From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:45457) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1f9RVv-0005Or-DV for qemu-devel@nongnu.org; Fri, 20 Apr 2018 04:35:28 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1f9RVs-0002US-9B for qemu-devel@nongnu.org; Fri, 20 Apr 2018 04:35:27 -0400 Received: from mx3-rdu2.redhat.com ([66.187.233.73]:37532 helo=mx1.redhat.com) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1f9RVs-0002U1-4f for qemu-devel@nongnu.org; Fri, 20 Apr 2018 04:35:24 -0400 From: Markus Armbruster References: <20180419223151.623174-1-eblake@redhat.com> <20180419223151.623174-3-eblake@redhat.com> Date: Fri, 20 Apr 2018 10:35:14 +0200 In-Reply-To: <20180419223151.623174-3-eblake@redhat.com> (Eric Blake's message of "Thu, 19 Apr 2018 17:31:51 -0500") Message-ID: <87d0yu1dj1.fsf@dusky.pond.sub.org> MIME-Version: 1.0 Content-Type: text/plain Subject: Re: [Qemu-devel] [PATCH v2 2/2] qapi: Rename .json to .qapi List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eric Blake Cc: qemu-devel@nongnu.org, Juan Quintela , Jason Wang , Michael Roth , Gerd Hoffmann , =?utf-8?Q?Marc-Andr=C3=A9?= Lureau , Paolo Bonzini , "Dr. David Alan Gilbert" Eric Blake writes: > Our QAPI input files aren't quite true JSON (comments being one obvious > difference); naming things .json doesn't make that obvious. Use a new > suffix .qapi that makes it easy to identify our files. I agree that .json is sub-optimal. Evidence: we add Emacs file variables to get Python mode, because that's less wrong than JavaScript mode. Except for "trivial" files, where we can't be bothered. The rename lessens the human confusion, but increases the Editor confusion somewhat: we get Fundamental mode for "trivial" files, which is more wrong than JavaScript mode. One way to avoid that would be adding file variables to all files. Permit me to digress. I figure the (sane part of the) rationale for picking JSON as programming language syntax was "do not invent syntax, do profit from existing tooling". Except JSON is unfit for the purpose as is; you need at least comments. Our "fixes" for "unfit" turned the thing into an original work, destroying the "profit" part. To paraphrase a German saying, what we built with our hands we upset with the ass. Too late to correct, I guess, but I'm open to different opinions. [...]