Server IP : 66.29.132.122 / Your IP : 18.225.92.95 Web Server : LiteSpeed System : Linux business142.web-hosting.com 4.18.0-553.lve.el8.x86_64 #1 SMP Mon May 27 15:27:34 UTC 2024 x86_64 User : admazpex ( 531) PHP Version : 7.2.34 Disable Function : NONE MySQL : OFF | cURL : ON | WGET : ON | Perl : ON | Python : ON | Sudo : OFF | Pkexec : OFF Directory : /proc/self/root/opt/alt/alt-nodejs11/root/lib/node_modules/npm/node_modules/node-gyp/gyp/pylib/gyp/ |
Upload File : |
# Copyright (c) 2013 Google Inc. All rights reserved. # Use of this source code is governed by a BSD-style license that can be # found in the LICENSE file. """Utility functions shared amongst the Windows generators.""" import copy import os # A dictionary mapping supported target types to extensions. TARGET_TYPE_EXT = { 'executable': 'exe', 'loadable_module': 'dll', 'shared_library': 'dll', 'static_library': 'lib', } def _GetLargePdbShimCcPath(): """Returns the path of the large_pdb_shim.cc file.""" this_dir = os.path.abspath(os.path.dirname(__file__)) src_dir = os.path.abspath(os.path.join(this_dir, '..', '..')) win_data_dir = os.path.join(src_dir, 'data', 'win') large_pdb_shim_cc = os.path.join(win_data_dir, 'large-pdb-shim.cc') return large_pdb_shim_cc def _DeepCopySomeKeys(in_dict, keys): """Performs a partial deep-copy on |in_dict|, only copying the keys in |keys|. Arguments: in_dict: The dictionary to copy. keys: The keys to be copied. If a key is in this list and doesn't exist in |in_dict| this is not an error. Returns: The partially deep-copied dictionary. """ d = {} for key in keys: if key not in in_dict: continue d[key] = copy.deepcopy(in_dict[key]) return d def _SuffixName(name, suffix): """Add a suffix to the end of a target. Arguments: name: name of the target (foo#target) suffix: the suffix to be added Returns: Target name with suffix added (foo_suffix#target) """ parts = name.rsplit('#', 1) parts[0] = '%s_%s' % (parts[0], suffix) return '#'.join(parts) def _ShardName(name, number): """Add a shard number to the end of a target. Arguments: name: name of the target (foo#target) number: shard number Returns: Target name with shard added (foo_1#target) """ return _SuffixName(name, str(number)) def ShardTargets(target_list, target_dicts): """Shard some targets apart to work around the linkers limits. Arguments: target_list: List of target pairs: 'base/base.gyp:base'. target_dicts: Dict of target properties keyed on target pair. Returns: Tuple of the new sharded versions of the inputs. """ # Gather the targets to shard, and how many pieces. targets_to_shard = {} for t in target_dicts: shards = int(target_dicts[t].get('msvs_shard', 0)) if shards: targets_to_shard[t] = shards # Shard target_list. new_target_list = [] for t in target_list: if t in targets_to_shard: for i in range(targets_to_shard[t]): new_target_list.append(_ShardName(t, i)) else: new_target_list.append(t) # Shard target_dict. new_target_dicts = {} for t in target_dicts: if t in targets_to_shard: for i in range(targets_to_shard[t]): name = _ShardName(t, i) new_target_dicts[name] = copy.copy(target_dicts[t]) new_target_dicts[name]['target_name'] = _ShardName( new_target_dicts[name]['target_name'], i) sources = new_target_dicts[name].get('sources', []) new_sources = [] for pos in range(i, len(sources), targets_to_shard[t]): new_sources.append(sources[pos]) new_target_dicts[name]['sources'] = new_sources else: new_target_dicts[t] = target_dicts[t] # Shard dependencies. for t in new_target_dicts: for deptype in ('dependencies', 'dependencies_original'): dependencies = copy.copy(new_target_dicts[t].get(deptype, [])) new_dependencies = [] for d in dependencies: if d in targets_to_shard: for i in range(targets_to_shard[d]): new_dependencies.append(_ShardName(d, i)) else: new_dependencies.append(d) new_target_dicts[t][deptype] = new_dependencies return (new_target_list, new_target_dicts) def _GetPdbPath(target_dict, config_name, vars): """Returns the path to the PDB file that will be generated by a given configuration. The lookup proceeds as follows: - Look for an explicit path in the VCLinkerTool configuration block. - Look for an 'msvs_large_pdb_path' variable. - Use '<(PRODUCT_DIR)/<(product_name).(exe|dll).pdb' if 'product_name' is specified. - Use '<(PRODUCT_DIR)/<(target_name).(exe|dll).pdb'. Arguments: target_dict: The target dictionary to be searched. config_name: The name of the configuration of interest. vars: A dictionary of common GYP variables with generator-specific values. Returns: The path of the corresponding PDB file. """ config = target_dict['configurations'][config_name] msvs = config.setdefault('msvs_settings', {}) linker = msvs.get('VCLinkerTool', {}) pdb_path = linker.get('ProgramDatabaseFile') if pdb_path: return pdb_path variables = target_dict.get('variables', {}) pdb_path = variables.get('msvs_large_pdb_path', None) if pdb_path: return pdb_path pdb_base = target_dict.get('product_name', target_dict['target_name']) pdb_base = '%s.%s.pdb' % (pdb_base, TARGET_TYPE_EXT[target_dict['type']]) pdb_path = vars['PRODUCT_DIR'] + '/' + pdb_base return pdb_path def InsertLargePdbShims(target_list, target_dicts, vars): """Insert a shim target that forces the linker to use 4KB pagesize PDBs. This is a workaround for targets with PDBs greater than 1GB in size, the limit for the 1KB pagesize PDBs created by the linker by default. Arguments: target_list: List of target pairs: 'base/base.gyp:base'. target_dicts: Dict of target properties keyed on target pair. vars: A dictionary of common GYP variables with generator-specific values. Returns: Tuple of the shimmed version of the inputs. """ # Determine which targets need shimming. targets_to_shim = [] for t in target_dicts: target_dict = target_dicts[t] # We only want to shim targets that have msvs_large_pdb enabled. if not int(target_dict.get('msvs_large_pdb', 0)): continue # This is intended for executable, shared_library and loadable_module # targets where every configuration is set up to produce a PDB output. # If any of these conditions is not true then the shim logic will fail # below. targets_to_shim.append(t) large_pdb_shim_cc = _GetLargePdbShimCcPath() for t in targets_to_shim: target_dict = target_dicts[t] target_name = target_dict.get('target_name') base_dict = _DeepCopySomeKeys(target_dict, ['configurations', 'default_configuration', 'toolset']) # This is the dict for copying the source file (part of the GYP tree) # to the intermediate directory of the project. This is necessary because # we can't always build a relative path to the shim source file (on Windows # GYP and the project may be on different drives), and Ninja hates absolute # paths (it ends up generating the .obj and .obj.d alongside the source # file, polluting GYPs tree). copy_suffix = 'large_pdb_copy' copy_target_name = target_name + '_' + copy_suffix full_copy_target_name = _SuffixName(t, copy_suffix) shim_cc_basename = os.path.basename(large_pdb_shim_cc) shim_cc_dir = vars['SHARED_INTERMEDIATE_DIR'] + '/' + copy_target_name shim_cc_path = shim_cc_dir + '/' + shim_cc_basename copy_dict = copy.deepcopy(base_dict) copy_dict['target_name'] = copy_target_name copy_dict['type'] = 'none' copy_dict['sources'] = [ large_pdb_shim_cc ] copy_dict['copies'] = [{ 'destination': shim_cc_dir, 'files': [ large_pdb_shim_cc ] }] # This is the dict for the PDB generating shim target. It depends on the # copy target. shim_suffix = 'large_pdb_shim' shim_target_name = target_name + '_' + shim_suffix full_shim_target_name = _SuffixName(t, shim_suffix) shim_dict = copy.deepcopy(base_dict) shim_dict['target_name'] = shim_target_name shim_dict['type'] = 'static_library' shim_dict['sources'] = [ shim_cc_path ] shim_dict['dependencies'] = [ full_copy_target_name ] # Set up the shim to output its PDB to the same location as the final linker # target. for config_name, config in shim_dict.get('configurations').iteritems(): pdb_path = _GetPdbPath(target_dict, config_name, vars) # A few keys that we don't want to propagate. for key in ['msvs_precompiled_header', 'msvs_precompiled_source', 'test']: config.pop(key, None) msvs = config.setdefault('msvs_settings', {}) # Update the compiler directives in the shim target. compiler = msvs.setdefault('VCCLCompilerTool', {}) compiler['DebugInformationFormat'] = '3' compiler['ProgramDataBaseFileName'] = pdb_path # Set the explicit PDB path in the appropriate configuration of the # original target. config = target_dict['configurations'][config_name] msvs = config.setdefault('msvs_settings', {}) linker = msvs.setdefault('VCLinkerTool', {}) linker['GenerateDebugInformation'] = 'true' linker['ProgramDatabaseFile'] = pdb_path # Add the new targets. They must go to the beginning of the list so that # the dependency generation works as expected in ninja. target_list.insert(0, full_copy_target_name) target_list.insert(0, full_shim_target_name) target_dicts[full_copy_target_name] = copy_dict target_dicts[full_shim_target_name] = shim_dict # Update the original target to depend on the shim target. target_dict.setdefault('dependencies', []).append(full_shim_target_name) return (target_list, target_dicts)