[lit] Use inodes instead of realpaths in the config map

[zturner]

Summary

This should better address symlinks and other weird situations instead of relying on path string comparisons.

[zturner]

Added missing file.

[dlj]

I think the previous logic should have yielded a fairly stable identifier... do you have an example off-hand of a case that fails?

(i.e., I would have expected the call to os.path.realpath() to have done the moral equivalent of looking up a canonical path based on the inode, which normpath() alone won't do.)

llvm/utils/lit/lit/discovery.py

51

.st_ino?

[zturner]

In D38015#874692, @dlj wrote:

I think the previous logic should have yielded a fairly stable identifier... do you have an example off-hand of a case that fails?

(i.e., I would have expected the call to os.path.realpath() to have done the moral equivalent of looking up a canonical path based on the inode, which normpath() alone won't do.)

I don't have an example of where the former normalization breaks, but it just seemed like this is exactly what inodes were designed for. Plus if realpath is going to look up the inode anyway, might as well use it instead of then converting back into some string. So I think the code will work either way, I just felt better about having a map of integers instead of a map of strings since there's fewer things that could go wrong.

[zturner]

So apparently on Windows st_ino is 0 until Python 3.4. So, that's a deal breaker.

llvm/utils/lit/lit/discovery.py

def search1(path):
         # This is a private builtin parameter which can be used to perform
         # translation of configuration paths.  Specifically, this parameter
         # can be set to a dictionary that the discovery process will consult
         # when it finds a configuration it is about to load.  If the given
         # path is in the map, the value of that key is a path to the
         # configuration to load instead.
         config_map = litConfig.params.get('config_map')
         if config_map:
-            cfgpath = os.path.realpath(cfgpath)
-            cfgpath = os.path.normpath(cfgpath)
-            cfgpath = os.path.normcase(cfgpath)
-            target = config_map.get(cfgpath)
+            source_inode = os.stat(cfgpath)
+            target = config_map.get(source_inode)
             if target:
                 cfgpath = target
 
         # We found a test suite, create a new config for it and load it.
         if litConfig.debug:
             litConfig.note('loading suite config %r' % cfgpath)
 
         cfg = TestingConfig.fromdefaults(litConfig)

llvm/utils/llvm-lit/llvm-lit.in

 #!/usr/bin/env python
 
 import os
 import sys
 
 config_map = {}
 
 def map_config(source_dir, site_config):
     global config_map
-    source_dir = os.path.realpath(source_dir)
-    source_dir = os.path.normpath(source_dir)
-    source_dir = os.path.normcase(source_dir)
+    source_inode = os.stat(source_dir).st_ino
     site_config = os.path.normpath(site_config)
-    config_map[source_dir] = site_config
+    config_map[source_inode] = site_config
 
 # Variables configured at build time.
 llvm_source_root = "@LLVM_SOURCE_DIR@"
 llvm_obj_root = "@LLVM_BINARY_DIR@"
 
 # Make sure we can find the lit package.
 sys.path.insert(0, os.path.join(llvm_source_root, 'utils', 'lit'))