From: David Bremner <david@tethera.net>
To: notmuch@notmuchmail.org
Subject: v4 sexp query parser
Date: Thu, 12 Aug 2021 10:06:57 -0700 [thread overview]
Message-ID: <20210812170728.1348333-1-david@tethera.net> (raw)
Two main user visible changes since v3.
1) I changed the option from
--query-syntax=sexp
to
--query=sexp
This is arguably a bit more ambiguous (are we specifying a query
"sexp" to search for?) but considerably less painful to type. I
previously mentioned `--squery`, but it is a bit cryptic, and collides with completing other options that start with 's'.
2) The new form (macro (param1 param2) body). This is a bit of feature
creep, but I realized it is one option for dealing with requests like id:87k0nl8nvi.fsf@linux.ibm.com.
The series still only enables the new syntax in notmuch-search and
notmuch-address. I was pondering making the query syntax a global
option handled in the same was as --uuid, but I didn't get that far
yet.
Interdiff
---------
The change to the command line option introduces a bunch of noise to
the interdiff, which I've omitted here. A full diffstat is
doc/man1/notmuch-config.rst | 5 +
doc/man7/notmuch-sexp-queries.rst | 48 ++++-
lib/parse-sexp.cc | 199 ++++++++++++++++++--
notmuch-config.c | 1 +
notmuch-search.c | 4 +-
test/T081-sexpr-search.sh | 431 ++++++++++++++++++++++++++++++-------------
test/T095-address.sh | 2 +-
The core changes are as follows:
diff --git a/lib/parse-sexp.cc b/lib/parse-sexp.cc
index e582e350..eea4d4da 100644
--- a/lib/parse-sexp.cc
+++ b/lib/parse-sexp.cc
@@ -7,6 +7,20 @@
/* _sexp is used for file scope symbols to avoid clashing with
* definitions from sexp.h */
+/* sexp_binding structs attach name to a sexp and a defining
+ * context. The latter allows lazy evaluation of parameters whose
+ * definition contains other parameters. Lazy evaluation is needed
+ * because a primary goal of macros is to change the parent field for
+ * a sexp.
+ */
+
+typedef struct sexp_binding {
+ const char *name;
+ const sexp_t *sx;
+ const struct sexp_binding *context;
+ const struct sexp_binding *next;
+} _sexp_binding_t;
+
typedef enum {
SEXP_FLAG_NONE = 0,
SEXP_FLAG_FIELD = 1 << 0,
@@ -99,12 +113,14 @@ static _sexp_prefix_t prefixes[] =
static notmuch_status_t _sexp_to_xapian_query (notmuch_database_t *notmuch,
const _sexp_prefix_t *parent,
+ const _sexp_binding_t *env,
const sexp_t *sx,
Xapian::Query &output);
static notmuch_status_t
_sexp_combine_query (notmuch_database_t *notmuch,
const _sexp_prefix_t *parent,
+ const _sexp_binding_t *env,
Xapian::Query::op operation,
Xapian::Query left,
const sexp_t *sx,
@@ -121,12 +137,13 @@ _sexp_combine_query (notmuch_database_t *notmuch,
return NOTMUCH_STATUS_SUCCESS;
}
- status = _sexp_to_xapian_query (notmuch, parent, sx, subquery);
+ status = _sexp_to_xapian_query (notmuch, parent, env, sx, subquery);
if (status)
return status;
return _sexp_combine_query (notmuch,
parent,
+ env,
operation,
Xapian::Query (operation, left, subquery),
sx->next, output);
@@ -165,6 +182,7 @@ _sexp_parse_phrase (std::string term_prefix, const char *phrase, Xapian::Query &
static notmuch_status_t
_sexp_parse_wildcard (notmuch_database_t *notmuch,
const _sexp_prefix_t *parent,
+ unused(const _sexp_binding_t *env),
std::string match,
Xapian::Query &output)
{
@@ -201,6 +219,7 @@ _sexp_parse_one_term (notmuch_database_t *notmuch, std::string term_prefix, cons
notmuch_status_t
_sexp_parse_regex (notmuch_database_t *notmuch,
const _sexp_prefix_t *prefix, const _sexp_prefix_t *parent,
+ unused(const _sexp_binding_t *env),
std::string val, Xapian::Query &output)
{
if (! parent) {
@@ -225,7 +244,7 @@ _sexp_parse_regex (notmuch_database_t *notmuch,
static notmuch_status_t
_sexp_expand_query (notmuch_database_t *notmuch,
const _sexp_prefix_t *prefix, const _sexp_prefix_t *parent,
- const sexp_t *sx, Xapian::Query &output)
+ unused(const _sexp_binding_t *env), const sexp_t *sx, Xapian::Query &output)
{
Xapian::Query subquery;
notmuch_status_t status;
@@ -236,7 +255,8 @@ _sexp_expand_query (notmuch_database_t *notmuch,
return NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
}
- status = _sexp_combine_query (notmuch, NULL, prefix->xapian_op, prefix->initial, sx, subquery);
+ status = _sexp_combine_query (notmuch, NULL, NULL, prefix->xapian_op, prefix->initial, sx,
+ subquery);
if (status)
return status;
@@ -272,7 +292,7 @@ _sexp_parse_infix (notmuch_database_t *notmuch, const sexp_t *sx, Xapian::Query
static notmuch_status_t
_sexp_parse_header (notmuch_database_t *notmuch, const _sexp_prefix_t *parent,
- const sexp_t *sx, Xapian::Query &output)
+ const _sexp_binding_t *env, const sexp_t *sx, Xapian::Query &output)
{
_sexp_prefix_t user_prefix;
@@ -287,23 +307,164 @@ _sexp_parse_header (notmuch_database_t *notmuch, const _sexp_prefix_t *parent,
parent = &user_prefix;
- return _sexp_combine_query (notmuch, parent, Xapian::Query::OP_AND, Xapian::Query::MatchAll,
+ return _sexp_combine_query (notmuch, parent, env, Xapian::Query::OP_AND, Xapian::Query::MatchAll,
sx->list->next, output);
}
+static _sexp_binding_t *
+_sexp_bind (void *ctx, const _sexp_binding_t *env, const char *name, const sexp_t *sx, const
+ _sexp_binding_t *context)
+{
+ _sexp_binding_t *binding = talloc (ctx, _sexp_binding_t);
+
+ binding->name = talloc_strdup (ctx, name);
+ binding->sx = sx;
+ binding->context = context;
+ binding->next = env;
+ return binding;
+}
+
+static notmuch_status_t
+maybe_apply_macro (notmuch_database_t *notmuch, const _sexp_prefix_t *parent,
+ const _sexp_binding_t *env, const sexp_t *sx, const sexp_t *args,
+ Xapian::Query &output)
+{
+ const sexp_t *params, *param, *arg, *body;
+ void *local = talloc_new (notmuch);
+ _sexp_binding_t *new_env = NULL;
+ notmuch_status_t status = NOTMUCH_STATUS_SUCCESS;
+
+ if (sx->list->ty != SEXP_VALUE || strcmp (sx->list->val, "macro") != 0) {
+ status = NOTMUCH_STATUS_IGNORED;
+ goto DONE;
+ }
+
+ params = sx->list->next;
+
+ if (! params || (params->ty != SEXP_LIST)) {
+ _notmuch_database_log (notmuch, "missing (possibly empty) list of arguments to macro\n");
+ return NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
+ }
+
+ body = params->next;
+
+ if (! body) {
+ _notmuch_database_log (notmuch, "missing body of macro\n");
+ status = NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
+ goto DONE;
+ }
+
+ for (param = params->list, arg = args;
+ param && arg;
+ param = param->next, arg = arg->next) {
+ if (param->ty != SEXP_VALUE || param->aty != SEXP_BASIC) {
+ _notmuch_database_log (notmuch, "macro parameters must be unquoted atoms\n");
+ status = NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
+ goto DONE;
+ }
+ new_env = _sexp_bind (local, new_env, param->val, arg, env);
+ }
+
+ if (param && ! arg) {
+ _notmuch_database_log (notmuch, "too few arguments to macro\n");
+ status = NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
+ goto DONE;
+ }
+
+ if (! param && arg) {
+ _notmuch_database_log (notmuch, "too many arguments to macro\n");
+ status = NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
+ goto DONE;
+ }
+
+ status = _sexp_to_xapian_query (notmuch, parent, new_env, body, output);
+
+ DONE:
+ if (local)
+ talloc_free (local);
+
+ return status;
+}
+
+static notmuch_status_t
+maybe_saved_squery (notmuch_database_t *notmuch, const _sexp_prefix_t *parent,
+ const _sexp_binding_t *env, const sexp_t *sx, Xapian::Query &output)
+{
+ char *key;
+ char *expansion = NULL;
+ notmuch_status_t status;
+ sexp_t *saved_sexp;
+ void *local = talloc_new (notmuch);
+ char *buf;
+
+ key = talloc_asprintf (local, "squery.%s", sx->list->val);
+ if (! key) {
+ status = NOTMUCH_STATUS_OUT_OF_MEMORY;
+ goto DONE;
+ }
+
+ status = notmuch_database_get_config (notmuch, key, &expansion);
+ if (status)
+ goto DONE;
+ if (EMPTY_STRING (expansion)) {
+ status = NOTMUCH_STATUS_IGNORED;
+ goto DONE;
+ }
+
+ buf = talloc_strdup (local, expansion);
+ /* XXX TODO: free this memory */
+ saved_sexp = parse_sexp (buf, strlen (expansion));
+ if (! saved_sexp) {
+ _notmuch_database_log (notmuch, "invalid saved s-expression query: '%s'\n", expansion);
+ status = NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
+ goto DONE;
+ }
+
+ status = maybe_apply_macro (notmuch, parent, env, saved_sexp, sx->list->next, output);
+ if (status == NOTMUCH_STATUS_IGNORED)
+ status = _sexp_to_xapian_query (notmuch, parent, env, saved_sexp, output);
+
+ DONE:
+ if (local)
+ talloc_free (local);
+
+ return status;
+}
+
+static notmuch_status_t
+_sexp_expand_param (notmuch_database_t *notmuch, const _sexp_prefix_t *parent,
+ const _sexp_binding_t *env, const char *name,
+ Xapian::Query &output)
+{
+ for (; env; env = env->next) {
+ if (strcmp (name, env->name) == 0) {
+ return _sexp_to_xapian_query (notmuch, parent, env->context, env->sx,
+ output);
+ }
+ }
+ _notmuch_database_log (notmuch, "undefined parameter %s\n", name);
+ return NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
+}
+
/* Here we expect the s-expression to be a proper list, with first
* element defining and operation, or as a special case the empty
* list */
static notmuch_status_t
-_sexp_to_xapian_query (notmuch_database_t *notmuch, const _sexp_prefix_t *parent, const sexp_t *sx,
- Xapian::Query &output)
+_sexp_to_xapian_query (notmuch_database_t *notmuch, const _sexp_prefix_t *parent,
+ const _sexp_binding_t *env, const sexp_t *sx, Xapian::Query &output)
{
+ notmuch_status_t status;
+
+ if (sx->ty == SEXP_VALUE && sx->aty == SEXP_BASIC && sx->val[0] == ',') {
+ return _sexp_expand_param (notmuch, parent, env, sx->val + 1, output);
+ }
+
if (sx->ty == SEXP_VALUE) {
std::string term_prefix = parent ? _notmuch_database_prefix (notmuch, parent->name) : "";
if (sx->aty == SEXP_BASIC && strcmp (sx->val, "*") == 0) {
- return _sexp_parse_wildcard (notmuch, parent, "", output);
+ return _sexp_parse_wildcard (notmuch, parent, env, "", output);
}
if (parent && (parent->flags & SEXP_FLAG_BOOLEAN)) {
@@ -317,7 +478,6 @@ _sexp_to_xapian_query (notmuch_database_t *notmuch, const _sexp_prefix_t *parent
Xapian::Query accumulator;
for (_sexp_prefix_t *prefix = prefixes; prefix->name; prefix++) {
if (prefix->flags & SEXP_FLAG_FIELD) {
- notmuch_status_t status;
Xapian::Query subquery;
term_prefix = _notmuch_database_prefix (notmuch, prefix->name);
status = _sexp_parse_one_term (notmuch, term_prefix, sx, subquery);
@@ -343,9 +503,18 @@ _sexp_to_xapian_query (notmuch_database_t *notmuch, const _sexp_prefix_t *parent
return NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
}
+ status = maybe_saved_squery (notmuch, parent, env, sx, output);
+ if (status != NOTMUCH_STATUS_IGNORED)
+ return status;
+
/* Check for user defined field */
if (_notmuch_string_map_get (notmuch->user_prefix, sx->list->val)) {
- return _sexp_parse_header (notmuch, parent, sx, output);
+ return _sexp_parse_header (notmuch, parent, env, sx, output);
+ }
+
+ if (strcmp (sx->list->val, "macro") == 0) {
+ _notmuch_database_log (notmuch, "macro definition not permitted here\n");
+ return NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
}
for (_sexp_prefix_t *prefix = prefixes; prefix && prefix->name; prefix++) {
@@ -381,17 +550,17 @@ _sexp_to_xapian_query (notmuch_database_t *notmuch, const _sexp_prefix_t *parent
}
if (prefix->xapian_op == Xapian::Query::OP_WILDCARD)
- return _sexp_parse_wildcard (notmuch, parent, sx->list->next->val, output);
+ return _sexp_parse_wildcard (notmuch, parent, env, sx->list->next->val, output);
if (prefix->flags & SEXP_FLAG_DO_REGEX) {
- return _sexp_parse_regex (notmuch, prefix, parent, sx->list->next->val, output);
+ return _sexp_parse_regex (notmuch, prefix, parent, env, sx->list->next->val, output);
}
if (prefix->flags & SEXP_FLAG_DO_EXPAND) {
- return _sexp_expand_query (notmuch, prefix, parent, sx->list->next, output);
+ return _sexp_expand_query (notmuch, prefix, parent, env, sx->list->next, output);
}
- return _sexp_combine_query (notmuch, parent, prefix->xapian_op, prefix->initial,
+ return _sexp_combine_query (notmuch, parent, env, prefix->xapian_op, prefix->initial,
sx->list->next, output);
}
}
@@ -413,6 +582,6 @@ _notmuch_sexp_string_to_xapian_query (notmuch_database_t *notmuch, const char *q
return NOTMUCH_STATUS_BAD_QUERY_SYNTAX;
}
- return _sexp_to_xapian_query (notmuch, NULL, sx, output);
+ return _sexp_to_xapian_query (notmuch, NULL, NULL, sx, output);
}
#endif
next reply other threads:[~2021-08-12 17:08 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-08-12 17:06 David Bremner [this message]
2021-08-12 17:06 ` [PATCH 01/31] configure: optional library sfsexp David Bremner
2021-08-12 17:06 ` [PATCH 02/31] lib: split notmuch_query_create David Bremner
2021-08-12 17:07 ` [PATCH 03/31] lib: define notmuch_query_create_with_syntax David Bremner
2021-08-12 17:07 ` [PATCH 04/31] CLI/search+address: support sexpr queries David Bremner
2021-08-12 17:07 ` [PATCH 05/31] lib: add new status code for query syntax errors David Bremner
2021-08-12 17:07 ` [PATCH 06/31] lib/parse-sexp: parse single terms and the empty list David Bremner
2021-08-12 17:07 ` [PATCH 07/31] lib: leave stemmer object accessible David Bremner
2021-08-12 17:07 ` [PATCH 08/31] lib/parse-sexp: stem unquoted atoms David Bremner
2021-08-12 17:07 ` [PATCH 09/31] lib/parse-sexp: support and, not, and or David Bremner
2021-08-12 17:07 ` [PATCH 10/31] lib/parse-sexp: support subject field David Bremner
2021-08-12 17:07 ` [PATCH 11/31] util/unicode: allow calling from C++ David Bremner
2021-08-12 17:07 ` [PATCH 12/31] lib/parse-sexp: support phrase queries David Bremner
2021-08-12 17:07 ` [PATCH 13/31] lib/parse-sexp: add term prefix backed fields David Bremner
2021-08-12 17:07 ` [PATCH 14/31] lib/parse-sexp: 'starts-with' wildcard searches David Bremner
2021-08-12 17:07 ` [PATCH 15/31] lib/parse-sexp: add '*' as syntactic sugar for '(starts-with "")' David Bremner
2021-08-12 17:07 ` [PATCH 16/31] lib/parse-sexp: handle unprefixed terms David Bremner
2021-08-12 17:07 ` [PATCH 17/31] lib/query: generalize exclude handling to s-expression queries David Bremner
2021-08-12 17:07 ` [PATCH 18/31] lib: factor out query construction from regexp David Bremner
2021-08-12 17:07 ` [PATCH 19/31] lib/parse-sexp: support regular expressions David Bremner
2021-08-12 17:07 ` [PATCH 20/31] lib: generate actual Xapian query for "*" and "" David Bremner
2021-08-12 17:07 ` [PATCH 21/31] lib/query: factor out _notmuch_query_string_to_xapian_query David Bremner
2021-08-12 17:07 ` [PATCH 22/31] lib/thread-fp: factor out query expansion, rewrite in Xapian David Bremner
2021-08-12 17:07 ` [PATCH 23/31] lib/parse-sexp: expand queries David Bremner
2021-08-12 17:07 ` [PATCH 24/31] lib/parse-sexp: support infix subqueries David Bremner
2021-08-12 17:07 ` [PATCH 25/31] lib/parse-sexp: parse user headers David Bremner
2021-08-12 17:07 ` [PATCH 26/31] lib: factor out expansion of saved queries David Bremner
2021-08-12 17:07 ` [PATCH 27/31] lib/parse-sexp: handle " David Bremner
2021-08-12 17:07 ` [PATCH 28/31] CLI/config support saving s-expression queries David Bremner
2021-08-12 17:07 ` [PATCH 29/31] lib/parse-sexp: support saved " David Bremner
2021-08-12 17:07 ` [PATCH 30/31] lib/parse-sexp: thread environment argument through parser David Bremner
2021-08-12 17:07 ` [PATCH 31/31] lib/parse-sexp: apply macros David Bremner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: https://notmuchmail.org/
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20210812170728.1348333-1-david@tethera.net \
--to=david@tethera.net \
--cc=notmuch@notmuchmail.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://yhetil.org/notmuch.git/
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).