Replaces manual string splitting with filepath.Ext for determining the file extension in renderFileContentAsMarkdown. This improves accuracy, especially for files with multiple dots in their names.
Refactors log rendering to print errors to stderr and continue processing on JSON parse or rendering failures, instead of returning early. Updates tests to check both stdout and stderr outputs, and adds new test cases and testdata for tolerant parsing and error scenarios.
Shell commands in log output are now prefixed with '$ ' for improved readability and consistency with common shell output conventions. Updated related test data and test helper comments to reflect this change.
Moved the genericToolCallNamesToTitles map from inside the renderGenericToolCall function to a package-level variable for improved readability and potential reuse. This change also updates the function signature order for consistency.
Introduces use of a pager for log output in the printLogs function. If the pager fails to start, an error message is printed to stderr. This improves log readability for long outputs.
Uncomments and activates support for Bash session tool calls (write_bash, read_bash, stop_bash, async_bash, read_async_bash, stop_async_bash) in the renderLogEntry function. Also defines the corresponding argument structs, enabling proper handling and display of these tool calls in the log output.
Renamed sample log input and expected output files in testdata to use a consistent 'log-*-input.txt' and 'log-*-want.txt' naming scheme. Updated references in log_test.go to match the new file names for improved clarity and maintainability.
A comment was added to TestFollow with instructions on how to update the .want testdata files when test outputs change. This helps maintainers regenerate expected output files more easily.
Added and expanded Go doc comments for multiple functions in pkg/cmd/agent-task/shared/log.go to clarify their purpose and usage. Also refactored renderToolCallTitle and related logic for improved naming consistency and robustness.
Refactors the function renderToolCall to renderToolCallTitle and updates all its usages for clarity. This improves function naming to better reflect its purpose of rendering tool call titles.
Refactored the function name from relativePath to relativeFilePath for clarity and updated all usages accordingly in log rendering functions. Also improved comments in stripDiffFormat for better context.
Renamed renderMarkdownWithPadding to renderMarkdownWithFormat and related variables for clarity. Updated comments and TODOs for better context on GUI support and code behavior.
Enhances log rendering by stripping diff formatting from viewed file content and improving JSON rendering with optional labels. Expands the list of recognized tool calls with descriptive titles, especially for Playwright and GitHub MCP server tools. Refactors related helper functions for clarity and robustness.
Extracted JSON content rendering into a new helper function renderContentAsJSONMarkdown for reuse. Updated 'report_progress' and 'create' tool call handling to use this helper, improving code clarity and reducing duplication. Also added support for rendering PR descriptions in 'report_progress' tool calls.
Updated the LogRenderer interface and implementations to accept *iostreams.IOStreams instead of *iostreams.ColorScheme, enabling access to terminal theme and width for improved markdown rendering. Refactored related code, tests, and mocks to support this change, and enhanced log rendering to better handle markdown and code output for various tool calls.
Introduces a new 'wantErrIs' field to test cases in Test_createRun to assert specific error types using require.ErrorIs. This enhances test coverage by verifying not only error messages but also error types.
Adds a test case to ensure that when a problem statement is provided as an argument non-interactively, the command does not prompt or return an error, and the correct job is created.
Modified the test in create_test.go to return and expect 'edited task description' instead of concatenating the prompt string. This clarifies the test's intent and expected behavior for the MarkdownEditorFunc and CreateJobFunc.
