Mailing-List: contact zsh-workers-help@zsh.org; run by ezmlm
Precedence: bulk
From: Bart Schaefer <schaefer@brasslantern.com>
Message-Id: <160701090529.ZM9013@torch.brasslantern.com>
Date: Fri, 1 Jul 2016 09:05:29 -0700
In-Reply-To: <20160701051106.GA31422@tarsus.local2>
Comments: In reply to Daniel Shahaf <d.s@daniel.shahaf.name>
        "Re: [PATCH v2 2/3] Fix the ':A' word modifier on paths with '..' components." (Jul  1,  5:11am)
References: <20160613085218.GA9572@tarsus.local2> 
	<1466474004-4669-1-git-send-email-danielsh@tarsus.local2> 
	<1466474004-4669-2-git-send-email-danielsh@tarsus.local2> 
	<CAHYJk3QPM+wKmzoT5jhPD90e=kYpdT0Yf2hNXcZd9woUSov+Mg@mail.gmail.com> 
	<20160625162807.GA9840@tarsus.local2> 
	<CAHYJk3SrjHP3mT+c7xWH49ozhh18WHvLC5LRUzoZwWnkSSv8fA@mail.gmail.com> 
	<20160627002031.GA20366@tarsus.local2> 
	<160628074851.ZM26955@torch.brasslantern.com> 
	<20160701051106.GA31422@tarsus.local2>
To: zsh-workers@zsh.org
Subject: Re: [PATCH v2 2/3] Fix the ':A' word modifier on paths with '..' components.
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

On Jul 1,  5:11am, Daniel Shahaf wrote:
} Subject: Re: [PATCH v2 2/3] Fix the ':A' word modifier on paths with '..' 
}
} What's the use-case for the "resolve '..' before symlinks" behaviour?
} 
} Was it an intentional design feature, or simply a documented bug?
} (Honest question.)

PWS will have to weigh in on that one.  The use-case may simply have
been the intention to make :A a superset of :a.

} (Perhaps instead of a new option, a new syntax; e.g., have $foo:A retain
} its meaning and $foo:A:A have the new meaning.  Or make :A take an
} optional argument.)

Not ideal IMO because either would make :A differ from all other history
modifiers in this respect.  I.e. :A:A should mean to resolve to real
path and then resolve to real path again, like :h:h means to chop two
elements off the tail (and thus the second :A should be a no-op like
repeating :t); and no other has an optional argument.  However, worth
considering.

Sorry to just sort of leave things there.  I don't have a strong
objection to changing the :A default given the :a:A equivalence, and
that seems the least confusing final outcome if we conclude that the
likelihood of breaking scripts is small, but I would like to reach a
consensus decision.