Merge branch 'topic/bbannier/bump-tree-sitter-zeek'

.pre-commit-config.yaml

@@ -10,7 +10,7 @@ repos:
   - id: check-added-large-files
 
 - repo: https://github.com/astral-sh/ruff-pre-commit
-  rev: v0.9.2
+  rev: v0.9.3
   hooks:
     - id: ruff
     - id: ruff-format

CHANGES

@@ -1,3 +1,28 @@
+1.3.0-18 | 2025-01-24 17:02:50 +0100
+
+  * Bump pre-commit hooks (Benjamin Bannier, Corelight)
+
+  * Add test for formatting of IDs (Benjamin Bannier, Corelight)
+
+  * Fix formatting of redefs changing field attributes (Benjamin Bannier, Corelight)
+
+  * Add formatter for `@pragma` (Benjamin Bannier, Corelight)
+
+  * Make sample numbering deterministic (Benjamin Bannier, Corelight)
+
+  * Do not include path in snapshot name (Benjamin Bannier, Corelight)
+
+  * Bump tree-sitter-zeek (Benjamin Bannier, Corelight)
+
+    This bump brings in e.g., support for parsing more attributes and
+    support for parsing some additional forms of IDs, record field removal
+    redefs and some forms of pragmas.
+
+    It also introduces some changes to behavior on errors. Since technically
+    `zeek-format` should not format inputs with parse errors this does seem
+    slightly invariant, but since we currently test for the exact error
+    behavior here I updated test baselines for that.
+
 1.3.0-10 | 2025-01-22 07:32:36 +0100
 
   * build(deps-dev): bump tree-sitter from 0.23.2 to 0.24.0 (dependabot[bot])

VERSION

@@ -1 +1 @@
-1.3.0-10
+1.3.0-18

pyproject.toml

@@ -28,7 +28,7 @@ requires-python = ">= 3.10"
 
 dependencies = [
     "tree-sitter==0.24.0",
-    "tree-sitter-zeek==0.2.1",
+    "tree-sitter-zeek==0.2.2",
 ]
 
 [project.optional-dependencies]

tests/__snapshots__/test_samples/test_samples[sample0][id.zeek].raw

@@ -0,0 +1,3 @@
+global a: bool;
+global mod::a: bool;
+global mod::sub_mod::a: bool;

tests/__snapshots__/test_samples/test_samples[sample1][pragma.zeek].raw

@@ -0,0 +1,7 @@
+event run_sync_hook()
+	{
+	hook Telemetry::sync();
+@pragma push ignore-deprecations
+	schedule sync_interval { run_sync_hook() };
+@pragma pop ignore-deprecations
+	}

tests/__snapshots__/test_samples/test_samples[sample2][redef.zeek].raw

@@ -0,0 +1,10 @@
+redef record fa_file += {
+	pe: Info &optional;
+};
+
+redef enum Log::ID += { LOG };
+
+# Single.
+redef record Conn::Info$ip_proto -= { &log } ;
+# Multiple (actually only `&log` can be redefined).
+redef record Conn::Info$ip_proto -= { &log &optional } ;

tests/samples/id.zeek

@@ -0,0 +1,3 @@
+global a: bool;
+global mod::a: bool;
+global mod::sub_mod::a: bool;

tests/samples/pragma.zeek

@@ -0,0 +1,6 @@
+event run_sync_hook() {
+        hook Telemetry::sync();
+@pragma push ignore-deprecations
+        schedule sync_interval { run_sync_hook() };
+@pragma pop ignore-deprecations
+}

tests/samples/redef.zeek

@@ -0,0 +1,10 @@
+redef record fa_file += {
+	pe: Info &optional;
+};
+
+redef enum Log::ID += { LOG };
+
+# Single.
+redef record Conn::Info$ip_proto-={&log};
+# Multiple (actually only `&log` can be redefined).
+redef record Conn::Info$ip_proto-={&log &optional};

tests/test_formatting.py

@@ -143,15 +143,15 @@ def test_mid_record_error(self):
 """,
             """type foo: record {
 	a: count; ##< A field
-	b count; ##< A broken field
+b count; ##< A broken field
 	c: count; ##< Another field, better not skipped!
 	d: count; ##< Ditto.
 };
 """,
             (
                 "\tb count; ##< A broken field",
                 2,
-                'cannot parse line 2, col 1: "b count;"',
+                'cannot parse line 2, col 1: "b count; ##< A broken field\n\tc: count; ##< Another[...]"',
             ),
         )
 

tests/test_samples.py

@@ -9,6 +9,7 @@
 """
 
 import io
+import os
 from pathlib import Path
 
 import pytest
@@ -39,7 +40,8 @@ def _get_samples():
     # We exclude directories since we store snapshots along with the samples.
     # This assumes that there are no tests in subdirectories of `SAMPLES_DIR`.
     try:
-        return [sample for sample in SAMPLES_DIR.iterdir() if sample.is_file()]
+        # Make sample order deterministic.
+        return sorted([sample for sample in SAMPLES_DIR.iterdir() if sample.is_file()])
     except FileNotFoundError:
         return []
 
@@ -52,7 +54,7 @@ def test_samples(sample: Path, snapshot: SnapshotAssertion):
     assert input_.parse(), f"failed to parse input {sample}"
     assert not input_.has_error(), f"parse result for {sample} has parse errors"
 
-    name = str(sample.relative_to(SAMPLES_DIR.parent.parent))
+    name = str(os.path.basename(sample))
 
     output = _format(input_)
     assert output == snapshot(name=name), (

zeekscript/__init__.py

@@ -1,6 +1,6 @@
 """Wrapper around more low-level tests."""
 
-__version__ = "1.3.0-10"
+__version__ = "1.3.0-18"
 __all__ = [
     "Formatter",
     "Script",

zeekscript/formatter.py

@@ -461,6 +461,11 @@ def format(self):
         self.ostream.use_linebreaks(True)
 
 
+class PragmaFormatter(LineFormatter):
+    def format(self):
+        self._format_token()
+
+
 class ModuleDeclFormatter(Formatter):
     def format(self):
         self._format_child()  # 'module'
@@ -598,19 +603,28 @@ def format(self):
         self._write_sp()
         self._format_child()  # 'record'
         self._write_sp()
-        self._format_child()  # <id>
+
+        # We could either be change fields in a type, or change attributes on a
+        # field. In the first case we would hold an id here, else and expr.
+        is_redef_attr = self._get_child_name() == "expr"
+        self._format_child()  # <id>/<expr>
+
         self._write_sp()
-        self._format_child()  # '+='
+        self._format_child()  # '+='/'-='
         self._write_sp()
-        self._format_child()  # '{'
-        self._write_nl()
-        while self._get_child_name() == "type_spec":  # any number of type_specs
-            self._format_child(indent=True)
-        self._format_child()  # '}'
-        if self._get_child_name() == "attr_list":
-            self._write_sp()
-            self._format_child()  # <attr_list>
-        self._format_child(hints=Hint.NO_LB_BEFORE)  # ';'
+
+        if is_redef_attr:
+            self._format_children(sep=" ")
+        else:
+            self._format_child()  # '{'
+            self._write_nl()
+            while self._get_child_name() == "type_spec":
+                self._format_child(indent=True)
+            self._format_child()  # '}'
+            if self._get_child_name() == "attr_list":
+                self._write_sp()
+                self._format_child()  # <attr_list>
+            self._format_child(hints=Hint.NO_LB_BEFORE)  # ';'
         self._write_nl()