Merge pull request #5 from furious-luke/master

Processing entry-point dependencies.

pipeline_browserify/compiler.py

@@ -1,6 +1,8 @@
 from pipeline.compilers import SubProcessCompiler
 from os.path import dirname
+import json
 from django.conf import settings
+from django.core.exceptions import SuspiciousFileOperation
 
 class BrowserifyCompiler(SubProcessCompiler):
     output_extension = 'browserified.js'
@@ -23,3 +25,61 @@ def compile_file(self, infile, outfile, outdated=False, force=False):
         )
         print('\ncommand:', command)
         return self.execute_command(command.split(), cwd=dirname(infile))
+
+    def is_outdated(self, infile, outfile):
+        """Check if the input file is outdated.
+
+        The difficulty with the default implementation is that any file that is
+        `require`d from the entry-point file will not trigger a recompile if it
+        is modified. This overloaded version of the method corrects this by generating
+        a list of all required files that are also a part of the storage manifest
+        and checking if they've been modified since the last compile.
+
+        The command used to generate the list of dependencies is the same as the compile
+        command but includes the `--deps` option.
+
+        WARNING: It seems to me that just generating the dependencies may take just
+        as long as actually compiling, which would mean we would be better off just
+        forcing a compile every time.
+        """
+
+        # Check for missing file or modified entry-point file.
+        if super(BrowserifyCompiler, self).is_outdated(infile, outfile):
+            return True
+
+        # Check if we've already calculated dependencies.
+        deps = getattr(self, '_deps', None)
+        if not deps:
+
+            # Collect dependency information.
+            command = "%s %s %s --deps %s" % (
+                getattr(settings, 'PIPELINE_BROWSERIFY_VARS', ''),
+                getattr(settings, 'PIPELINE_BROWSERIFY_BINARY', '/usr/bin/env browserify'),
+                getattr(settings, 'PIPELINE_BROWSERIFY_ARGUMENTS', ''),
+                self.storage.path(infile),
+            )
+            dep_json = self.execute_command(command) #, cwd=dirname(infile))
+
+            # Process the output data. It's JSON, and the file's path is coded
+            # in the "file" field. We also want to save the content of each file
+            # so we can check if they're outdated, which is coded under "source".
+            deps = []
+            for dep in json.loads(dep_json.decode()):
+
+                # Is this file managed by the storage?
+                try:
+                    exists = self.storage.exists(dep['file'])
+                except SuspiciousFileOperation:
+                    exists = None
+                if exists == True or exists == False:
+                    deps.append(dep['file'])
+
+            # Cache the dependencies for the next possible run.
+            self._deps = deps
+
+        # Test the dependencies to see if they're out of date.
+        for dep in deps:
+            if super(BrowserifyCompiler, self).is_outdated(dep, outfile):
+                return True
+
+        return False