From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on inbox.vuxu.org
X-Spam-Level: 
X-Spam-Status: No, score=0.0 required=5.0 tests=none autolearn=ham
	autolearn_force=no version=3.4.4
Received: (qmail 3929 invoked from network); 18 Nov 2022 04:34:17 -0000
Received: from 9front.inri.net (168.235.81.73)
  by inbox.vuxu.org with ESMTPUTF8; 18 Nov 2022 04:34:17 -0000
Received: from mimir.eigenstate.org ([206.124.132.107]) by 9front; Thu Nov 17 23:30:15 -0500 2022
Received: from abbatoir (pool-108-27-53-161.nycmny.fios.verizon.net [108.27.53.161])
	by mimir.eigenstate.org (OpenSMTPD) with ESMTPSA id c4d32079 (TLSv1.2:ECDHE-RSA-AES256-SHA:256:NO)
	for <9front@9front.org>;
	Thu, 17 Nov 2022 20:30:14 -0800 (PST)
Message-ID: <4F579BAA82A5A3386F286262ED2A7806@eigenstate.org>
To: 9front@9front.org
Date: Thu, 17 Nov 2022 23:30:12 -0500
From: ori@eigenstate.org
In-Reply-To: <923270033661C2AB404C7473549B4808@driusan.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
List-ID: <9front.9front.org>
List-Help: <http://lists.9front.org>
X-Glyph: ➈
X-Bullshit: RESTful API-based database 
Subject: Re: [9front] libjson can print invalid json
Reply-To: 9front@9front.org
Precedence: bulk

Quoth Dave MacFarlane <driusan@driusan.net>:
> print("%J", json); can print invalid json if a JSON string in the object has a quotation mark in it.
> 
> I just threw together this patch to try and escape the strings from printjson() if anyone's interested
> in it.
> 
> diff 30c5296f32b87d83529d772732726891e1261c9c uncommitted
> --- a/sys/src/libjson/printjson.c
> +++ b/sys/src/libjson/printjson.c
> @@ -52,6 +52,38 @@
>  }
>  
>  static int
> +printstring(Fmt *f, char *s)
> +{
> +	int slen = strlen(s);
> +	int i, nq=0;
> +	int r;
> +	char *dup, lastq;
> +	for(i = 0; i < slen; i++)
> +		if (s[i] == '"') nq++;
> +
> +	if (nq == 0)
> +		return fmtprint(f, "\"%s\"", s);
> +
> +	r = fmtprint(f, "\"");
> +
> +	dup = strdup(s);
> +	lastq = -1;
> +	for(i = 0; i < slen; i++){
> +		if(dup[i] == '"') {
> +			dup[i] = 0;
> +			r += fmtprint(f, "%s\\\"", &dup[lastq+1]);
> +			lastq = i;
> +		}
> +	}
> +
> +	if (lastq != slen-1)
> +		r += fmtprint(f, "%s", &dup[lastq+1]);
> +	r += fmtprint(f, "\"");
> +	free(dup);
> +	return r;
> +}
> +

I think we have a problem with more than just quotes; we're
not allowed any control characters, and we need to escape '\';

how's something like this (untested)?


	static int
	printstring(Fmt *f, char *s)
	{
		int n;
	
		n = 0;
		for(; *s; s++){
			switch(*s){
			case '\\':	n += fmtstrcpy(f, "\\\\");	break;
			case '\f':	n += fmtstrcpy(f, "\\f");	break;
			case '\b':	n += fmtstrcpy(f, "\\b");	break;
			case '\n':	n += fmtstrcpy(f, "\\n");	break;
			case '\r':	n += fmtstrcpy(f, "\\r");	break;
			case '\"':	n += fmtstrcpy(f, "\\\"");	break;
			default:
				if(*s < 0x20)
					n += fmtprint(f, "\\u%04x", *s);
				else
					n += fmtrune(f, *s);
			}
		}
		return n;
	}