequoted = not doublequoted
+
+            newline += ch
+
+    return newline
+
+
+with open(".testrevnum", "rb") as f:
+    exec(f.read().decode("utf-8"))
+
+for path, lines in contents.items():
+    sys.stderr.write("\rwriting %s\n" % path)
+    sys.stderr.flush()
+    with open(path, "wb") as f:
+        f.write(b"".join(lines))
+
+with open(".testrevnum", "wb") as f:
+    f.write(b"")

eden/scm/tests/run-tests.py

@@ -1488,6 +1488,7 @@ class Test(unittest.TestCase):
                 env["PYTHONUSERBASE"] = sysconfig.get_config_var("userbase")
         env["HGEMITWARNINGS"] = "1"
         env["TESTTMP"] = self._testtmp
+        env["TESTFILE"] = self.path
         env["HOME"] = self._testtmp
         if self._usechg:
             env["CHGDISABLE"] = "0"
@@ -2044,6 +2045,7 @@ class TTest(Test):
                 cmd = l[4:].split()
                 if len(cmd) == 2 and cmd[0] == b"cd":
                     l = b"  $ cd %s || exit 1\n" % cmd[1]
+                script.append(b"export TESTLINE=%d\n" % n)
                 script.append(l[4:])
             elif l.startswith(b"  > "):  # continuations
                 after.setdefault(prepos, []).append(l)