diff options
author | William Deegan <bill@baddogconsulting.com> | 2019-07-14 21:19:49 (GMT) |
---|---|---|
committer | GitHub <noreply@github.com> | 2019-07-14 21:19:49 (GMT) |
commit | df0b5d7336b586481aace243925d688505a0dcba (patch) | |
tree | f0dfd5c4b9c89e617d5949bc59890b324fadede4 /src/engine/SCons | |
parent | 2f38ee65e95cdcea441f028d7c3ce9d1b4746862 (diff) | |
parent | a2bb59f78fb41ca5b81e220032245d4c732ba566 (diff) | |
download | SCons-df0b5d7336b586481aace243925d688505a0dcba.zip SCons-df0b5d7336b586481aace243925d688505a0dcba.tar.gz SCons-df0b5d7336b586481aace243925d688505a0dcba.tar.bz2 |
Merge pull request #3400 from bdbaddog/fix_slow_md5_decider
Fix slow MD5-timestamp decider
Diffstat (limited to 'src/engine/SCons')
-rw-r--r-- | src/engine/SCons/Action.py | 2 | ||||
-rw-r--r-- | src/engine/SCons/Environment.py | 39 | ||||
-rw-r--r-- | src/engine/SCons/Environment.xml | 13 | ||||
-rw-r--r-- | src/engine/SCons/Node/FS.py | 110 | ||||
-rw-r--r-- | src/engine/SCons/Node/__init__.py | 47 | ||||
-rw-r--r-- | src/engine/SCons/SConf.py | 9 |
6 files changed, 123 insertions, 97 deletions
diff --git a/src/engine/SCons/Action.py b/src/engine/SCons/Action.py index 3ec8a4c..c6fc575 100644 --- a/src/engine/SCons/Action.py +++ b/src/engine/SCons/Action.py @@ -534,7 +534,7 @@ class ActionBase(object): result = self.get_presig(target, source, env) if not isinstance(result,(bytes, bytearray)): - result = bytearray("",'utf-8').join([ SCons.Util.to_bytes(r) for r in result ]) + result = bytearray(result, 'utf-8') else: # Make a copy and put in bytearray, without this the contents returned by get_presig # can be changed by the logic below, appending with each call and causing very diff --git a/src/engine/SCons/Environment.py b/src/engine/SCons/Environment.py index 395a6a7..8d06af7 100644 --- a/src/engine/SCons/Environment.py +++ b/src/engine/SCons/Environment.py @@ -864,18 +864,21 @@ class SubstitutionEnvironment(object): return self -def default_decide_source(dependency, target, prev_ni): +def default_decide_source(dependency, target, prev_ni, repo_node=None): f = SCons.Defaults.DefaultEnvironment().decide_source - return f(dependency, target, prev_ni) + return f(dependency, target, prev_ni, repo_node) -def default_decide_target(dependency, target, prev_ni): + +def default_decide_target(dependency, target, prev_ni, repo_node=None): f = SCons.Defaults.DefaultEnvironment().decide_target - return f(dependency, target, prev_ni) + return f(dependency, target, prev_ni, repo_node) + def default_copy_from_cache(src, dst): f = SCons.Defaults.DefaultEnvironment().copy_from_cache return f(src, dst) + class Base(SubstitutionEnvironment): """Base class for "real" construction Environments. These are the primary objects used to communicate dependency and construction @@ -1434,30 +1437,30 @@ class Base(SubstitutionEnvironment): _warn_copy_deprecated = False return self.Clone(*args, **kw) - def _changed_build(self, dependency, target, prev_ni): - if dependency.changed_state(target, prev_ni): + def _changed_build(self, dependency, target, prev_ni, repo_node=None): + if dependency.changed_state(target, prev_ni, repo_node): return 1 - return self.decide_source(dependency, target, prev_ni) + return self.decide_source(dependency, target, prev_ni, repo_node) - def _changed_content(self, dependency, target, prev_ni): - return dependency.changed_content(target, prev_ni) + def _changed_content(self, dependency, target, prev_ni, repo_node=None): + return dependency.changed_content(target, prev_ni, repo_node) - def _changed_source(self, dependency, target, prev_ni): + def _changed_source(self, dependency, target, prev_ni, repo_node=None): target_env = dependency.get_build_env() type = target_env.get_tgt_sig_type() if type == 'source': - return target_env.decide_source(dependency, target, prev_ni) + return target_env.decide_source(dependency, target, prev_ni, repo_node) else: - return target_env.decide_target(dependency, target, prev_ni) + return target_env.decide_target(dependency, target, prev_ni, repo_node) - def _changed_timestamp_then_content(self, dependency, target, prev_ni): - return dependency.changed_timestamp_then_content(target, prev_ni) + def _changed_timestamp_then_content(self, dependency, target, prev_ni, repo_node=None): + return dependency.changed_timestamp_then_content(target, prev_ni, repo_node) - def _changed_timestamp_newer(self, dependency, target, prev_ni): - return dependency.changed_timestamp_newer(target, prev_ni) + def _changed_timestamp_newer(self, dependency, target, prev_ni, repo_node=None): + return dependency.changed_timestamp_newer(target, prev_ni, repo_node) - def _changed_timestamp_match(self, dependency, target, prev_ni): - return dependency.changed_timestamp_match(target, prev_ni) + def _changed_timestamp_match(self, dependency, target, prev_ni, repo_node=None): + return dependency.changed_timestamp_match(target, prev_ni, repo_node) def _copy_from_cache(self, src, dst): return self.fs.copy(src, dst) diff --git a/src/engine/SCons/Environment.xml b/src/engine/SCons/Environment.xml index 0982c31..1b0a04c 100644 --- a/src/engine/SCons/Environment.xml +++ b/src/engine/SCons/Environment.xml @@ -1213,6 +1213,17 @@ size, or content signature. </para> </listitem> </varlistentry> + <varlistentry> +<term><parameter>repo_node</parameter></term> +<listitem> +<para> +Use this node instead of the one specified by +<varname>dependency</varname> + to determine if the dependency has changed. +</para> +</listitem> +</varlistentry> + </variablelist> </para> @@ -1249,7 +1260,7 @@ Example: </para> <example_commands> -def my_decider(dependency, target, prev_ni): +def my_decider(dependency, target, prev_ni, repo_node=None): return not os.path.exists(str(target)) env.Decider(my_decider) diff --git a/src/engine/SCons/Node/FS.py b/src/engine/SCons/Node/FS.py index 91d349d..33105fb 100644 --- a/src/engine/SCons/Node/FS.py +++ b/src/engine/SCons/Node/FS.py @@ -57,7 +57,6 @@ import SCons.Util import SCons.Warnings from SCons.Debug import Trace -from . import DeciderNeedsNode print_duplicate = 0 @@ -2261,7 +2260,7 @@ class RootDir(Dir): this directory. """ - __slots__ = ['_lookupDict'] + __slots__ = ('_lookupDict', ) def __init__(self, drive, fs): if SCons.Debug.track_instances: logInstanceCreation(self, 'Node.FS.RootDir') @@ -2467,7 +2466,7 @@ class FileNodeInfo(SCons.Node.NodeInfoBase): """ state = getattr(self, '__dict__', {}).copy() for obj in type(self).mro(): - for name in getattr(obj,'__slots__',()): + for name in getattr(obj, '__slots__', ()): if hasattr(self, name): state[name] = getattr(self, name) @@ -2511,7 +2510,7 @@ class FileBuildInfo(SCons.Node.BuildInfoBase): or count of any of these could yield writing wrong csig, and then false positive rebuilds """ - __slots__ = ('dependency_map') + __slots__ = ['dependency_map', ] current_version_id = 2 def __setattr__(self, key, value): @@ -3283,14 +3282,14 @@ class File(Base): self._memo['changed'] = has_changed return has_changed - def changed_content(self, target, prev_ni): + def changed_content(self, target, prev_ni, repo_node=None): cur_csig = self.get_csig() try: return cur_csig != prev_ni.csig except AttributeError: return 1 - def changed_state(self, target, prev_ni): + def changed_state(self, target, prev_ni, repo_node=None): return self.state != SCons.Node.up_to_date @@ -3317,13 +3316,26 @@ class File(Base): len(binfo.bimplicitsigs)) == 0: return {} - - # store this info so we can avoid regenerating it. - binfo.dependency_map = { str(child):signature for child, signature in zip(chain(binfo.bsources, binfo.bdepends, binfo.bimplicit), + binfo.dependency_map = { child:signature for child, signature in zip(chain(binfo.bsources, binfo.bdepends, binfo.bimplicit), chain(binfo.bsourcesigs, binfo.bdependsigs, binfo.bimplicitsigs))} return binfo.dependency_map + # @profile + def _add_strings_to_dependency_map(self, dmap): + """ + In the case comparing node objects isn't sufficient, we'll add the strings for the nodes to the dependency map + :return: + """ + + first_string = str(next(iter(dmap))) + + # print("DMAP:%s"%id(dmap)) + if first_string not in dmap: + string_dict = {str(child): signature for child, signature in dmap.items()} + dmap.update(string_dict) + return dmap + def _get_previous_signatures(self, dmap): """ Return a list of corresponding csigs from previous @@ -3342,37 +3354,62 @@ class File(Base): if len(dmap) == 0: if MD5_TIMESTAMP_DEBUG: print("Nothing dmap shortcutting") return None + elif MD5_TIMESTAMP_DEBUG: print("len(dmap):%d"%len(dmap)) + - if MD5_TIMESTAMP_DEBUG: print("len(dmap):%d"%len(dmap)) - # First try the simple name for node - c_str = str(self) - if MD5_TIMESTAMP_DEBUG: print("Checking :%s"%c_str) - df = dmap.get(c_str, None) + # First try retrieving via Node + if MD5_TIMESTAMP_DEBUG: print("Checking if self is in map:%s id:%s type:%s"%(str(self), id(self), type(self))) + df = dmap.get(self, False) if df: return df + # Now check if self's repository file is in map. + rf = self.rfile() + if MD5_TIMESTAMP_DEBUG: print("Checking if self.rfile is in map:%s id:%s type:%s"%(str(rf), id(rf), type(rf))) + rfm = dmap.get(rf, False) + if rfm: + return rfm + + # get default string for node and then also string swapping os.altsep for os.sep (/ for \) + c_strs = [str(self)] + if os.altsep: - c_str = c_str.replace(os.sep, os.altsep) - df = dmap.get(c_str, None) - if MD5_TIMESTAMP_DEBUG: print("-->%s"%df) + c_strs.append(c_strs[0].replace(os.sep, os.altsep)) + + # In some cases the dependency_maps' keys are already strings check. + # Check if either string is now in dmap. + for s in c_strs: + if MD5_TIMESTAMP_DEBUG: print("Checking if str(self) is in map :%s" % s) + df = dmap.get(s, False) if df: return df + # Strings don't exist in map, add them and try again + # If there are no strings in this dmap, then add them. + # This may not be necessary, we could walk the nodes in the dmap and check each string + # rather than adding ALL the strings to dmap. In theory that would be n/2 vs 2n str() calls on node + # if not dmap.has_strings: + dmap = self._add_strings_to_dependency_map(dmap) + + # In some cases the dependency_maps' keys are already strings check. + # Check if either string is now in dmap. + for s in c_strs: + if MD5_TIMESTAMP_DEBUG: print("Checking if str(self) is in map (now with strings) :%s" % s) + df = dmap.get(s, False) + if df: + return df + + # Lastly use nodes get_path() to generate string and see if that's in dmap if not df: try: # this should yield a path which matches what's in the sconsign c_str = self.get_path() - df = dmap.get(c_str, None) - if MD5_TIMESTAMP_DEBUG: print("-->%s"%df) - if df: - return df - if os.altsep: c_str = c_str.replace(os.sep, os.altsep) - df = dmap.get(c_str, None) - if MD5_TIMESTAMP_DEBUG: print("-->%s"%df) - if df: - return df + + if MD5_TIMESTAMP_DEBUG: print("Checking if self.get_path is in map (now with strings) :%s" % s) + + df = dmap.get(c_str, None) except AttributeError as e: raise FileBuildInfoFileToCsigMappingError("No mapping from file name to content signature for :%s"%c_str) @@ -3392,16 +3429,12 @@ class File(Base): self - dependency target - target prev_ni - The NodeInfo object loaded from previous builds .sconsign - node - Node instance. This is the only changed* function which requires - node to function. So if we detect that it's not passed. - we throw DeciderNeedsNode, and caller should handle this and pass node. + node - Node instance. Check this node for file existence/timestamp + if specified. Returns: Boolean - Indicates if node(File) has changed. """ - if node is None: - # We need required node argument to get BuildInfo to function - raise DeciderNeedsNode(self.changed_timestamp_then_content) # Now get sconsign name -> csig map and then get proper prev_ni if possible bi = node.get_stored_info().binfo @@ -3433,7 +3466,6 @@ class File(Base): print("Mismatch self.changed_timestamp_match(%s, prev_ni) old:%s new:%s"%(str(target), old, new)) new_prev_ni = self._get_previous_signatures(dependency_map) - if not new: try: # NOTE: We're modifying the current node's csig in a query. @@ -3443,13 +3475,13 @@ class File(Base): return False return self.changed_content(target, new_prev_ni) - def changed_timestamp_newer(self, target, prev_ni): + def changed_timestamp_newer(self, target, prev_ni, repo_node=None): try: return self.get_timestamp() > target.get_timestamp() except AttributeError: return 1 - def changed_timestamp_match(self, target, prev_ni): + def changed_timestamp_match(self, target, prev_ni, repo_node=None): """ Return True if the timestamps don't match or if there is no previous timestamp :param target: @@ -3462,14 +3494,18 @@ class File(Base): return 1 def is_up_to_date(self): + """Check for whether the Node is current + In all cases self is the target we're checking to see if it's up to date + """ + T = 0 if T: Trace('is_up_to_date(%s):' % self) if not self.exists(): if T: Trace(' not self.exists():') - # The file doesn't exist locally... + # The file (always a target) doesn't exist locally... r = self.rfile() if r != self: - # ...but there is one in a Repository... + # ...but there is one (always a target) in a Repository... if not self.changed(r): if T: Trace(' changed(%s):' % r) # ...and it's even up-to-date... diff --git a/src/engine/SCons/Node/__init__.py b/src/engine/SCons/Node/__init__.py index aeb7092..3073d59 100644 --- a/src/engine/SCons/Node/__init__.py +++ b/src/engine/SCons/Node/__init__.py @@ -253,25 +253,10 @@ _target_from_source_map = {0 : target_from_source_none, # used by it. # - -class DeciderNeedsNode(Exception): - """ - Indicate that the decider needs the node as well as the target and the dependency. - Normally the node and the target are the same, but in the case of repository - They may be different. Also the NodeInfo is retrieved from the node - """ - def __init__(self, call_this_decider): - """ - :param call_this_decider: to return the decider to call directly since deciders - are called through several levels of indirection - """ - self.decider = call_this_decider - - # # First, the single decider functions # -def changed_since_last_build_node(node, target, prev_ni): +def changed_since_last_build_node(node, target, prev_ni, repo_node=None): """ Must be overridden in a specific subclass to return True if this @@ -292,7 +277,7 @@ def changed_since_last_build_node(node, target, prev_ni): raise NotImplementedError -def changed_since_last_build_alias(node, target, prev_ni): +def changed_since_last_build_alias(node, target, prev_ni, repo_node=None): cur_csig = node.get_csig() try: return cur_csig != prev_ni.csig @@ -300,24 +285,24 @@ def changed_since_last_build_alias(node, target, prev_ni): return 1 -def changed_since_last_build_entry(node, target, prev_ni): +def changed_since_last_build_entry(node, target, prev_ni, repo_node=None): node.disambiguate() - return _decider_map[node.changed_since_last_build](node, target, prev_ni) + return _decider_map[node.changed_since_last_build](node, target, prev_ni, repo_node) -def changed_since_last_build_state_changed(node, target, prev_ni): +def changed_since_last_build_state_changed(node, target, prev_ni, repo_node=None): return node.state != SCons.Node.up_to_date -def decide_source(node, target, prev_ni): - return target.get_build_env().decide_source(node, target, prev_ni) +def decide_source(node, target, prev_ni, repo_node=None): + return target.get_build_env().decide_source(node, target, prev_ni, repo_node) -def decide_target(node, target, prev_ni): - return target.get_build_env().decide_target(node, target, prev_ni) +def decide_target(node, target, prev_ni, repo_node=None): + return target.get_build_env().decide_target(node, target, prev_ni, repo_node) -def changed_since_last_build_python(node, target, prev_ni): +def changed_since_last_build_python(node, target, prev_ni, repo_node=None): cur_csig = node.get_csig() try: return cur_csig != prev_ni.csig @@ -1505,17 +1490,11 @@ class Node(object, with_metaclass(NoSlotsPyPy)): result = True for child, prev_ni in zip(children, then): - try: - if _decider_map[child.changed_since_last_build](child, self, prev_ni): - if t: Trace(': %s changed' % child) - result = True - except DeciderNeedsNode as e: - if e.decider(self, prev_ni, node=node): - if t: Trace(': %s changed' % child) - result = True + if _decider_map[child.changed_since_last_build](child, self, prev_ni, node): + if t: Trace(': %s changed' % child) + result = True if self.has_builder(): - import SCons.Util contents = self.get_executor().get_contents() newsig = SCons.Util.MD5signature(contents) if bi.bactsig != newsig: diff --git a/src/engine/SCons/SConf.py b/src/engine/SCons/SConf.py index 59afb40..71729c9 100644 --- a/src/engine/SCons/SConf.py +++ b/src/engine/SCons/SConf.py @@ -56,7 +56,6 @@ import SCons.Warnings import SCons.Conftest from SCons.Debug import Trace -from SCons.Node import DeciderNeedsNode # Turn off the Conftest error logging SCons.Conftest.LogInputFiles = 0 @@ -407,12 +406,10 @@ class SConfBase(object): # that the correct .sconsign info will get calculated # and keep the build state consistent. def force_build(dependency, target, prev_ni, - env_decider=env.decide_source, - node=None): + repo_node=None, + env_decider=env.decide_source): try: - env_decider(dependency, target, prev_ni) - except DeciderNeedsNode as e: - e.decider(target, prev_ni, node=target) + env_decider(dependency, target, prev_ni, repo_node) except Exception as e: raise e return True |