The machinery to map paths through aliases for merging coverage data from disparate machines. Part of fixing #17.

2 files changed, 87 insertions, 3 deletions

diff --git a/coverage/data.py b/coverage/data.py
index 3263cb38..1cc4c950 100644
--- a/coverage/data.py
+++ b/coverage/data.py
@@ -1,8 +1,10 @@
 """Coverage data for Coverage."""
 
-import os
+import fnmatch, os, re
 
 from coverage.backward import pickle, sorted        # pylint: disable=W0622
+from coverage.files import PathAliases
+from coverage.misc import CoverageException
 
 
 class CoverageData(object):
@@ -169,13 +171,17 @@ class CoverageData(object):
             pass
         return lines, arcs
 
-    def combine_parallel_data(self):
+    def combine_parallel_data(self, aliases=None):
         """Combine a number of data files together.
 
         Treat `self.filename` as a file prefix, and combine the data from all
         of the data files starting with that prefix plus a dot.
 
+        If `aliases` is provided, it's a PathAliases object that is used to
+        re-map paths to match the local machine's.
+
         """
+        aliases = aliases or PathAliases()
         data_dir, local = os.path.split(self.filename)
         localdot = local + '.'
         for f in os.listdir(data_dir or '.'):
@@ -183,8 +189,10 @@ class CoverageData(object):
                 full_path = os.path.join(data_dir, f)
                 new_lines, new_arcs = self._read_file(full_path)
                 for filename, file_data in new_lines.items():
+                    filename = aliases.map(filename)
                     self.lines.setdefault(filename, {}).update(file_data)
                 for filename, file_data in new_arcs.items():
+                    filename = aliases.map(filename)
                     self.arcs.setdefault(filename, {}).update(file_data)
                 if f != local:
                     os.remove(full_path)
diff --git a/coverage/files.py b/coverage/files.py
index a68a0a7f..f1046ed1 100644
--- a/coverage/files.py
+++ b/coverage/files.py
@@ -1,7 +1,8 @@
 """File wrangling."""
 
 from coverage.backward import to_string
-import fnmatch, os, sys
+from coverage.misc import CoverageException
+import fnmatch, os, re, sys
 
 class FileLocator(object):
     """Understand how filenames work."""
@@ -118,6 +119,81 @@ class FnmatchMatcher(object):
         return False
 
 
+class PathAliases(object):
+    """A collection of aliases for paths.
+
+    When combining data files from remote machines, often the paths to source
+    code are different, for example, due to OS differences, or because of
+    serialized checkouts on continuous integration machines.
+
+    A `PathAliases` object tracks a list of pattern/result pairs, and can
+    map a path through those aliases to produce a unified path.
+
+    """
+    def __init__(self):
+        self.aliases = []
+
+    def _sep(self, s):
+        """Find the path separator used in this string, or os.sep if none."""
+        sep_match = re.search(r"[\\/]", s)
+        if sep_match:
+            sep = sep_match.group(0)
+        else:
+            sep = os.sep
+        return sep
+
+    def add(self, pattern, result):
+        """Add the `pattern`/`result` pair to the list of aliases.
+
+        `pattern` is an `fnmatch`-style pattern.  `result` is a simple
+        string.  When mapping paths, if a path starts with a match against
+        `pattern`, then that match is replaced with `result`.  This models
+        isomorphic source trees being rooted at different places on two
+        different machines.
+
+        `pattern` can't end with a wildcard component, since that would
+        match an entire tree, and not just its root.
+
+        """
+        # The pattern can't end with a wildcard component.
+        pattern = pattern.rstrip(r"\/")
+        if pattern.endswith("*"):
+            raise CoverageException("Pattern must not end with wildcards.")
+        pattern_sep = self._sep(pattern)
+        pattern += pattern_sep
+
+        # Make a regex from the pattern.  fnmatch always adds a \Z to match
+        # the whole string, which we don't want.
+        regex_pat = fnmatch.translate(pattern).replace(r'\Z', '')
+        regex = re.compile("(?i)" + regex_pat)
+
+        # Normalize the result: it must end with a path separator.
+        result_sep = self._sep(result)
+        result = result.rstrip(r"\/") + result_sep
+        self.aliases.append((regex, result, pattern_sep, result_sep))
+
+    def map(self, path):
+        """Map `path` through the aliases.
+
+        `path` is checked against all of the patterns.  The first pattern to
+        match is used to replace the root of the path with the result root.
+        Only one pattern is ever used.  If no patterns match, `path` is
+        returned unchanged.
+
+        The separator style in the result is made to match that of the result
+        in the alias.
+
+        """
+        for regex, result, pattern_sep, result_sep in self.aliases:
+            m = regex.match(path)
+            if m:
+                new = path.replace(m.group(0), result)
+                if pattern_sep != result_sep:
+                    new = new.replace(pattern_sep, result_sep)
+                return new
+        return path
+
+
 def find_python_files(dirname):
     """Yield all of the importable Python files in `dirname`, recursively."""
     for dirpath, dirnames, filenames in os.walk(dirname, topdown=True):