[Patch v7 05/14] quote_and_decode_query: new function to quote hex-decoded queries

[Mark Walters]

On Fri, 14 Dec 2012, david at tethera.net wrote:
> From: David Bremner <bremner at debian.org>
>
> The query is split into tokens, with ' ' and ':' as delimiters.  Any
> token containing some hex-escaped character is quoted according to
> Xapian rules.  This maps id:foo%20%22bar to id:"foo ""bar".
> This intentionally does not quote prefixes, so they still work as prefixes.
> ---
>  tag-util.c |   50 ++++++++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 50 insertions(+)
>
> diff --git a/tag-util.c b/tag-util.c
> index f89669a..e1181f8 100644
> --- a/tag-util.c
> +++ b/tag-util.c
> @@ -56,6 +56,56 @@ illegal_tag (const char *tag, notmuch_bool_t remove)
>      return NULL;
>  }
>  
> +static tag_parse_status_t
> +quote_and_decode_query (void *ctx, char *encoded, const char *line_for_error,
> +			char **query_string)
> +{

Would decode_and_quote_query be a better name given the order these two
happen? Also a comment describing the function would be nice.

> +    char *tok = encoded;
> +    size_t tok_len = 0;
> +    char *buf = NULL;
> +    size_t buf_len = 0;
> +    tag_parse_status_t ret = TAG_PARSE_SUCCESS;
> +
> +    *query_string = talloc_strdup (ctx, "");
> +
> +    while (*query_string &&
> +	   (tok = strtok_len (tok + tok_len, ": ", &tok_len)) != NULL) {
> +	char delim = tok[tok_len];
> +
> +	*(tok + tok_len++) = '\0';

These two look a little odd: I would prefer either array or pointer in
both cases.

> +
> +	if (strcspn (tok, "%") < tok_len - 1) {
> +	    /* something to decode */
> +	    if (hex_decode_inplace (tok) != HEX_SUCCESS) {
> +		ret = line_error (TAG_PARSE_INVALID, line_for_error,
> +				  "hex decoding of token '%s' failed", tok);
> +		goto DONE;
> +	    }
> +
> +	    if (double_quote_str (ctx, tok, &buf, &buf_len)) {
> +		ret = line_error (TAG_PARSE_OUT_OF_MEMORY,
> +				  line_for_error, "aborting");
> +		goto DONE;
> +	    }
> +	    *query_string = talloc_asprintf_append_buffer (
> +		*query_string, "%s%c", buf, delim);
> +
> +	} else {
> +	    /* This is not just an optimization, but used to preserve
> +	     * prefixes like id:, which cannot be quoted.
> +	     */
> +	    *query_string = talloc_asprintf_append_buffer (
> +		*query_string, "%s%c", tok, delim);
> +	}

What happens if a message id (for example) contains a ':'? Is a query of
the form 

id:stuff"encoded_stuff" 

acceptable? (As far as I can see from the man page ':' does not need to
be in hex.)

Best wishes

Mark


> +
> +    }
> +
> +  DONE:
> +    if (ret != TAG_PARSE_SUCCESS && *query_string)
> +	talloc_free (*query_string);
> +    return ret;
> +}
> +
>  tag_parse_status_t
>  parse_tag_line (void *ctx, char *line,
>  		tag_op_flag_t flags,
> -- 
> 1.7.10.4
>
> _______________________________________________
> notmuch mailing list
> notmuch at notmuchmail.org
> http://notmuchmail.org/mailman/listinfo/notmuch