durin42 added a comment.
From the list:
On Sep 1, 2017, at 02:16, Adrian Buehlmann <adr...@cadifra.com> wrote:
> Ugh. Can I reply to a phabricator notification by email?
>
> Adding gregory.sz...@gmail.com manually, as I'm not sure replaying to
> those nasty phabricator emails is going to work...
>
>> On 2017-09-01 00:32, indygreg (Gregory Szorc) wrote:
>> indygreg created this revision.
>> Herald added a subscriber: mercurial-devel.
>> Herald added a reviewer: hg-reviewers.
>
> [...]
>
>> - if os.path.isdir(f):
>> - # use EPERM because it is POSIX prescribed value, even though
>> - # unlink(2) on directories returns EISDIR on Linux
>> - raise IOError(errno.EPERM,
>> - "Unlinking directory not permitted: '%s'" % f) + # If the path
doesn't exist, raise that exception. + # If it is a directory, emulate POSIX
behavior. + try: + st = os.stat(f) + if
stat.S_ISDIR(st.st_mode): + # use EPERM because it is POSIX
prescribed value, even though + # unlink(2) on directories returns
EISDIR on Linux + raise IOError(errno.EPERM, +
"Unlinking directory not permitted: '%s'" % f) + except OSError as e: +
if e.errno == errno.ENOENT: + raise + + # In the common
case, a normal unlink will work. Try that first and fall + # back to more
complexity if and only if we need to. + try: + os.unlink(f) +
return + except (IOError, OSError) as e: + pass
>> 1. POSIX allows to unlink and rename open files. Windows has serious
>> 2. problems with doing that:
>
> Do you get an error at all, if a file, which is in open state, is unlinked?
>
> My fear is: You won't get an error, but instead the filename is blocked
> by the file being held in place by the other process, until the other
> process closes it. Which means: You already lost the game.
>
> Which would explain why we didn't do things like you propose here.
>
> See also
>
> https://www.mercurial-scm.org/wiki/UnlinkingFilesOnWindows
>
>
> (specifically, heading 2)
REPOSITORY
rHG Mercurial
REVISION DETAIL
https://phab.mercurial-scm.org/D588
To: indygreg, #hg-reviewers, quark
Cc: durin42, quark, mercurial-devel
_______________________________________________
Mercurial-devel mailing list
Mercurial-devel@mercurial-scm.org
https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel
