From 9b77ae4543fc230296b559f486582d0697cef1e2 Mon Sep 17 00:00:00 2001 From: "Danilo M." Date: Thu, 13 Aug 2026 11:09:22 +0200 Subject: feat(rulequery): parse a flat and/or chain --- src/rulequery.cpp | 199 ++++++++++++++++++++++++++++++++++++++++++++++- tests/test_rulequery.cpp | 93 ++++++++++++++++++++++ 2 files changed, 290 insertions(+), 2 deletions(-) diff --git a/src/rulequery.cpp b/src/rulequery.cpp index ac8bf18..d4d8966 100644 --- a/src/rulequery.cpp +++ b/src/rulequery.cpp @@ -18,7 +18,9 @@ #include "rulequery.h" +#include #include +#include namespace { @@ -89,6 +91,130 @@ QString compileTerm(const RuleTerm &term) return isNegated(term.op) ? QStringLiteral("not ") + body : body; } +/// Splits on whitespace, keeping a double-quoted run as one token. Returns +/// false when a quote is left open, which is a query this builder will not +/// represent. +bool tokenise(const QString &query, QStringList *out) +{ + QString current; + bool inQuotes = false; + bool has = false; + + for (int i = 0; i < query.size(); ++i) { + const QChar c = query.at(i); + if (c == QLatin1Char('"')) { + inQuotes = !inQuotes; + current += c; + has = true; + } else if (!inQuotes && c.isSpace()) { + if (has) { + out->append(current); + current.clear(); + has = false; + } + } else { + current += c; + has = true; + } + } + + if (inQuotes) + return false; + if (has) + out->append(current); + return true; +} + +bool fieldForPrefix(const QString &prefix, RuleTerm::Field *out) +{ + static const QVector> table = { + {QStringLiteral("from"), RuleTerm::From}, + {QStringLiteral("to"), RuleTerm::To}, + {QStringLiteral("cc"), RuleTerm::Cc}, + {QStringLiteral("subject"), RuleTerm::Subject}, + {QStringLiteral("tag"), RuleTerm::Tag}, + {QStringLiteral("path"), RuleTerm::Folder}, + {QStringLiteral("attachment"), RuleTerm::Attachment}, + {QStringLiteral("date"), RuleTerm::Date}, + }; + + for (const auto &entry : table) { + if (entry.first == prefix) { + *out = entry.second; + return true; + } + } + return false; +} + +/// Parses ONE token into a term. Returns false for anything this builder does +/// not represent, which is not the same as invalid: notmuch accepts far more +/// than this. +bool parseTerm(const QString &token, RuleTerm *out) +{ + const int colon = token.indexOf(QLatin1Char(':')); + if (colon <= 0) + return false; + + RuleTerm::Field field; + if (!fieldForPrefix(token.left(colon), &field)) + return false; + + QString value = token.mid(colon + 1); + if (value.isEmpty()) + return false; + + bool quoted = false; + if (value.size() >= 2 && value.startsWith(QLatin1Char('"')) + && value.endsWith(QLatin1Char('"'))) { + value = value.mid(1, value.size() - 2); + quoted = true; + } + // A quote anywhere else means a shape this builder does not emit. + if (value.contains(QLatin1Char('"'))) + return false; + + out->field = field; + + if (field == RuleTerm::Date) { + if (value.startsWith(QStringLiteral(".."))) { + out->op = RuleTerm::Before; + out->value = value.mid(2); + } else if (value.endsWith(QStringLiteral(".."))) { + out->op = RuleTerm::After; + out->value = value.chopped(2); + } else { + return false; // A two-sided range is not a row. + } + return !out->value.isEmpty(); + } + + if (field == RuleTerm::Folder) { + // Only the recursive form is representable; a bare path means + // something different to notmuch and must not be silently rewritten. + if (!value.endsWith(QStringLiteral("/**"))) + return false; + value = value.chopped(3); + out->op = RuleTerm::Is; + out->value = value; + return !value.isEmpty(); + } + + // Tag and Attachment compile unquoted (see needsQuotes), so their + // operator must not be inferred from the quoting: reading a quoted tag + // back as a quoting operator would compile it unquoted and change the + // stored string. + if (field == RuleTerm::Attachment) + out->op = RuleTerm::Has; + else if (field == RuleTerm::Tag) + out->op = RuleTerm::Is; + else + out->op = quoted ? RuleTerm::Is : RuleTerm::Contains; + + out->value = value; + return true; +} + } // namespace bool operator==(const RuleTerm &a, const RuleTerm &b) @@ -133,6 +259,75 @@ QString RuleQuery::compile() const RuleQuery RuleQuery::parse(const QString &query) { - Q_UNUSED(query); - return RuleQuery(); + RuleQuery out; + + const QString trimmed = query.trimmed(); + if (trimmed.isEmpty()) { + // An empty query is a rule with no rows yet, not a failure. + out.parsed = true; + return out; + } + + QStringList tokens; + if (!tokenise(trimmed, &tokens)) + return RuleQuery(); + + // Walk the chain: term, operator, term, ... Anything else rejects whole. + bool sawOr = false; + bool sawAnd = false; + int i = 0; + + while (i < tokens.size()) { + bool negated = false; + if (tokens.at(i).compare(QStringLiteral("not"), + Qt::CaseInsensitive) == 0) { + negated = true; + ++i; + if (i >= tokens.size()) + return RuleQuery(); + } + + RuleTerm term; + if (!parseTerm(tokens.at(i), &term)) + return RuleQuery(); + + if (negated) { + // `not date:` has no row form: "not before" is "after", which the + // unnegated operators already express. + if (term.field == RuleTerm::Date) + return RuleQuery(); + // The block IS the negation, so the row is stored un-negated and + // compile() re-applies the `and not`. + out.exclusions.append(term); + } else { + out.terms.append(term); + } + ++i; + + if (i >= tokens.size()) + break; + + const QString glue = tokens.at(i).toLower(); + if (glue == QStringLiteral("and")) { + sawAnd = true; + } else if (glue == QStringLiteral("or")) { + sawOr = true; + } else { + return RuleQuery(); // Not a joining word: unrepresentable. + } + ++i; + if (i >= tokens.size()) + return RuleQuery(); // Trailing operator. + } + + // Mixed and/or without parentheses is ambiguous to a reader and binds in + // a way the rows cannot show. Reject rather than guess. + if (sawAnd && sawOr) + return RuleQuery(); + if (out.terms.isEmpty()) + return RuleQuery(); + + out.join = sawOr ? Any : All; + out.parsed = true; + return out; } diff --git a/tests/test_rulequery.cpp b/tests/test_rulequery.cpp index 2eb34ed..b8d6f22 100644 --- a/tests/test_rulequery.cpp +++ b/tests/test_rulequery.cpp @@ -41,6 +41,12 @@ private slots: void exclusionsAppendAsAndNot(); void anyIsParenthesisedOnlyWhenExclusionsFollow(); void anEmptyQueryCompilesToAnEmptyString(); + void aFlatAndChainParses(); + void aFlatOrChainParses(); + void anEmptyQueryParsesToNoRows(); + void quotedValuesLoseTheirQuotes(); + void aNegatedTermParsesAsANegatedOperator(); + void whatParsesCompilesBackUnchanged(); }; void TestRuleQuery::aSingleContainsTermCompiles() @@ -207,5 +213,92 @@ void TestRuleQuery::anEmptyQueryCompilesToAnEmptyString() QCOMPARE(q.compile(), QString()); } +void TestRuleQuery::aFlatAndChainParses() +{ + const RuleQuery q = RuleQuery::parse( + QStringLiteral("from:vendor.example.org and subject:receipt")); + + QVERIFY(q.parsed); + QCOMPARE(q.join, RuleQuery::All); + QCOMPARE(q.terms.size(), 2); + QCOMPARE(q.terms.at(0).field, RuleTerm::From); + QCOMPARE(q.terms.at(0).op, RuleTerm::Contains); + QCOMPARE(q.terms.at(0).value, QStringLiteral("vendor.example.org")); + QCOMPARE(q.terms.at(1).field, RuleTerm::Subject); + QVERIFY(q.exclusions.isEmpty()); +} + +void TestRuleQuery::aFlatOrChainParses() +{ + const RuleQuery q = RuleQuery::parse( + QStringLiteral("from:one.example.org or from:two.example.org")); + + QVERIFY(q.parsed); + QCOMPARE(q.join, RuleQuery::Any); + QCOMPARE(q.terms.size(), 2); +} + +void TestRuleQuery::anEmptyQueryParsesToNoRows() +{ + // One shipped rule has an empty query. It must open in the builder ready + // to receive a row, not fall back to text mode. + const RuleQuery q = RuleQuery::parse(QString()); + + QVERIFY(q.parsed); + QVERIFY(q.terms.isEmpty()); +} + +void TestRuleQuery::quotedValuesLoseTheirQuotes() +{ + const RuleQuery q = RuleQuery::parse( + QStringLiteral("subject:\"your receipt\"")); + + QVERIFY(q.parsed); + QCOMPARE(q.terms.size(), 1); + QCOMPARE(q.terms.at(0).value, QStringLiteral("your receipt")); + QCOMPARE(q.terms.at(0).op, RuleTerm::Is); +} + +void TestRuleQuery::aNegatedTermParsesAsANegatedOperator() +{ + const RuleQuery q = RuleQuery::parse( + QStringLiteral("from:vendor.example.org and not tag:inbox")); + + QVERIFY(q.parsed); + // A trailing negation on an `and` chain becomes an exclusion: that is how + // the user describes these rules, and the design records the preference. + QCOMPARE(q.terms.size(), 1); + QCOMPARE(q.exclusions.size(), 1); + QCOMPARE(q.exclusions.at(0).field, RuleTerm::Tag); + QCOMPARE(q.exclusions.at(0).op, RuleTerm::Is); +} + +void TestRuleQuery::whatParsesCompilesBackUnchanged() +{ + // The dialog decides whether to rewrite the stored string by comparing + // against what it parsed, so a compile that differs by so much as a quote + // would churn a file a second tool reads. + const QStringList queries = { + QStringLiteral("from:vendor.example.org"), + QStringLiteral("from:vendor.example.org and subject:receipt"), + QStringLiteral("from:one.example.org or from:two.example.org"), + QStringLiteral("subject:\"your receipt\""), + QStringLiteral("path:\"account-one/**\""), + QStringLiteral("tag:inbox"), + QStringLiteral("attachment:pdf"), + QStringLiteral("date:..2026-01-01"), + QStringLiteral("date:2026-01-01.."), + QStringLiteral("from:vendor.example.org and not tag:inbox"), + QStringLiteral("from:vendor.example.org and not subject:receipt " + "and not subject:refund"), + }; + + for (const QString &query : queries) { + const RuleQuery parsed = RuleQuery::parse(query); + QVERIFY2(parsed.parsed, qPrintable(query)); + QCOMPARE(parsed.compile(), query); + } +} + QTEST_MAIN(TestRuleQuery) #include "test_rulequery.moc" -- cgit v1.2.3