From a2785c3919c56a370a7f860f4eeb93846f4a1e63 Mon Sep 17 00:00:00 2001 From: David Bremner Date: Tue, 24 Aug 2021 08:17:18 -0700 Subject: [PATCH] lib/parse-sexp: stem unquoted atoms This is somewhat less DWIM than the Xapian query parser, but it has the advantage of simplicity. --- doc/man7/notmuch-sexp-queries.rst | 10 ++++++++-- lib/parse-sexp.cc | 10 +++++++--- test/T081-sexpr-search.sh | 5 ++++- 3 files changed, 19 insertions(+), 6 deletions(-) diff --git a/doc/man7/notmuch-sexp-queries.rst b/doc/man7/notmuch-sexp-queries.rst index 1118f854..d177934d 100644 --- a/doc/man7/notmuch-sexp-queries.rst +++ b/doc/man7/notmuch-sexp-queries.rst @@ -41,8 +41,10 @@ subqueries. The empty list matches all messages *term* - Match all messages containing *term*, possibly after stemming - or phase splitting. + Match all messages containing *term*, possibly after + stemming or phase splitting. For discussion of stemming in + notmuch see :any:`notmuch-search-terms(7)`. Stemming only applies + to unquoted terms (basic values) in s-expression queries. ``(`` *field* |q1| |q2| ... |qn| ``)`` Restrict the queries |q1| to |qn| to *field*, and combine with *and* @@ -76,6 +78,10 @@ EXAMPLES ``Wizard`` Match all messages containing the word "wizard", ignoring case. +``added`` + Match all messages containing "added", but also those containing "add", "additional", + "Additional", "adds", etc... via stemming. + .. |q1| replace:: :math:`q_1` .. |q2| replace:: :math:`q_2` .. |qn| replace:: :math:`q_n` diff --git a/lib/parse-sexp.cc b/lib/parse-sexp.cc index 66dbdb41..f031d790 100644 --- a/lib/parse-sexp.cc +++ b/lib/parse-sexp.cc @@ -1,5 +1,4 @@ -#include -#include "notmuch-private.h" +#include "database-private.h" #if HAVE_SFSEXP #include "sexp.h" @@ -18,7 +17,12 @@ _sexp_to_xapian_query (notmuch_database_t *notmuch, const sexp_t *sx, { if (sx->ty == SEXP_VALUE) { - output = Xapian::Query (Xapian::Unicode::tolower (sx->val)); + std::string term = Xapian::Unicode::tolower (sx->val); + Xapian::Stem stem = *(notmuch->stemmer); + if (sx->aty == SEXP_BASIC) + term = "Z" + stem (term); + + output = Xapian::Query (term); return NOTMUCH_STATUS_SUCCESS; } diff --git a/test/T081-sexpr-search.sh b/test/T081-sexpr-search.sh index 46cc712c..4e3d755c 100755 --- a/test/T081-sexpr-search.sh +++ b/test/T081-sexpr-search.sh @@ -29,11 +29,14 @@ EOF test_expect_equal_file EXPECTED OUTPUT test_begin_subtest "single term in body, stemmed version" -test_subtest_known_broken notmuch search arriv > EXPECTED notmuch search --query=sexp arriv > OUTPUT test_expect_equal_file EXPECTED OUTPUT +test_begin_subtest "single term in body, unstemmed version" +notmuch search --query=sexp '"arriv"' > OUTPUT +test_expect_equal_file /dev/null OUTPUT + test_begin_subtest "Unbalanced parens" # A code 1 indicates the error was handled (a crash will return e.g. 139). test_expect_code 1 "notmuch search --query=sexp '('" -- 2.43.0