%PDF- %PDF-
Direktori : /lib/calibre/calibre/utils/ |
Current File : //lib/calibre/calibre/utils/filenames.py |
''' Make strings safe for use as ASCII filenames, while trying to preserve as much meaning as possible. ''' import errno import os import shutil import time from math import ceil from contextlib import suppress, closing from calibre import force_unicode, isbytestring, prints, sanitize_file_name from calibre.constants import ( filesystem_encoding, iswindows, preferred_encoding, ismacos ) from calibre.utils.localization import get_udc from polyglot.builtins import iteritems, itervalues def ascii_text(orig): udc = get_udc() try: ascii = udc.decode(orig) except Exception: if isinstance(orig, str): orig = orig.encode('ascii', 'replace') ascii = orig.decode(preferred_encoding, 'replace') if isinstance(ascii, bytes): ascii = ascii.decode('ascii', 'replace') return ascii def ascii_filename(orig, substitute='_'): if isinstance(substitute, bytes): substitute = substitute.decode(filesystem_encoding) orig = ascii_text(orig).replace('?', '_') ans = ''.join(x if ord(x) >= 32 else substitute for x in orig) return sanitize_file_name(ans, substitute=substitute) def shorten_component(s, by_what): l = len(s) if l < by_what: return s l = (l - by_what)//2 if l <= 0: return s return s[:l] + s[-l:] def limit_component(x, limit=254): # windows and macs use ytf-16 codepoints for length, linux uses arbitrary # binary data, but we will assume utf-8 filename_encoding_for_length = 'utf-16' if iswindows or ismacos else 'utf-8' def encoded_length(): q = x if isinstance(x, bytes) else x.encode(filename_encoding_for_length) return len(q) while encoded_length() > limit: delta = encoded_length() - limit x = shorten_component(x, max(2, delta // 2)) return x def shorten_components_to(length, components, more_to_take=0, last_has_extension=True): components = [limit_component(cx) for cx in components] filepath = os.sep.join(components) extra = len(filepath) - (length - more_to_take) if extra < 1: return components deltas = [] for x in components: pct = len(x)/float(len(filepath)) deltas.append(int(ceil(pct*extra))) ans = [] for i, x in enumerate(components): delta = deltas[i] if delta > len(x): r = x[0] if x is components[-1] else '' else: if last_has_extension and x is components[-1]: b, e = os.path.splitext(x) if e == '.': e = '' r = shorten_component(b, delta)+e if r.startswith('.'): r = x[0]+r else: r = shorten_component(x, delta) r = r.strip() if not r: r = x.strip()[0] if x.strip() else 'x' ans.append(r) if len(os.sep.join(ans)) > length: return shorten_components_to(length, components, more_to_take+2) return ans def find_executable_in_path(name, path=None): if path is None: path = os.environ.get('PATH', '') exts = '.exe .cmd .bat'.split() if iswindows and not name.endswith('.exe') else ('',) path = path.split(os.pathsep) for x in path: for ext in exts: q = os.path.abspath(os.path.join(x, name)) + ext if os.access(q, os.X_OK): return q def is_case_sensitive(path): ''' Return True if the filesystem is case sensitive. path must be the path to an existing directory. You must have permission to create and delete files in this directory. The results of this test apply to the filesystem containing the directory in path. ''' is_case_sensitive = False if not iswindows: name1, name2 = ('calibre_test_case_sensitivity.txt', 'calibre_TesT_CaSe_sensitiVitY.Txt') f1, f2 = os.path.join(path, name1), os.path.join(path, name2) with suppress(OSError): os.remove(f1) open(f1, 'w').close() is_case_sensitive = not os.path.exists(f2) os.remove(f1) return is_case_sensitive def case_preserving_open_file(path, mode='wb', mkdir_mode=0o777): ''' Open the file pointed to by path with the specified mode. If any directories in path do not exist, they are created. Returns the opened file object and the path to the opened file object. This path is guaranteed to have the same case as the on disk path. For case insensitive filesystems, the returned path may be different from the passed in path. The returned path is always unicode and always an absolute path. If mode is None, then this function assumes that path points to a directory and return the path to the directory as the file object. mkdir_mode specifies the mode with which any missing directories in path are created. ''' if isbytestring(path): path = path.decode(filesystem_encoding) path = os.path.abspath(path) sep = force_unicode(os.sep, 'ascii') if path.endswith(sep): path = path[:-1] if not path: raise ValueError('Path must not point to root') components = path.split(sep) if not components: raise ValueError('Invalid path: %r'%path) cpath = sep if iswindows: # Always upper case the drive letter and add a trailing slash so that # the first os.listdir works correctly cpath = components[0].upper() + sep bdir = path if mode is None else os.path.dirname(path) if not os.path.exists(bdir): os.makedirs(bdir, mkdir_mode) # Walk all the directories in path, putting the on disk case version of # the directory into cpath dirs = components[1:] if mode is None else components[1:-1] for comp in dirs: cdir = os.path.join(cpath, comp) cl = comp.lower() try: candidates = [c for c in os.listdir(cpath) if c.lower() == cl] except: # Dont have permission to do the listdir, assume the case is # correct as we have no way to check it. pass else: if len(candidates) == 1: cdir = os.path.join(cpath, candidates[0]) # else: We are on a case sensitive file system so cdir must already # be correct cpath = cdir if mode is None: ans = fpath = cpath else: fname = components[-1] ans = lopen(os.path.join(cpath, fname), mode) # Ensure file and all its metadata is written to disk so that subsequent # listdir() has file name in it. I don't know if this is actually # necessary, but given the diversity of platforms, best to be safe. ans.flush() os.fsync(ans.fileno()) cl = fname.lower() try: candidates = [c for c in os.listdir(cpath) if c.lower() == cl] except OSError: # The containing directory, somehow disappeared? candidates = [] if len(candidates) == 1: fpath = os.path.join(cpath, candidates[0]) else: # We are on a case sensitive filesystem fpath = os.path.join(cpath, fname) return ans, fpath def windows_get_fileid(path): ''' The fileid uniquely identifies actual file contents (it is the same for all hardlinks to a file). Similar to inode number on linux. ''' from calibre_extensions.winutil import get_file_id if isbytestring(path): path = path.decode(filesystem_encoding) with suppress(OSError): return get_file_id(path) def samefile_windows(src, dst): samestring = (os.path.normcase(os.path.abspath(src)) == os.path.normcase(os.path.abspath(dst))) if samestring: return True a, b = windows_get_fileid(src), windows_get_fileid(dst) if a is None and b is None: return False return a == b def samefile(src, dst): ''' Check if two paths point to the same actual file on the filesystem. Handles symlinks, case insensitivity, mapped drives, etc. Returns True iff both paths exist and point to the same file on disk. Note: On windows will return True if the two string are identical (up to case) even if the file does not exist. This is because I have no way of knowing how reliable the GetFileInformationByHandle method is. ''' if iswindows: return samefile_windows(src, dst) if hasattr(os.path, 'samefile'): # Unix try: return os.path.samefile(src, dst) except OSError: return False # All other platforms: check for same pathname. samestring = (os.path.normcase(os.path.abspath(src)) == os.path.normcase(os.path.abspath(dst))) return samestring def windows_get_size(path): ''' On windows file sizes are only accurately stored in the actual file, not in the directory entry (which could be out of date). So we open the file, and get the actual size. ''' from calibre_extensions import winutil if isbytestring(path): path = path.decode(filesystem_encoding) with closing(winutil.create_file( path, 0, winutil.FILE_SHARE_READ | winutil.FILE_SHARE_WRITE | winutil.FILE_SHARE_DELETE, winutil.OPEN_EXISTING, 0) ) as h: return winutil.get_file_size(h) def windows_hardlink(src, dest): from calibre_extensions import winutil winutil.create_hard_link(dest, src) src_size = os.path.getsize(src) # We open and close dest, to ensure its directory entry is updated # see http://blogs.msdn.com/b/oldnewthing/archive/2011/12/26/10251026.aspx for i in range(10): # If we are on a network filesystem, we have to wait for some indeterminate time, since # network file systems are the best thing since sliced bread try: if windows_get_size(dest) == src_size: return except OSError: pass time.sleep(0.3) sz = windows_get_size(dest) if sz != src_size: msg = f'Creating hardlink from {src} to {dest} failed: %s' raise OSError(msg % ('hardlink size: %d not the same as source size' % sz)) def windows_fast_hardlink(src, dest): from calibre_extensions import winutil winutil.create_hard_link(dest, src) ssz, dsz = windows_get_size(src), windows_get_size(dest) if ssz != dsz: msg = f'Creating hardlink from {src} to {dest} failed: %s' raise OSError(msg % ('hardlink size: %d not the same as source size: %s' % (dsz, ssz))) def windows_nlinks(path): from calibre_extensions import winutil if isbytestring(path): path = path.decode(filesystem_encoding) return winutil.nlinks(path) class WindowsAtomicFolderMove: ''' Move all the files inside a specified folder in an atomic fashion, preventing any other process from locking a file while the operation is incomplete. Raises an IOError if another process has locked a file before the operation starts. Note that this only operates on the files in the folder, not any sub-folders. ''' def __init__(self, path): from collections import defaultdict from calibre_extensions import winutil self.handle_map = {} if isbytestring(path): path = path.decode(filesystem_encoding) if not os.path.exists(path): return names = os.listdir(path) name_to_fileid = {x:windows_get_fileid(os.path.join(path, x)) for x in names} fileid_to_names = defaultdict(set) for name, fileid in iteritems(name_to_fileid): fileid_to_names[fileid].add(name) for x in names: f = os.path.normcase(os.path.abspath(os.path.join(path, x))) if not os.path.isfile(f): continue with suppress(OSError): # Ensure the file is not read-only winutil.set_file_attributes(f, winutil.FILE_ATTRIBUTE_NORMAL) try: h = winutil.create_file(f, winutil.GENERIC_READ, winutil.FILE_SHARE_DELETE, winutil.OPEN_EXISTING, winutil.FILE_FLAG_SEQUENTIAL_SCAN) except OSError as e: if e.winerror == winutil.ERROR_SHARING_VIOLATION: # The file could be a hardlink to an already opened file, # in which case we use the same handle for both files fileid = name_to_fileid[x] found = False if fileid is not None: for other in fileid_to_names[fileid]: other = os.path.normcase(os.path.abspath(os.path.join(path, other))) if other in self.handle_map: self.handle_map[f] = self.handle_map[other] found = True break if found: continue self.close_handles() if e.winerror == winutil.ERROR_SHARING_VIOLATION: err = IOError(errno.EACCES, _('File is open in another process')) err.filename = f raise err prints('CreateFile failed for: %r' % f) raise except: self.close_handles() prints('CreateFile failed for: %r' % f) raise self.handle_map[f] = h def copy_path_to(self, path, dest): from calibre_extensions import winutil handle = None for p, h in self.handle_map.items(): if samefile_windows(path, p): handle = h break if handle is None: if os.path.exists(path): raise ValueError('The file %r did not exist when this move' ' operation was started'%path) else: raise ValueError('The file %r does not exist'%path) with suppress(OSError): windows_hardlink(path, dest) return winutil.set_file_pointer(handle, 0, winutil.FILE_BEGIN) with lopen(dest, 'wb') as f: sz = 1024 * 1024 while True: raw = winutil.read_file(handle, sz) if not raw: break f.write(raw) def release_file(self, path): ' Release the lock on the file pointed to by path. Will also release the lock on any hardlinks to path ' key = None for p, h in iteritems(self.handle_map): if samefile_windows(path, p): key = (p, h) break if key is not None: key[1].close() remove = [f for f, h in iteritems(self.handle_map) if h is key[1]] for x in remove: self.handle_map.pop(x) def close_handles(self): for h in itervalues(self.handle_map): h.close() self.handle_map = {} def delete_originals(self): from calibre_extensions import winutil for path in self.handle_map: winutil.delete_file(path) self.close_handles() def hardlink_file(src, dest): if iswindows: windows_hardlink(src, dest) return os.link(src, dest) def nlinks_file(path): ' Return number of hardlinks to the file ' if iswindows: return windows_nlinks(path) return os.stat(path).st_nlink if iswindows: from calibre_extensions.winutil import move_file def rename_file(a, b): if isinstance(a, bytes): a = os.fsdecode(a) if isinstance(b, bytes): b = os.fsdecode(b) move_file(a, b) def retry_on_fail(func, *args, count=10, sleep_time=0.2): for i in range(count): try: func(*args) break except OSError: if i > count - 2: raise # Try the operation repeatedly in case something like a virus # scanner has opened one of the files (I love windows) time.sleep(sleep_time) def atomic_rename(oldpath, newpath): '''Replace the file newpath with the file oldpath. Can fail if the files are on different volumes. If succeeds, guaranteed to be atomic. newpath may or may not exist. If it exists, it is replaced. ''' if iswindows: retry_on_fail(rename_file, oldpath, newpath) else: os.rename(oldpath, newpath) def remove_dir_if_empty(path, ignore_metadata_caches=False): ''' Remove a directory if it is empty or contains only the folder metadata caches from different OSes. To delete the folder if it contains only metadata caches, set ignore_metadata_caches to True.''' try: os.rmdir(path) except OSError as e: if e.errno == errno.ENOTEMPTY or len(os.listdir(path)) > 0: # Some linux systems appear to raise an EPERM instead of an # ENOTEMPTY, see https://bugs.launchpad.net/bugs/1240797 if ignore_metadata_caches: try: found = False for x in os.listdir(path): if x.lower() in {'.ds_store', 'thumbs.db'}: found = True x = os.path.join(path, x) if os.path.isdir(x): import shutil shutil.rmtree(x) else: os.remove(x) except Exception: # We could get an error, if, for example, windows has locked Thumbs.db found = False if found: remove_dir_if_empty(path) return raise expanduser = os.path.expanduser def format_permissions(st_mode): import stat for func, letter in (x.split(':') for x in 'REG:- DIR:d BLK:b CHR:c FIFO:p LNK:l SOCK:s'.split()): if getattr(stat, 'S_IS' + func)(st_mode): break else: letter = '?' rwx = ('---', '--x', '-w-', '-wx', 'r--', 'r-x', 'rw-', 'rwx') ans = [letter] + list(rwx[(st_mode >> 6) & 7]) + list(rwx[(st_mode >> 3) & 7]) + list(rwx[(st_mode & 7)]) if st_mode & stat.S_ISUID: ans[3] = 's' if (st_mode & stat.S_IXUSR) else 'S' if st_mode & stat.S_ISGID: ans[6] = 's' if (st_mode & stat.S_IXGRP) else 'l' if st_mode & stat.S_ISVTX: ans[9] = 't' if (st_mode & stat.S_IXUSR) else 'T' return ''.join(ans) def copyfile(src, dest): shutil.copyfile(src, dest) try: shutil.copystat(src, dest) except Exception: pass def get_hardlink_function(src, dest): if not iswindows: return os.link from calibre_extensions import winutil root = dest[0] + ':\\' if src[0].lower() == dest[0].lower() and hasattr(winutil, 'supports_hardlinks') and winutil.supports_hardlinks(root): return windows_fast_hardlink def copyfile_using_links(path, dest, dest_is_dir=True, filecopyfunc=copyfile): path, dest = os.path.abspath(path), os.path.abspath(dest) if dest_is_dir: dest = os.path.join(dest, os.path.basename(path)) hardlink = get_hardlink_function(path, dest) try: hardlink(path, dest) except Exception: filecopyfunc(path, dest) def copytree_using_links(path, dest, dest_is_parent=True, filecopyfunc=copyfile): path, dest = os.path.abspath(path), os.path.abspath(dest) if dest_is_parent: dest = os.path.join(dest, os.path.basename(path)) hardlink = get_hardlink_function(path, dest) try: os.makedirs(dest) except OSError as e: if e.errno != errno.EEXIST: raise for dirpath, dirnames, filenames in os.walk(path): base = os.path.relpath(dirpath, path) dest_base = os.path.join(dest, base) for dname in dirnames: try: os.mkdir(os.path.join(dest_base, dname)) except OSError as e: if e.errno != errno.EEXIST: raise for fname in filenames: src, df = os.path.join(dirpath, fname), os.path.join(dest_base, fname) try: hardlink(src, df) except Exception: filecopyfunc(src, df) rmtree = shutil.rmtree if iswindows: long_path_prefix = '\\\\?\\' def make_long_path_useable(path): if len(path) > 200 and os.path.isabs(path) and not path.startswith(long_path_prefix): path = long_path_prefix + os.path.normpath(path) return path else: def make_long_path_useable(path): return path