Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT cygwin DOT com Mail-Followup-To: cygwin AT cygwin DOT com Delivered-To: mailing list cygwin AT cygwin DOT com From: ericblake AT comcast DOT net (Eric Blake) To: cygwin AT cygwin DOT com Subject: Re: Please test 2005-Apr-12 snapshot Date: Tue, 12 Apr 2005 23:22:00 +0000 Message-Id: <041220052322.20346.425C5818000A08C200004F7A22058891160A050E040D0C079D0A@comcast.net> X-Authenticated-Sender: ZXJpY2JsYWtlQGNvbWNhc3QubmV0 > I'd like to ask people to test the latest snapshot as of today. It contains > two patches, one of them with some impact on how timestamps are used. > > The first patch is less important, it should just stop floppy access when > sync is called on Win2K boxes. Doesn't seem to be any problem with this on my Win2k box at work (I still need to test XP and 98 when I get home) > > The second patch is this: > > In 1.5.13 and 1.5.14 we introduced touching the CreationTime stamp in > a way, which simulates a POSIX ctime using the CreationTime stamp. > This resulted in some complaints, the most important one that native > Windows applications might misbehave because of strange CreationTimes. > > The snapshot now contains a patch which changes the ctime handling as > follows: > > - Windows NT supports a fourth timestamp which is inaccessible from the > Win32 API. The NTFS filesystem actually implements it. It behaves > as a ctime in a POSIX-like fashion. Cygwin's st_ctime stat member now > contains this ChangeTime, if it's available. > > - Any other file system, which doesn't support the ChangeTime stamp > uses the LastWriteTime stamp as ctime. This comes relatively close > to the way ctime behaves in POSIX. > > - The CreationTime stamp is neither read nor changed programatically by > Cygwin now. This should solve the aforementioned problems for native > Win32 applications. > > Please download from http://www.cygwin.com/snapshots/ and test. > > The second patch caused a regression in the coreutils test suite: $ touch -c none touch: setting times of `none': Permission denied $ echo $? 1 This used to work in 1.5.14, and POSIX requires that it is a nop with exit status 0. -- Eric Blake -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/