Saving to disk: Fix errors on Linux/macOS if the title/authors are long enough to make individual path components larger than 255 characters. Fixes #1807525 [[Enhancement] Filename length during saving on disk](https://bugs.launchpad.net/calibre/+bug/1807525)

This commit is contained in:
Kovid Goyal 2018-12-29 16:20:14 +05:30
parent 19b02bd063
commit 2e33cb5cb3
No known key found for this signature in database
GPG Key ID: 06BC317B515ACE7C
2 changed files with 20 additions and 15 deletions

View File

@ -12,7 +12,7 @@ from calibre.constants import DEBUG
from calibre.db.errors import NoSuchFormat
from calibre.utils.config import Config, StringConfig, tweaks
from calibre.utils.formatter import TemplateFormatter
from calibre.utils.filenames import shorten_components_to, supports_long_names, ascii_filename
from calibre.utils.filenames import shorten_components_to, ascii_filename
from calibre.constants import preferred_encoding
from calibre.ebooks.metadata import fmt_sidx
from calibre.ebooks.metadata import title_sort
@ -378,7 +378,7 @@ def sanitize_args(root, opts):
root = os.path.abspath(root)
opts.template = preprocess_template(opts.template)
length = 1000 if supports_long_names(root) else 240
length = 240
length -= len(root)
if length < 5:
raise ValueError('%r is too long.'%root)

View File

@ -11,7 +11,7 @@ from math import ceil
from calibre import force_unicode, isbytestring, prints, sanitize_file_name
from calibre.constants import (
filesystem_encoding, iswindows, plugins, preferred_encoding
filesystem_encoding, iswindows, plugins, preferred_encoding, isosx
)
from calibre.utils.localization import get_udc
@ -38,18 +38,6 @@ def ascii_filename(orig, substitute='_'):
return sanitize_file_name(''.join(ans), substitute=substitute)
def supports_long_names(path):
t = ('a'*300)+'.txt'
try:
p = os.path.join(path, t)
open(p, 'wb').close()
os.remove(p)
except:
return False
else:
return True
def shorten_component(s, by_what):
l = len(s)
if l < by_what:
@ -60,7 +48,24 @@ def shorten_component(s, by_what):
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 isosx 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: