Filter support added #16

ccp-chargeback · 2026-01-31T15:20:09Z

Allow filtering of included resource files based on rules defined in one or more supplied filter.ini files.
Same changes as in PR/14 (done to try to trigger TC build on new project after latest changes from template project were added)

The change contains:

Reading and parsing of .ini filter files.
Allow overrides of include/exclude rules
Support wildcard filters:
- "*" = current folder
- "..." = any subfolder
Define priorities for include/exclude filter:
- Based on folder/file hierarcy
- Include/Exclude sameName file based on its hierarcy in each
Add CLI integration
Add loads of test coverage

Make use of existing filePath test functions

Next step, add missing tests.

Also fixed clang-tidy suggestions

- Do some function name changes to reduce stuttering.

- mostly <stdexcept>

- It worked (wrong) on Windows - Failed on macOS

- Caused a warning as error on macOS - Wasn't being used so safe to remove.

- Fails to build on macOS (incorrect casing, warning as errors)

- Change includes both tools and tests

…ePaths - Temporarily adding debug information to test in order to see why it is failing for macOS in TeamCity

Changed tests: - ResourceFilter_Load_validSimpleExample1_ini_usingRelativePaths - ResourceFilter_Load_validSimpleExample1_ini_usingAbsolutePaths - ResourceFilter_Load_validComplexExample1_ini_usingRelativePaths - ResourceFilter_Load2iniFiles_validComplexExample1_and_validSimpleExample1 New tests: - ResourceFilter_Load_validComplexExample1_ini_usingAbsolutePaths

- To be removed later

- Done to generate the missing "gold" compare files on macOS

CCPCookies

Looks nice, left some comments.

Not managed to go through it completely yet but just wanted to submit this for now as I ran out of day.

CCPCookies · 2026-02-03T15:53:05Z

tests/src/CliTestFixture.cpp

-}
+}
+
+void CliTestFixture::CleanupTestOutputFiles( const std::vector<std::filesystem::path>& filesToRemove )


Should probably be called just RemoveFiles as it could be called with any file, not just test output files.

CCPCookies · 2026-02-03T15:54:44Z

tests/src/ResourceFilterTest.cpp

+#include <FilterResourceFile.h>
+#include <ResourceFilter.h>
+
+TEST_F( ResourceFilterTest, Example1IniParsing )


Test name is not descriptive; from test name it should be obvious what the test is testing.

CCPCookies · 2026-02-03T15:55:47Z

tests/src/ResourceFilterTest.cpp

+
+// -----------------------------------------
+
+TEST_F( ResourceFilterTest, FilterResourceFilter_OnlyIncludeFilter )


From title I'm not sure what this test is supposed to do.

CCPCookies · 2026-02-03T15:57:21Z

tests/src/ResourceFilterTest.cpp

+	EXPECT_TRUE( excludes.empty() );
+}
+
+TEST_F( ResourceFilterTest, FilterResourceFilter_OnlyExcludeFilter_Toplevel )


Again, filter name is confusing.

CCPCookies · 2026-02-03T15:57:57Z

tests/src/ResourceFilterTest.cpp

+	EXPECT_EQ( includes[0], "*" ); // Wild-card added when no include filter specified for TOP-LEVEL filter
+}
+
+TEST_F( ResourceFilterTest, FilterResourceFilter_OnlyExcludeFilter_Inline )


Same as above

CCPCookies · 2026-02-03T16:33:35Z

tools/include/FilterPrefixmap.h

+
+#include <string>
+#include <map>
+#include <vector>


CCPCookies · 2026-02-03T16:34:28Z

tools/include/FilterNamedSection.h

+#define FILTERNAMEDSECTION_H
+
+#include <string>
+#include <vector>


CCPCookies · 2026-02-03T16:36:55Z

tools/include/FilterResourceFile.h

+	// Resolved PathMap for all named sections defined in a resource .ini file
+	std::map<std::string, FilterResourceFilter> m_iniFileResolvedPathMap;
+
+	void ParseIniFile();


Not a fan of having a function below members.

Personally, I like to have two private: areas to split private functions and private members but I think our coding standards says not to do this.

CCPCookies · 2026-02-03T16:41:36Z

tools/include/ResourceFilter.h

+
+	bool HasFilters() const
+	{
+		return !m_filterFiles.empty();


Why inline this?

CCPCookies · 2026-02-03T16:44:51Z

tools/src/FilterResourceFile.cpp

+FilterResourceFile::FilterResourceFile( const std::filesystem::path& iniFilePath ) :
+	m_iniFilePath( iniFilePath )
+{
+	m_iniFileResolvedPathMap.clear();


Wont this always be clear at this point anyway?

- Done to try to fix C++20 arm64 osx triplet cmake error for inih

- Use string() instead of generic_string()

- Also change Filter cli tests to set an override

CCPCookies

Likely lots of stuff due to this being a draft.
Made a first pass of it now.

CCPCookies · 2026-02-04T13:51:53Z

tools/src/FilterResourceFilter.cpp

+{
+
+FilterResourceFilter::FilterResourceFilter( const std::string& rawFilter, bool isToplevelFilter /* = false */ ) :
+	m_rawFilter( rawFilter ),


std::string being passed by const ref here, fine with that.
Previously in other areas std::move was used which I am also fine with.

Not huge deal but a bit strange to have two styles in the changes.

CCPCookies · 2026-02-04T13:53:48Z

tools/src/FilterDefaultSection.cpp

@@ -0,0 +1,20 @@
+// Copyright © 2025 CCP ehf.
+
+#include <stdexcept>


Is this used?

CCPCookies · 2026-02-04T13:55:33Z

tools/src/FilterNamedSection.cpp

+	m_respaths( respaths, parentPrefixMap, m_filter ),
+	m_resfile( resfile.empty() ? std::nullopt : std::make_optional<FilterResourcePathFile>( resfile, parentPrefixMap, m_filter ) )
+{
+	m_resolvedCombinedPathMap.clear();


Isn't this already clear?

CCPCookies · 2026-02-04T14:40:59Z

tools/src/FilterNamedSection.cpp

+	}
+
+	// Return empty map if no resfile present
+	static const std::map<std::string, FilterResourceFilter> emptyResfileMap;


This feels odd.Assuming this was a way for this function to sort of continue on fail.

Is it important to hide this failure from the caller?

Have it return a bool and take in a ref to the map which gets populated on success. Though probably requires a copy.

Perhaps it doesn't make full sense to have a function to GetResourceResfileMap from a named section? Perhaps caller should access resfile externally to this and so this function no longer is a wrapper around this data? (Hard to fully grasp from GitHub Diff sorry)

CCPCookies · 2026-02-04T14:44:00Z

tools/src/FilterPrefixmap.cpp

+
+FilterPrefixMap::FilterPrefixMap( const std::string& rawPrefixMap )
+{
+	m_prefixMapEntries.clear();


Again, isn't this clear already?

CCPCookies · 2026-02-04T16:43:18Z

tools/src/ResourceFilter.cpp

+{
+	// Replace ... with a unique token
+	std::string pat = pattern;
+	std::string token = "\x01";


Needs to follow our constants coding guidelines

CCPCookies · 2026-02-04T16:44:22Z

tools/CMakeLists.txt

        include/RollingChecksum.h
        include/ScopedFile.h
        include/StatusCallback.h
+        include/FilterResourceFile.h


The list was alphabetical :D but this can go down as a nitpick :D

CCPCookies · 2026-02-04T16:44:42Z

tools/CMakeLists.txt

        src/ScopedFile.cpp
        src/Patching.cpp
        src/RollingChecksum.cpp
+        src/FilterResourceFile.cpp


Same as above :D

CCPCookies · 2026-02-04T16:45:33Z

.gitignore

 vcpkg_registry_cache/
+
+### Ignored Test Run Files ###
+/tests/testData/IgnoredTestOutputFiles/


? There are now output files generated in the repo during test?

CCPCookies · 2026-02-04T16:49:30Z

CMakeLists.txt

 include(cmake/CcpBuildConfigurations.cmake)

 find_package(yaml-cpp CONFIG REQUIRED)
+find_package(unofficial-inih CONFIG REQUIRED)


What's with the unofficial name? Is it that we use features that are not strictly supported through just inih or something?

ccp-chargeback added 30 commits December 1, 2025 14:42

Add example1.ini file (for tests)

fbb3830

Add vcpkg inih to project

31479c2

Add first ResourceFilterTest

9ce31fa

Update ExampleIniParsing test

4a0a0d8

Make use of existing filePath test functions

Rename the ResourceFilterTest class

dbbfb83

Add scaffolding for the Filter classes

fd26db0

Include exclude filter implementation without regex

40ec602

Simplify FilterResourceFilter::ParseFilters()

ede9636

Add error tests for FilterResourceFilter::ParseFilters()

b79fe7e

Add tests for Condensed but valid filter strings

a2bd883

Add Doxygen markdown comments to FilterResourceFilter class.

1a406b3

Change scaffolding of FilterResourceFilter

c65e3c4

Move Filter logic into tools

5d151d0

Make filter tool classes part of ResourceTools namespace

c8b8bcd

Fix Filter tests to now use new ResourceTools namespace

5fed2b9

Add test for multiline respaths with empty lines

369e3c8

Refactor of ResourceLing and PrefixMap

f9d4965

Add FilterPrefixMap and Entry classes

8a68f53

Next step, add missing tests.

Add tests for FilterPrefixMap and Entry

e90027b

Include and function signature cleanup

cd967c2

Add tests for class FilterDefaultSection

a84ebd9

Fixup NamedSection and related classes.

12e4d74

Add test for Duplicate Filter in Overrides

3ce25ae

Change m_parentPrefixMap to a const reference

f96197d

Also fixed clang-tidy suggestions

Change m_parentSectionFilter to a const reference

698874b

Remove un-necessary default constructors

74de122

Rename and gorup Resource Filter tests by class tested

022b0cf

Add FilterNamedSection tests

7b31e28

Add FilterResourceFile class

073a220

- Do some function name changes to reduce stuttering.

Add test for FilterResourceFile class (example1.ini)

c885982

ccp-chargeback added 4 commits January 26, 2026 17:17

Fix formatting of FilterResources CLI tests

d80b9e6

Remove unnecessary comments in class CurrentWorkingDirectoryChanger

56dbfb0

Merge branch 'main' into add-resource-filtering

215803d

Update include statements

ae101d9

- mostly <stdexcept>

ccp-chargeback requested a review from CCPCookies January 31, 2026 15:20

ccp-chargeback added 16 commits January 31, 2026 15:42

Fix wrong case of #include <FilterPrefixmap.h>

b19dba2

- It worked (wrong) on Windows - Failed on macOS

Remove unused m_parentPrefixMap member in NamedSection

8a4b712

- Caused a warning as error on macOS - Wasn't being used so safe to remove.

Change Kotlin Github access token

2cddd56

Move include statement to trigger TC build

8459427

Fix casing of #include <FilterPrefixmap.h> (missed one from earlier)

e260153

- Fails to build on macOS (incorrect casing, warning as errors)

Use generic_string() for all paths involving Filter operations

a943d6a

- Change includes both tools and tests

Debug info - ResourceFilter_Load_validSimpleExample1_ini_usingRelativ…

d8ee506

…ePaths - Temporarily adding debug information to test in order to see why it is failing for macOS in TeamCity

Normalize path strings before WildcardMatching

25d2083

Change WildcardMatch regex exception handling

e237775

Fix formatting of FilterResourcePathFile.cpp

b983c2a

Add debug info in ResourceFilter.cpp

8841d93

- To be removed later

Fix handling on recursive folder wildcard "..."

d7c7bda

Temp disable expected checks on Filter CLI tests

b0d0b53

- Done to generate the missing "gold" compare files on macOS

Add debug statements for CLI Filter tests

85db31b

Normalize paths in CLI Filter test arguments

02c5c0f

CCPCookies reviewed Feb 3, 2026

View reviewed changes

ccp-chargeback added 7 commits February 3, 2026 21:24

Update microsoft vcpkg baseline

4b0fa83

- Done to try to fix C++20 arm64 osx triplet cmake error for inih

Normalize paths in CLI Filter test - take 2

5f64f32

Normalize outputFilePath in CLI Filter tests

9873d32

Use system specific filePaths for CLI Filter tests

bf0f295

- Use string() instead of generic_string()

Change RunCli() to populate stderr as well

9d302ec

Populate std::cerr on Filter errors

a658470

Add override of working directory in CliTestFixture::RunCli()

0cc5064

- Also change Filter cli tests to set an override

CCPCookies reviewed Feb 4, 2026

View reviewed changes


		// -----------------------------------------

		TEST_F( ResourceFilterTest, FilterResourceFilter_OnlyIncludeFilter )

		@@ -0,0 +1,20 @@
		// Copyright © 2025 CCP ehf.

		#include <stdexcept>

Filter support added #16

Are you sure you want to change the base?

Filter support added #16

Uh oh!

Conversation

ccp-chargeback commented Jan 31, 2026

Uh oh!

CCPCookies left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

CCPCookies left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants