[scanner] Only store types in cache store

Only store types that can be referenced in the cache store
this reduces the size of the serialized Gtk-3.0.gir in the store
from 5.7M to 366k on my system.

It also reduces the time it takes to create a gir in gtksourceview
by 35% and the time to run the warnings test by more than 50%

This also disables the cache for passthrough mode since it needs
access to the whole serialized tree.

3 files changed, 45 insertions, 19 deletions

diff --git a/giscanner/girparser.py b/giscanner/girparser.py
index aedae3d3..8b638bf1 100644
--- a/giscanner/girparser.py
+++ b/giscanner/girparser.py
@@ -44,7 +44,8 @@ def _cns(tag):
 
 class GIRParser(object):
 
-    def __init__(self):
+    def __init__(self, types_only=False):
+        self._types_only = types_only
         self._shared_libraries = []
         self._includes = set()
         self._pkgconfig_packages = set()
@@ -149,8 +150,6 @@ class GIRParser(object):
             _corens('bitfield'): self._parse_enumeration_bitfield,
             _corens('callback'): self._parse_callback,
             _corens('class'): self._parse_object_interface,
-            _corens('constant'): self._parse_constant,
-            _corens('function'): self._parse_function,
             _corens('enumeration'): self._parse_enumeration_bitfield,
             _corens('interface'): self._parse_object_interface,
             _corens('record'): self._parse_record,
@@ -158,6 +157,10 @@ class GIRParser(object):
             _glibns('boxed'): self._parse_boxed,
             }
 
+        if not self._types_only:
+            parser_methods[_corens('constant')] = self._parse_constant
+            parser_methods[_corens('function')] = self._parse_function
+
         for node in ns.getchildren():
             method = parser_methods.get(node.tag)
             if method is not None:
@@ -184,6 +187,11 @@ class GIRParser(object):
 
     def _parse_generic_attribs(self, node, obj):
         assert isinstance(obj, ast.Annotated)
+        introspectable = node.attrib.get('introspectable')
+        if introspectable:
+            obj.introspectable = int(introspectable) > 0
+        if self._types_only:
+            return
         doc = node.find(_corens('doc'))
         if doc is not None:
             obj.doc = doc.text
@@ -193,9 +201,6 @@ class GIRParser(object):
         deprecated = node.attrib.get('deprecated')
         if deprecated:
             obj.deprecated = deprecated
-        introspectable = node.attrib.get('introspectable')
-        if introspectable:
-            obj.introspectable = int(introspectable) > 0
 
     def _parse_object_interface(self, node):
         parent = node.attrib.get('parent')
@@ -227,6 +232,8 @@ class GIRParser(object):
             obj.glib_type_struct = self._namespace.type_from_name(type_struct)
         self._namespace.append(obj)
 
+        if self._types_only:
+            return
         for iface in self._find_children(node, _corens('implements')):
             obj.interfaces.append(self._namespace.type_from_name(iface.attrib['name']))
         for iface in self._find_children(node, _corens('prerequisite')):
@@ -340,16 +347,17 @@ class GIRParser(object):
         if node.attrib.get('foreign') == '1':
             compound.foreign = True
         self._parse_generic_attribs(node, compound)
-        compound.fields.extend(self._parse_fields(node))
-        for method in self._find_children(node, _corens('method')):
-            compound.methods.append(
-                self._parse_function_common(method, ast.Function))
-        for func in self._find_children(node, _corens('function')):
-            compound.static_methods.append(
-                self._parse_function_common(func, ast.Function))
-        for ctor in self._find_children(node, _corens('constructor')):
-            compound.constructors.append(
-                self._parse_function_common(ctor, ast.Function))
+        if not self._types_only:
+            compound.fields.extend(self._parse_fields(node))
+            for method in self._find_children(node, _corens('method')):
+                compound.methods.append(
+                    self._parse_function_common(method, ast.Function))
+            for func in self._find_children(node, _corens('function')):
+                compound.static_methods.append(
+                    self._parse_function_common(func, ast.Function))
+            for ctor in self._find_children(node, _corens('constructor')):
+                compound.constructors.append(
+                    self._parse_function_common(ctor, ast.Function))
         return compound
 
     def _parse_record(self, node, anonymous=False):
@@ -444,6 +452,8 @@ class GIRParser(object):
                         c_symbol_prefix=node.attrib.get(_cns('symbol-prefix')))
         self._parse_generic_attribs(node, obj)
         self._namespace.append(obj)
+        if self._types_only:
+            return
         for method in self._find_children(node, _corens('method')):
             func = self._parse_function_common(method, ast.Function)
             func.is_method = True
@@ -529,5 +539,7 @@ class GIRParser(object):
         self._parse_generic_attribs(node, obj)
         self._namespace.append(obj)
 
+        if self._types_only:
+            return
         for member in self._find_children(node, _corens('member')):
             members.append(self._parse_member(member))
diff --git a/giscanner/scannermain.py b/giscanner/scannermain.py
index 556e3c95..3a9f27c9 100644
--- a/giscanner/scannermain.py
+++ b/giscanner/scannermain.py
@@ -291,6 +291,10 @@ see --identifier-prefix and --symbol-prefix."""
     transformer = Transformer(namespace,
                               accept_unprefixed=options.accept_unprefixed)
     transformer.set_include_paths(options.include_paths)
+    if options.passthrough_gir:
+        transformer.disable_cache()
+        transformer.set_passthrough_mode()
+
     shown_include_warning = False
     for include in options.includes:
         if os.sep in include:
diff --git a/giscanner/transformer.py b/giscanner/transformer.py
index be34ff5f..70addb26 100644
--- a/giscanner/transformer.py
+++ b/giscanner/transformer.py
@@ -54,6 +54,7 @@ class Transformer(object):
         self._includes = {}
         self._include_names = set()
         self._includepaths = []
+        self._passthrough_mode = False
 
     def get_includes(self):
         return self._include_names
@@ -61,6 +62,12 @@ class Transformer(object):
     def get_pkgconfig_packages(self):
         return self._pkg_config_packages
 
+    def disable_cache(self):
+        self._cachestore = None
+
+    def set_passthrough_mode(self):
+        self._passthrough_mode = True
+
     def _append_new_node(self, node):
         original = self._namespace.get(node.name)
         # Special case constants here; we allow duplication to sort-of
@@ -169,11 +176,14 @@ None."""
         sys.exit(1)
 
     def _parse_include(self, filename, uninstalled=False):
-        parser = self._cachestore.load(filename)
+        parser = None
+        if self._cachestore is not None:
+            parser = self._cachestore.load(filename)
         if parser is None:
-            parser = GIRParser()
+            parser = GIRParser(types_only=not self._passthrough_mode)
             parser.parse(filename)
-            self._cachestore.store(filename, parser)
+            if self._cachestore is not None:
+                self._cachestore.store(filename, parser)
 
         for include in parser.get_includes():
             self.register_include(include)