implement 'ReadAttribute' inside the codegen-data-model IM/DM split (…

…#33877) * Initial copy with a clean history * make linter happy * Restyle * Fix typo * Add nolint: assert will return before we use the underlying value * 2 more fixes regarding unchecked access * Switch some asserts to expects, for better test logic * Start implementing a read ... wip while I switch to a few other tests * More implementation and we seem to need privilege check as well * More mock ember functions implemented, to make it at least link for now * A bit of work on AAI support. Still need to split out global attribute readers * Split out global AAI...for now * Add global attribute access interface as part of ember data model definitions ... they seem needed * Drop RequiredPrivilegeStubs ... those are odd and if you use mock ember you may as well mock these out as well * Restyle * Added sufficient overrides for things to compile ... this is a MESS * Some more changes to compile * Things link now ... but boy do we have many workarounds... * Remove double return * Restyle * Full ember handling * Comment updates and slight restyle * remove commented out code * Remove read-state from arguments - attributevalueencoder maintains its own state, so we should not duplicate things at this time * more comments and some logic cleanup on alternatives ... code should be less broken now * clean one more comment that seemed odd * Fix off by one bug for long strin processing * More generics * Split out io storage and type conversion as shared functions for ember-compatibility * Restyle * Added more comments * Fix lint errors ... these files are not in libraries since they depend on metadata * Restyle * Fix the access denied error logic ... the translation to UnsupportedAccess is to be done by the caller at a later time, since caller may skip * Shorter code is better * Some comments, cleaner code * Fix auto-include * add some TODO because access control is needed * Model renames * Add renamed files * Add some attribute iteration hint * Make use of the attribute cache * Restyle * Add a cluster iteration hint * Add a few more hints. Ember code still contains loops though, so this may not be ideal still * Add some TODO items for using faster iterations for data. Ember index vs value duality still needs some work * Add a cluster type cache as well. This relies on ember being reasonably static * Add global attribute handling * Fix typing u16 vs unsigned * Unit test preparation: make ACL pass and have an ACL test as well * Some progress in testability. No mock reads, however at least some progress * Make some unit tests pass with ember overrides * Restyle * Attempt to decode data. It does NOT work yet * Validation actually passes * Start splitting into unrelated files ...this is painful * Restyle * Better test namespacing - be consistent * Re-organize a bit ... boilerplate still seems A LOT * Minor comments to start organizing the code better - I NEED more readable things * Restyle and re-organize for readability * More code changes to support testing strings ... however this FAILS right now * Restyle * Fix a typo ... still broken though * one more typo fix ... test passes * Long octet string test as well * Restyle * Fix comment and change the size of the string * Add several ember-specific tests * unit tests and working with nullable values * Fix up the tests * Restyle * AAI unit test for read * Slight comment updates * Restyle * Use StringBuilderAdapters to format chip_error nicely * Format * More unit tests for lists ... test overflow as well * Add test support for partially encoded lists * add missing file * Fix name shadowing * Fix auto-added include names * Test global attribute read via AAI * More unit test coverage * Test nullable string reads * One more test for error pahs * More failure cases on failure path * Restyle * Remove back the initialization and make the comment more obvious * Undo odd include that got auto-added * Place files back past renames * A bit more renaming * Fix includes * More updates to the latest code * Fix unit test merging * Resolve data model ambiguity * Restyle * Do not try to compile ember tests on zephyr as it is not compatible with single large executable * Update placement and comment for including codegen test into one large binary * Restyle * Add static cast for ember calls for span sizes ... we know we are less than 65K * Fix clang-tidy error report * Restyle * Do not try to translate UnsupportedRead * Typo fixes * Code review feedback: correct the return code for global attribute missing the cluster * Some spelling updates and format * ChipDie if internal logic for read fails * Fix typo * Add unit test for invalid global attribute read * Code review feedback: comment about internal flags and implement path expansion logic for skipping data encoding * Restyle and unit test for expansion * Another unit test for AAI this time for unsupported read * Restyle * Update src/app/codegen-data-model/tests/TestAttributeReportIBsEncoding.cpp Co-authored-by: Boris Zbarsky <bzbarsky@apple.com> * Rename files to not have a Test prefix when not containing tests * Add comment as per review comment * Use macros for error returns * Fix typo * Add a macro for dynamic global IM status codes in chip_error * Reference issue in comment * Fix comment a bit * Undo module updates ... this time for real --------- Co-authored-by: Andrei Litvin <andreilitvin@google.com> Co-authored-by: Boris Zbarsky <bzbarsky@apple.com>
project-chip · Jun 28, 2024 · 28430f1 · 28430f1
1 parent f62d3bf
commit 28430f1
Show file tree

Hide file tree

Showing 17 changed files with 2,038 additions and 37 deletions.
diff --git a/examples/chef/chef.py b/examples/chef/chef.py
@@ -878,7 +878,7 @@ def main() -> int:
                         """))
             if options.do_clean:
                 shell.run_cmd("rm -rf out")
-            shell.run_cmd("gn gen out")
+            shell.run_cmd("gn gen --export-compile-commands out")
             shell.run_cmd("ninja -C out")
 
     #

diff --git a/src/BUILD.gn b/src/BUILD.gn
@@ -50,7 +50,6 @@ if (chip_build_tests) {
   chip_test_group("tests") {
     deps = []
     tests = [
-      "${chip_root}/src/app/codegen-data-model/tests",
       "${chip_root}/src/app/data-model-interface/tests",
       "${chip_root}/src/access/tests",
       "${chip_root}/src/crypto/tests",
@@ -82,7 +81,15 @@ if (chip_build_tests) {
 
     if (current_os != "zephyr" && current_os != "mbed" &&
         chip_device_platform != "efr32") {
+      # Avoid these items from "one single binary" test executions. Once tests
+      # are split, we can re-visit this (and likely many others)
+      #
+      # In particular:
+      #   "app/codegen-data-model/tests" contains symbols for ember mocks which
+      #                                  are used by other tests
+
       tests += [
+        "${chip_root}/src/app/codegen-data-model/tests",
         "${chip_root}/src/setup_payload/tests",
         "${chip_root}/src/transport/raw/tests",
       ]

diff --git a/src/app/codegen-data-model/BUILD.gn b/src/app/codegen-data-model/BUILD.gn
@@ -20,6 +20,7 @@ import("//build_overrides/chip.gni")
 #
 # Use `model.gni` to get access to:
 #   CodegenDataModel.cpp
+#   CodegenDataModel_Read.cpp
 #   CodegenDataModel.h
 #
 # The above list of files exists to satisfy the "dependency linter"

diff --git a/src/app/codegen-data-model/CodegenDataModel.cpp b/src/app/codegen-data-model/CodegenDataModel.cpp
@@ -231,13 +231,6 @@ bool CodegenDataModel::EmberCommandListIterator::Exists(const CommandId * list,
     return (*mCurrentHint == toCheck);
 }
 
-CHIP_ERROR CodegenDataModel::ReadAttribute(const InteractionModel::ReadAttributeRequest & request,
-                                           InteractionModel::ReadState & state, AttributeValueEncoder & encoder)
-{
-    // TODO: this needs an implementation
-    return CHIP_ERROR_NOT_IMPLEMENTED;
-}
-
 CHIP_ERROR CodegenDataModel::WriteAttribute(const InteractionModel::WriteAttributeRequest & request,
                                             AttributeValueDecoder & decoder)
 {

diff --git a/src/app/codegen-data-model/CodegenDataModel.h b/src/app/codegen-data-model/CodegenDataModel.h
@@ -68,8 +68,7 @@ class CodegenDataModel : public chip::app::InteractionModel::DataModel
     /// Generic model implementations
     CHIP_ERROR Shutdown() override { return CHIP_NO_ERROR; }
 
-    CHIP_ERROR ReadAttribute(const InteractionModel::ReadAttributeRequest & request, InteractionModel::ReadState & state,
-                             AttributeValueEncoder & encoder) override;
+    CHIP_ERROR ReadAttribute(const InteractionModel::ReadAttributeRequest & request, AttributeValueEncoder & encoder) override;
     CHIP_ERROR WriteAttribute(const InteractionModel::WriteAttributeRequest & request, AttributeValueDecoder & decoder) override;
     CHIP_ERROR Invoke(const InteractionModel::InvokeRequest & request, chip::TLV::TLVReader & input_arguments,
                       InteractionModel::InvokeReply & reply) override;