Re: Commit cce8d6fdb introduces file t/t5100/nul, git tree is now incompatible with Cygwin (and probably Windows)

From: Avery Pennarun <apenwarr@gmail.com>
Date: May 28, 2008, 17:52 UTC
Message-ID: <32541b130805281052o2a56f00m4b64299737d27091@mail.gmail.com>
In-Reply-To: <7vfxs2uytm.fsf@gitster.siamese.dyndns.org>

On 5/28/08, Junio C Hamano <gitster@pobox.com> wrote:

>  Please learn to think before typing, let alone sending such a message to
>  waste other people's time.

Phew, just mentioning Windows on this list seems to get people flamed.

Reading over my message, I'm not sure what I would have written to have it interpreted as being negative. I was simply suggesting that git could fail more gracefully here.

Show 5 quoted lines

>  We give the user an error message, and signal error by exiting with
>  non-zero.  You cannot have that path on the system, and we are being
>  honest about it.  It is not like we are suddenly painting the screen in
>  blue and refusing to get any more user input when you try to check out
>  such a tree.  Which part of that is _not_ surviving?

The part that's not surviving is git-bisect, which is a valid problem that resulted in someone asking to rewrite the history. Clearly rewriting the history is not a good solution, thus no good solution has yet been proposed, which is why I wrote my message.

>  The system with a *fatal* error is not git but the one that does not want
>  an not-so-unreasonable name "NUL" on it.

That is clearly true. But knowing that doesn't seem to be making this user's problem go away.

>  What alternatives do you want to implement?  Certainly not silently
>  creating "nul-garbage" file instead and pretend that nothing bad happened,
>  as that would lead to madness.

If the file failed to be created (with a warning), but we treated it as having been deleted in 'git status' instead of throwing an error, it would work much like the case sensitivity problem: the index is annoyingly dirty, but you can still get work done. I think that (perhaps with a little patch) would allow git-bisect to work.

Have fun,

Avery