From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from localhost (localhost [127.0.0.1]) by olra.theworths.org (Postfix) with ESMTP id 37B90431FBC for ; Sat, 15 Dec 2012 15:04:23 -0800 (PST) X-Virus-Scanned: Debian amavisd-new at olra.theworths.org X-Spam-Flag: NO X-Spam-Score: -0.7 X-Spam-Level: X-Spam-Status: No, score=-0.7 tagged_above=-999 required=5 tests=[RCVD_IN_DNSWL_LOW=-0.7] autolearn=disabled Received: from olra.theworths.org ([127.0.0.1]) by localhost (olra.theworths.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id G-Ei8KC+YenN for ; Sat, 15 Dec 2012 15:04:22 -0800 (PST) Received: from mail-lb0-f181.google.com (mail-lb0-f181.google.com [209.85.217.181]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by olra.theworths.org (Postfix) with ESMTPS id 245EC431FB6 for ; Sat, 15 Dec 2012 15:04:21 -0800 (PST) Received: by mail-lb0-f181.google.com with SMTP id ge1so3673115lbb.26 for ; Sat, 15 Dec 2012 15:04:20 -0800 (PST) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=from:to:cc:subject:in-reply-to:references:user-agent:date :message-id:mime-version:content-type:x-gm-message-state; bh=dSE00KTRCDMypaLFoiYsbbD6j+ynuW0qktf2QZ2UKuI=; b=Wuk+6xPntwbaFN10NUMFADmBpCjYiLLhWgqbUAvTjWgtFDfFnGI2fvrYOL34V1NJgL oDApPh/4RrvGtTqF8bxvbVO4ZiJdXaRZpuffo5ZeH6C3A0TWKT71Rahf93JfM9ZAT0KE 9D3kTRNoZdu0Bk8GIWuH9NYQFgZgRJ+syKSMIaVxuR89bCcG9I8ligJ4ZtTOEoi1jNxb qCFZ9miRnV4VYilsMwwUcamQSI+36+2ltcpyVEjlkHZOS0y6NRe4Y+Gf4H1T3qY5Ebr2 hbcx39dwkANXxX5gZGPnHV7ZW7xeXSRmCwcL6R8fDsi3zJzAMhn0BBE7h69QW1fdeORJ +u9Q== Received: by 10.112.11.33 with SMTP id n1mr4120732lbb.18.1355612660546; Sat, 15 Dec 2012 15:04:20 -0800 (PST) Received: from localhost (dsl-hkibrasgw4-50df51-27.dhcp.inet.fi. [80.223.81.27]) by mx.google.com with ESMTPS id er8sm3274601lbb.9.2012.12.15.15.04.18 (version=SSLv3 cipher=OTHER); Sat, 15 Dec 2012 15:04:19 -0800 (PST) From: Jani Nikula To: david@tethera.net, notmuch@notmuchmail.org Subject: Re: [Patch v7 06/14] notmuch-restore: move query handling for batch restore to parser In-Reply-To: <1355492062-7546-7-git-send-email-david@tethera.net> References: <1355492062-7546-1-git-send-email-david@tethera.net> <1355492062-7546-7-git-send-email-david@tethera.net> User-Agent: Notmuch/0.14+138~g7041c56 (http://notmuchmail.org) Emacs/23.4.1 (i686-pc-linux-gnu) Date: Sun, 16 Dec 2012 01:04:17 +0200 Message-ID: <87vcc2q5n2.fsf@nikula.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Gm-Message-State: ALoCoQlmkEtCMvMzHRLTDsHeT71XJjDsXmLmi5n7kAtEziatui3xTZQhIrKfr1X+A/9pnsghwaHL Cc: David Bremner X-BeenThere: notmuch@notmuchmail.org X-Mailman-Version: 2.1.13 Precedence: list List-Id: "Use and development of the notmuch mail system." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 15 Dec 2012 23:04:23 -0000 On Fri, 14 Dec 2012, david@tethera.net wrote: > From: David Bremner > > We are able to detect more errors by looking at the string before it > is hex-decoded. We also need this to avoid the query quoting for more > general queries (to be written) that will mess up raw message-ids. > --- > notmuch-restore.c | 18 +----------------- > tag-util.c | 26 ++++++++++++++++++++------ > tag-util.h | 5 ++++- > test/dump-restore | 5 ++--- > 4 files changed, 27 insertions(+), 27 deletions(-) > > diff --git a/notmuch-restore.c b/notmuch-restore.c > index 40596a8..112f2f3 100644 > --- a/notmuch-restore.c > +++ b/notmuch-restore.c > @@ -208,24 +208,8 @@ notmuch_restore_command (unused (void *ctx), int argc, char *argv[]) > if (input_format == DUMP_FORMAT_SUP) { > ret = parse_sup_line (ctx, line, &query_string, tag_ops); > } else { > - ret = parse_tag_line (ctx, line, TAG_FLAG_BE_GENEROUS, > + ret = parse_tag_line (ctx, line, TAG_FLAG_BE_GENEROUS | TAG_FLAG_ID_ONLY, > &query_string, tag_ops); I realize that we've screwed up a bit here already in the restore series. parse_sup_line() allocates query_string for each input line, but doesn't free it during parsing. parse_tag_line() sets query_string to point to the query string within line. And now it gets worse when query_string is allocated vs. set to point to another buffer depending on TAG_FLAG_ID_ONLY, so it's not an easy interface to use. > - > - if (ret == 0) { > - if (strncmp ("id:", query_string, 3) != 0) { > - fprintf (stderr, "Warning: unsupported query: %s\n", query_string); > - continue; > - } > - /* delete id: from front of string; tag_message > - * expects a raw message-id. > - * > - * XXX: Note that query string id:foo and bar will be > - * interpreted as a message id "foo and bar". This > - * should eventually be fixed to give a better error > - * message. > - */ > - query_string = query_string + 3; > - } > } > > if (ret > 0) > diff --git a/tag-util.c b/tag-util.c > index e1181f8..8fea76c 100644 > --- a/tag-util.c > +++ b/tag-util.c > @@ -201,14 +201,28 @@ parse_tag_line (void *ctx, char *line, > } > > /* tok now points to the query string */ > - if (hex_decode_inplace (tok) != HEX_SUCCESS) { > - ret = line_error (TAG_PARSE_INVALID, line_for_error, > - "hex decoding of query %s failed", tok); > - goto DONE; > + if (flags & TAG_FLAG_ID_ONLY) { > + /* this is under the assumption that any whitespace in the > + * message-id must be hex-encoded. The check is probably not > + * perfect for exotic unicode whitespace; as fallback the > + * search for strange message-ids will fail */ > + if ((strncmp ("id:", tok, 3) != 0) || The current wording is, "Any characters in and MAY be hex encoded with %NN...", but the above no longer matches that, as "id:" must not be encoded. In the interest of keeping the documentation concise, I think you should move the above check after hex_decode_inplace(), but keep the below check here. That should do it, right? > + (strcspn (tok, " \t") < strlen (tok))) { > + ret = line_error (TAG_PARSE_INVALID, line_for_error, > + "query '%s' is not 'id:'", tok); > + goto DONE; > + } > + if (hex_decode_inplace (tok) != HEX_SUCCESS) { > + ret = line_error (TAG_PARSE_INVALID, line_for_error, > + "hex decoding of query %s failed", tok); > + goto DONE; > + } > + /* skip 'id:' */ > + *query_string = tok + 3; > + } else { > + ret = quote_and_decode_query (ctx, tok, line_for_error, query_string); > } It's not pretty that query_string gets allocated or not depending on a flag that doesn't seem to have anything to do with that. I don't have a good suggestion now, though, and I'd like to keep the restore path like it is, without allocation. > > - *query_string = tok; > - > DONE: > talloc_free (line_for_error); > return ret; > diff --git a/tag-util.h b/tag-util.h > index 2889736..7674051 100644 > --- a/tag-util.h > +++ b/tag-util.h > @@ -26,7 +26,10 @@ typedef enum { > /* Accept strange tags that might be user error; > * intended for use by notmuch-restore. > */ > - TAG_FLAG_BE_GENEROUS = (1 << 3) > + TAG_FLAG_BE_GENEROUS = (1 << 3), > + > + /* Query consists of a single id:$message-id */ > + TAG_FLAG_ID_ONLY = (1 << 4) > > } tag_op_flag_t; > > diff --git a/test/dump-restore b/test/dump-restore > index 6a989b6..eb7933a 100755 > --- a/test/dump-restore > +++ b/test/dump-restore > @@ -199,19 +199,18 @@ a > # the next non-comment line should report an an empty tag error for > # batch tagging, but not for restore > + +e -- id:20091117232137.GA7669@griffis1.net > -# highlight the sketchy id parsing; this should be last > +g -- id:foo and bar > EOF > > cat < EXPECTED > -Warning: unsupported query: a > +Warning: query 'a' is not 'id:' [a] > Warning: no query string [+0] > Warning: no query string [+a +b] > Warning: missing query string [+a +b ] > Warning: no query string after -- [+c +d --] > Warning: hex decoding of tag %zz failed [+%zz -- id:whatever] > Warning: hex decoding of query id:%yy failed [+e +f id:%yy] > -Warning: cannot apply tags to missing message: foo and bar > +Warning: query 'id:foo and bar' is not 'id:' [+g -- id:foo and bar] > EOF > > test_expect_equal_file EXPECTED OUTPUT > -- > 1.7.10.4 > > _______________________________________________ > notmuch mailing list > notmuch@notmuchmail.org > http://notmuchmail.org/mailman/listinfo/notmuch