coq-club AT inria.fr
Subject: The Coq mailing list
List archive
- From: Thomas Schilling <nominolo AT googlemail.com>
- To: Adam Chlipala <adam AT chlipala.net>
- Cc: coq-club AT inria.fr
- Subject: Re: [Coq-Club] Problem with definition from CPDT
- Date: Tue, 15 Feb 2011 13:36:38 +0000
- Domainkey-signature: a=rsa-sha1; c=nofws; d=googlemail.com; s=gamma; h=subject:mime-version:content-type:from:in-reply-to:date:cc :content-transfer-encoding:message-id:references:to:x-mailer; b=VUBEv+qVL0iWDMzkEbMxcEWSSfD30wkFlmCBlSyXmI6ynLtpK5KwdxfIphebF1xXgM xJfG4qmkjWsOSxx+LD71rXpypZdIGsTHDvM1e4EjpsN/4TCduGj9JT5PXYydW3weBzLA fEQbtJemSwn195ZB4t/TbFJ1XZmEomRoJYZ4k=
On 15 Feb 2011, at 13:26, Adam Chlipala wrote:
> Thomas Schilling wrote:
>> I'm working through Adam's Certified Programming with Dependent Types and
>> Coq fails to accept the following definition from the book:
>> [...]
>>
>> Is this a limitation of my version of Coq? I'm using Coq 8.2pl1 (latest
>> on MacPorts). Or am I missing something obvious?
>>
>
> The book is only tested with Coq 8.2pl2 and 8.3pl1. There were definitely
> some bugs in 8.2pl1 that prevented some of the code from type-checking,
> though I don't remember seeing your particular example mentioned before.
Ok, I managed to fix it. I used "Print nat_rect" to figure out the
definition of the auto-generated induction principles and then via trial and
error I got it to accept the following. (The changes are the "as .. return
.." annotations.)
Fixpoint nat_tree_ind' (tr : nat_tree) : P tr :=
match tr as tr0 return (P tr0) with
| NLeaf => NLeaf_case
| NNode n ls => NNode_case n ls
((fix list_nat_tree_ind (ls : list nat_tree) : All P ls :=
match ls as ls0 return All P ls0 with
| nil => I
| cons tr rest =>
conj (nat_tree_ind' tr) (list_nat_tree_ind rest)
end) ls)
end.
It indeed appears to be a limitation of my version of Coq.
BTW, thank you for making the sources for your book available. It allowed me
to create a special version for easier reading on my Kindle.
/ Thomas
- [Coq-Club] Problem with definition from CPDT, Thomas Schilling
- Re: [Coq-Club] Problem with definition from CPDT,
Adam Chlipala
- Re: [Coq-Club] Problem with definition from CPDT, Thomas Schilling
- Re: [Coq-Club] Problem with definition from CPDT,
Adam Chlipala
Archive powered by MhonArc 2.6.16.