diff options
Diffstat (limited to 'contrib/tools/python3')
| -rw-r--r-- | contrib/tools/python3/Lib/inspect.py | 10 | ||||
| -rw-r--r-- | contrib/tools/python3/patches/skip-realpath-for-bundled-modules.patch | 54 |
2 files changed, 59 insertions, 5 deletions
diff --git a/contrib/tools/python3/Lib/inspect.py b/contrib/tools/python3/Lib/inspect.py index d74444e27be..48f89e8171c 100644 --- a/contrib/tools/python3/Lib/inspect.py +++ b/contrib/tools/python3/Lib/inspect.py @@ -980,7 +980,7 @@ def getsourcefile(object): # return a filename found in the linecache even if it doesn't exist on disk if filename in linecache.cache: return filename - if os.path.exists(filename): + if os.path.isabs(filename) and os.path.exists(filename): return filename # only return a non-existent filename if the module has a PEP 302 loader module = getmodule(object, filename) @@ -1027,10 +1027,10 @@ def getmodule(object, _filename=None): # Have already mapped this module, so skip it continue _filesbymodname[modname] = f - f = getabsfile(module) - # Always map to the name the module knows itself by - modulesbyfile[f] = modulesbyfile[ - os.path.realpath(f)] = module.__name__ + absfile = getabsfile(module) + modulesbyfile[absfile] = module.__name__ + if os.path.isabs(f): + modulesbyfile[os.path.realpath(absfile)] = module.__name__ if file in modulesbyfile: return sys.modules.get(modulesbyfile[file]) # Check the main module diff --git a/contrib/tools/python3/patches/skip-realpath-for-bundled-modules.patch b/contrib/tools/python3/patches/skip-realpath-for-bundled-modules.patch new file mode 100644 index 00000000000..e862d9c2069 --- /dev/null +++ b/contrib/tools/python3/patches/skip-realpath-for-bundled-modules.patch @@ -0,0 +1,54 @@ +Skip filesystem syscalls in inspect for binary-bundled modules. + +Modules bundled into an Arcadia binary (loaded via the ResourceImporter +in library/python/runtime_py3/__res.py) expose a relative __file__ such +as 'contrib/tools/python3/Lib/inspect.py'. There is no real file behind +that path, but inspect.getmodule() and inspect.getsourcefile() probe +the filesystem for it anyway: + + * getmodule() walks sys.modules and calls os.path.realpath() on each + module's path. realpath() in turn calls os.lstat() for every path + component (see posixpath.realpath), producing dozens to hundreds of + syscalls per invocation. For a relative path getabsfile() first + joins it with the current working directory, so the resolved path + depends on where the binary was launched from -- often pointing at + an unrelated file on disk. + + * getsourcefile() calls os.path.exists() on the same relative path, + which is again resolved against CWD and triggers an os.stat(). + +Both probes are useless for bundled modules: their source is served by +the PEP 302 loader, not the filesystem. We restrict the realpath() and +exists() calls to genuinely absolute paths, which is the case for +modules loaded from disk via Y_PYTHON_SOURCE_ROOT (see __res.py: those +are passed to compile() with an absolute filename, while bundled +modules keep the Arcadia-relative path baked in at compile time). + +linecache is already handled by fix-traceback.patch. + +--- a/Lib/inspect.py (index) ++++ b/Lib/inspect.py (working tree) +@@ -980,7 +980,7 @@ def getsourcefile(object): + # return a filename found in the linecache even if it doesn't exist on disk + if filename in linecache.cache: + return filename +- if os.path.exists(filename): ++ if os.path.isabs(filename) and os.path.exists(filename): + return filename + # only return a non-existent filename if the module has a PEP 302 loader + module = getmodule(object, filename) +@@ -1028,9 +1028,10 @@ def getmodule(object, _filename=None): + # Have already mapped this module, so skip it + continue + _filesbymodname[modname] = f +- f = getabsfile(module) +- # Always map to the name the module knows itself by +- modulesbyfile[f] = modulesbyfile[ +- os.path.realpath(f)] = module.__name__ ++ absfile = getabsfile(module) ++ modulesbyfile[absfile] = module.__name__ ++ if os.path.isabs(f): ++ modulesbyfile[os.path.realpath(absfile)] = module.__name__ + if file in modulesbyfile: + return sys.modules.get(modulesbyfile[file]) + # Check the main module |