RfC2822 mandates a line length limit of 998 byte + CRLF.
Python mail tools break lines at 990 byte. To prevent
that, we quoted-printable encode overly long lines.
The posixfile_nt code hits the win32 file API directly, which
essentially amounts to performing a system call for every read and
write. This is slow.
We add a C extension that lets us use a Python file object instead,
but preserve our desired POSIX-like semantics (the ability to rename
or delete a file that is being accessed).
If the C extension is not available (e.g. in a VPS environment
without a compiler), we fall back to the posixfile_nt code.
Remove hginside var and the test it relates to:
not( top or not hginside ) == (not top) and hginside, so the only case
when nothing will be yielded is when hginside is True and top is False.
Because of the returns placed upstream, this case will not happen anymore.
We can then safely remove hginside and the (if)s
When not in root repo, if we meet a .hg, bail out.
break was used, but return can be safely used instead: if we go through break
then (top or not hginside) == False -> nothing is done after the for loop
Do not yield (bool, tuple) in a subfunction and check later:
'if not bool: yield tuple'.
Instead simplify so the tuple doesnt get yielded on the first time, to avoid
systematic checks.
When we try to build manpages with xmlto and sed, but xmlto is
missing fail at the xmlto stage. Otherwise, one may run `cd doc;
make' and miss the warnings like:
xmlto: not found
sed: hg.1: No such file or directory
and end up with empty files installed as manpages.
The subprocess module is not thread safe. Spawning a thread to read
the output leads to exceptions like this when Mercurial exits:
Exception exceptions.TypeError: TypeError("'NoneType' object is not
callable",) in <bound method Popen.__del__ of <subprocess.Popen
object at 0x9ed0dcc>> ignored
The bug is already reported in the Python bug tracker:
http://bugs.python.org/issue1731717
Some modules (like revlog) would import util.sha1 as _sha1. This
defeats the purpose of having util.sha1 overwrite itself with a faster
version -- revlog would end up always calling the slow version. By
always delegating to util._fastsha1 we avoid this at the cost of an
extra (but unconditional) indirection.
This adds a change to the way that abort is processed, as it will not continue
truncating files beyond the first failure, otherwise the respective
functionality is maintained, i.e. abort will not unlink files, but rollback
will.
Co-contributor: Sune Foldager <cryo@cyanite.org>
Previously, accessing the filelinkrevs of a specific file that happens to
have already been unlinked from the filesystem, e.g. due to a partial rollback
having occurred, will trigger a KeyError being raised.
Co-contributor: Sune Foldager <cryo@cyanite.org>
All transactional methods on the transaction class have had a decorator
added that ensures the transaction is running.
Co-contributor: Sune Foldager <cryo@cyanite.org>
