V0.90.4

CONTRIBUTING.md

@@ -64,11 +64,14 @@ docker build -t cecli -f docker/Dockerfile .
 
 ## Coding Standards
 
-It really helps the merge process if your PR:
-
-1. complies with project coding standards
-2. includes test coverage
-3. updates the relevant user-facing documentation, including the output of `/help` and `--help` as well as notes in config files and the web-site.
+In order for your PR to be accepted it must:
+
+1. Be up to date with the main branch
+2. Comply with project coding standards (including running the pre-commit formatting hooks)
+3. Include test coverage
+4. Update relevant user-facing documentation:
+   - Primary documentation will live in `aider/website/docs/config/`
+   - Check new cli arguments with the output of `/help` and `--help`
 
 ### Python Compatibility
 

aider/__init__.py

@@ -1,6 +1,6 @@
 from packaging import version
 
-__version__ = "0.90.0.dev"
+__version__ = "0.90.4.dev"
 safe_version = __version__
 
 try:

aider/args.py

@@ -246,8 +246,8 @@ def get_parser(default_config_files, git_root):
     group = parser.add_argument_group("TUI Settings")
     group.add_argument(
         "--tui",
-        action="store_true",
-        default=False,
+        action=argparse.BooleanOptionalAction,
+        default=None,
         help="Launch Textual TUI interface (experimental)",
     )
     group.add_argument(
@@ -718,7 +718,7 @@ def get_parser(default_config_files, git_root):
         "--check-update",
         action=argparse.BooleanOptionalAction,
         help="Check for new aider versions on launch",
-        default=False,
+        default=True,
     )
     group.add_argument(
         "--show-release-notes",
@@ -803,10 +803,8 @@ def get_parser(default_config_files, git_root):
     )
     group.add_argument(
         "--linear-output",
-        action="store_true",
-        help=(
-            "Run input and output sequentially instead of us simultaneous streams (default: False)"
-        ),
+        action=argparse.BooleanOptionalAction,
+        help="Run input and output sequentially instead of us simultaneous streams (default: True)",
         default=True,
     )
     group.add_argument(

aider/coders/__init__.py

@@ -3,6 +3,7 @@
 from .ask_coder import AskCoder
 from .base_coder import Coder
 from .context_coder import ContextCoder
+from .copypaste_coder import CopyPasteCoder
 from .editblock_coder import EditBlockCoder
 from .editblock_fenced_coder import EditBlockFencedCoder
 from .editor_diff_fenced_coder import EditorDiffFencedCoder
@@ -33,4 +34,5 @@
     EditorDiffFencedCoder,
     ContextCoder,
     AgentCoder,
+    CopyPasteCoder,
 ]

aider/coders/agent_coder.py

@@ -81,12 +81,9 @@ class AgentCoder(Coder):
     """Mode where the LLM autonomously manages which files are in context."""
 
     edit_format = "agent"
+    gpt_prompts = AgentPrompts()
 
     def __init__(self, *args, **kwargs):
-        # Initialize appropriate prompt set before calling parent constructor
-        # This needs to happen before super().__init__ so the parent class has access to gpt_prompts
-        self.gpt_prompts = AgentPrompts()
-
         # Dictionary to track recently removed files
         self.recently_removed = {}
 
@@ -1203,6 +1200,8 @@ async def reply_completed(self):
 
         if self.agent_finished:
             self.tool_usage_history = []
+            if self.files_edited_by_tools:
+                _ = await self.auto_commit(self.files_edited_by_tools)
             return True
 
         # Since we are no longer suppressing, the partial_response_content IS the final content.

aider/coders/agent_prompts.py

@@ -17,9 +17,8 @@ class AgentPrompts(CoderPrompts):
 ## Core Directives
 - **Role**: Act as an expert software engineer.
 - **Act Proactively**: Autonomously use file discovery and context management tools (`ViewFilesAtGlob`, `ViewFilesMatching`, `Ls`, `View`, `Remove`) to gather information and fulfill the user's request. Chain tool calls across multiple turns to continue exploration.
-- **Be Decisive**: Do not ask the same question or search for the same term in multiple ways. Trust that your initial findings are valid.
-- **Be Concise**: Keep all responses brief and direct (1-3 sentences). Avoid preamble, postamble, and unnecessary explanations.
-- **Confirm Ambiguity**: Before applying complex or ambiguous edits, briefly state your plan. For simple, direct edits, proceed without confirmation.
+- **Be Decisive**: Trust that your initial findings are valid. Refrain from asking the same question or searching for the same term in multiple similar ways.
+- **Be Concise**: Keep all responses brief and direct (1-3 sentences). Avoid preamble, postamble, and unnecessary explanations. Do not repeat yourself.
 </context>
 
 <context name="workflow_and_tool_usage">

aider/coders/base_coder.py

@@ -172,7 +172,7 @@ async def create(
             if from_coder:
                 main_model = from_coder.main_model
             else:
-                main_model = models.Model(models.DEFAULT_MODEL_NAME)
+                main_model = models.Model(models.DEFAULT_MODEL_NAME, io=io)
 
         if edit_format == "code":
             edit_format = None
@@ -229,6 +229,15 @@ async def create(
             kwargs = use_kwargs
             from_coder.ok_to_warm_cache = False
 
+        if (
+            getattr(main_model, "copy_paste_mode", False)
+            and getattr(main_model, "copy_paste_transport", "api") == "clipboard"
+        ):
+            res = coders.CopyPasteCoder(main_model, io, args=args, **kwargs)
+            await res.initialize_mcp_tools()
+            res.original_kwargs = dict(kwargs)
+            return res
+
         for coder in coders.__all__:
             if hasattr(coder, "edit_format") and coder.edit_format == edit_format:
                 res = coder(main_model, io, args=args, **kwargs)
@@ -379,6 +388,9 @@ def __init__(
         self.io = io
         self.io.coder = weakref.ref(self)
 
+        self.manual_copy_paste = getattr(main_model, "copy_paste_transport", "api") == "clipboard"
+        self.copy_paste_mode = getattr(main_model, "copy_paste_mode", False) or auto_copy_context
+
         self.shell_commands = []
         self.partial_response_tool_calls = []
 
@@ -399,7 +411,7 @@ def __init__(
             self.main_model.reasoning_tag if self.main_model.reasoning_tag else REASONING_TAG
         )
 
-        self.stream = stream and main_model.streaming
+        self.stream = stream and main_model.streaming and not self.manual_copy_paste
 
         if cache_prompts and self.main_model.cache_control:
             self.add_cache_headers = True
@@ -581,6 +593,8 @@ def get_announcements(self):
             output += ", prompt cache"
         if main_model.info.get("supports_assistant_prefill"):
             output += ", infinite output"
+        if self.copy_paste_mode:
+            output += ", copy/paste mode"
 
         lines.append(output)
 
@@ -639,7 +653,7 @@ def get_announcements(self):
         if self.done_messages:
             lines.append("Restored previous conversation history.")
 
-        if self.io.multiline_mode:
+        if self.io.multiline_mode and not self.args.tui:
             lines.append("Multiline mode: Enabled. Enter inserts newline, Alt-Enter submits text")
 
         return lines
@@ -2823,8 +2837,8 @@ def add_assistant_reply_to_cur_messages(self):
                 # but response.dict() is the Pydantic V1 method name.
                 response_dict = dict(response)
             except TypeError:
-                print("Neither model_dump() nor dict() worked as expected.")
-                raise
+                print("Response parsing error.")
+                return
 
         msg = response_dict["choices"][0]["message"]