git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jeff King <peff@peff.net>
To: Chris Rorvick <chris@rorvick.com>
Cc: Ramsay Jones <ramsay@ramsay1.demon.co.uk>, git@vger.kernel.org
Subject: Re: What's cooking in git.git (Oct 2012, #09; Mon, 29)
Date: Fri, 2 Nov 2012 05:43:00 -0400	[thread overview]
Message-ID: <20121102094259.GA28414@sigill.intra.peff.net> (raw)
In-Reply-To: <CAEUsAPb7hUViLn7V7v65r6mOqRHr+180ynRX8K9t3nuJVyePfg@mail.gmail.com>

On Thu, Nov 01, 2012 at 08:12:20PM -0500, Chris Rorvick wrote:

> > Just FYI, t9604-cvsimport-timestamps.sh is still failing for me.
> >
> > I haven't spent too long on this yet, but I had hoped that setting
> > TZ would sidestep any DST issues. (I have downloaded new tzdata, but
> > have yet to install - actually I don't really want to!). It is not
> > clear from the tzset manpage what happens if you use the DST format
> > for TZ, but you don't provide the start/end date for DST, which is
> > what this test is doing.
> >
> > Perhaps the test should use the non-DST format? e.g. "TZ=CST6 git ..."
> > Does the test really care about DST? (*if* that is indeed the problem).
> 
> It actually looks like your TZ database is fine and the problem is
> with the conversion to a struct tm.  In each case, the time is
> localized to the previous TZ value while the offset for the current TZ
> value.  For example, look at the first commit in the first test.  It
> converted the timestamp to 18:00 (CST6) while all the rest came
> through as expected.    I suspect the previous version of cvsimport
> would exhibit similar behavior with the first imported commit.  What
> is your platform?

Yeah, I think that is it. IIRC, Ramsay is on cygwin, and I noticed this
in perl 5.16's POSIX.xs:

  #ifdef WIN32

  /*
   * (1) The CRT maintains its own copy of the environment, separate from
   * the Win32API copy.
   *
   * (2) CRT getenv() retrieves from this copy. CRT putenv() updates this
   * copy, and then calls SetEnvironmentVariableA() to update the Win32API
   * copy.
   *
   * (3) win32_getenv() and win32_putenv() call GetEnvironmentVariableA() and
   * SetEnvironmentVariableA() directly, bypassing the CRT copy of the
   * environment.
   *
   * (4) The CRT strftime() "%Z" implementation calls __tzset(). That
   * calls CRT tzset(), but only the first time it is called, and in turn
   * that uses CRT getenv("TZ") to retrieve the timezone info from the CRT
   * local copy of the environment and hence gets the original setting as
   * perl never updates the CRT copy when assigning to $ENV{TZ}.
   *
   * Therefore, we need to retrieve the value of $ENV{TZ} and call CRT
   * putenv() to update the CRT copy of the environment (if it is different)
   * whenever we're about to call tzset().
   [...]

I wonder if Ramsay has an older perl that does not do this special
hackery right. I'll see if I can dig up where it first appeared.

-Peff

  reply	other threads:[~2012-11-02  9:43 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-10-29 10:21 What's cooking in git.git (Oct 2012, #09; Mon, 29) Jeff King
2012-11-01 18:40 ` Ramsay Jones
2012-11-02  1:12   ` Chris Rorvick
2012-11-02  9:43     ` Jeff King [this message]
2012-11-02  9:56       ` Jeff King
2012-11-02 21:33         ` Ramsay Jones
2012-11-04 12:29           ` Jeff King
2012-11-10 18:33             ` Ramsay Jones
2012-11-11 17:41               ` Jeff King
2012-11-13 18:47                 ` Ramsay Jones

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20121102094259.GA28414@sigill.intra.peff.net \
    --to=peff@peff.net \
    --cc=chris@rorvick.com \
    --cc=git@vger.kernel.org \
    --cc=ramsay@ramsay1.demon.co.uk \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).