ghstack 0.9.6__tar.gz → 0.10.0__tar.gz

{ghstack-0.9.6 → ghstack-0.10.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: ghstack
-Version: 0.9.6
+Version: 0.10.0
 Summary: Stack diff support for GitHub
 Author: Edward Z. Yang
 Author-email: Edward Z. Yang <ezyang@mit.edu>
@@ -23,7 +23,7 @@ Description-Content-Type: text/markdown
 Conveniently submit stacks of diffs to GitHub as separate pull requests.
 
 ```
-pip3 install ghstack
+uv tool install ghstack
 ```
 
 ghstack is tested with 
@@ -75,7 +75,8 @@ normal GitHub UI, as their branch bases won't be master.  Use
 a ghstack'ed pull request.
 
 You can also setup a GitHub action to allow a bot land.
-https://github.com/Chillee/ghstack_land_example/blob/main/.github/workflows/issue_comment.yml is an end-to-end example of how to do this.
+[ghstack_land_example](https://github.com/Chillee/ghstack_land_example/blob/main/.github/workflows/issue_comment.yml)
+is an end-to-end example of how to do this.
 
 ## Structure of submitted pull requests
 

{ghstack-0.9.6 → ghstack-0.10.0}/README.md

@@ -3,7 +3,7 @@
 Conveniently submit stacks of diffs to GitHub as separate pull requests.
 
 ```
-pip3 install ghstack
+uv tool install ghstack
 ```
 
 ghstack is tested with 
@@ -55,7 +55,8 @@ normal GitHub UI, as their branch bases won't be master.  Use
 a ghstack'ed pull request.
 
 You can also setup a GitHub action to allow a bot land.
-https://github.com/Chillee/ghstack_land_example/blob/main/.github/workflows/issue_comment.yml is an end-to-end example of how to do this.
+[ghstack_land_example](https://github.com/Chillee/ghstack_land_example/blob/main/.github/workflows/issue_comment.yml)
+is an end-to-end example of how to do this.
 
 ## Structure of submitted pull requests
 

{ghstack-0.9.6 → ghstack-0.10.0}/pyproject.toml

@@ -13,7 +13,7 @@ dependencies = [
     "flake8<8.0.0,>=7.0.0",
 ]
 name = "ghstack"
-version = "0.9.6"
+version = "0.10.0"
 description = "Stack diff support for GitHub"
 readme = "README.md"
 classifiers = [

{ghstack-0.9.6 → ghstack-0.10.0}/src/ghstack/land.py

@@ -168,7 +168,9 @@ to complain to the ghstack authors."""
             # TODO: regex here so janky
             base_ref = re.sub(r"/orig$", "/base", orig_ref)
             head_ref = re.sub(r"/orig$", "/head", orig_ref)
-            sh.git("push", remote_name, f"{remote_name}/{head_ref}:{base_ref}")
+            sh.git(
+                "push", remote_name, f"{remote_name}/{head_ref}:refs/heads/{base_ref}"
+            )
             github.notify_merged(pr_resolved)
 
         # All good! Push!

{ghstack-0.9.6 → ghstack-0.10.0}/src/ghstack/submit.py

@@ -15,6 +15,7 @@ import ghstack.github_utils
 import ghstack.gpg_sign
 import ghstack.logs
 import ghstack.shell
+import ghstack.trailers
 from ghstack.types import GhNumber, GitCommitHash, GitHubNumber, GitHubRepositoryId
 
 # Either "base", "head" or "orig"; which of the ghstack generated
@@ -963,20 +964,18 @@ to disassociate the commit with the pull request, and then try again.
             commit_msg = self._update_source_id(diff.summary, elab_diff)
         else:
             # Need to insert metadata for the first time
-            # TODO: Probably quicker if we reimplement reinterpret-trailers
-            # in Python
-            commit_msg = self.sh.git(
-                "interpret-trailers",
-                "--trailer",
-                f"ghstack-source-id: {diff.source_id}",
-                *(
-                    ["--trailer", f"ghstack-comment-id: {elab_diff.comment_id}"]
-                    if self.direct
-                    else []
-                ),
-                "--trailer",
-                f"Pull-Request-resolved: {pull_request_resolved.url()}",
-                input=strip_mentions(diff.summary.rstrip()),
+            # Using our Python implementation of interpret-trailers
+            trailers_to_add = [f"ghstack-source-id: {diff.source_id}"]
+
+            if self.direct:
+                trailers_to_add.append(f"ghstack-comment-id: {elab_diff.comment_id}")
+
+            trailers_to_add.append(
+                f"Pull-Request-resolved: {pull_request_resolved.url()}"
+            )
+
+            commit_msg = ghstack.trailers.interpret_trailers(
+                strip_mentions(diff.summary.rstrip()), trailers_to_add
             )
 
         return DiffMeta(

ghstack-0.10.0/src/ghstack/trailers.py

@@ -0,0 +1,191 @@
+#!/usr/bin/env python3
+
+import re
+from typing import List, Tuple
+
+# Compile regexes once at module level for better performance
+TRAILER_RE = re.compile(r"^([A-Za-z0-9_-]+)(\s*:\s*)(.*)$")
+CONTINUATION_RE = re.compile(r"^\s+\S.*$")
+
+# Git-generated trailer prefixes
+GIT_GENERATED_PREFIXES = ["Signed-off-by: ", "(cherry picked from commit "]
+
+
+def parse_message(message: str) -> Tuple[str, str, str]:
+    """
+    Parse a Git commit message into subject, body, and trailers.
+
+    According to the Git documentation, trailers are:
+    - A group of one or more lines that is all trailers, or contains at least one
+      Git-generated or user-configured trailer and consists of at least 25% trailers.
+    - The group must be preceded by one or more empty (or whitespace-only) lines.
+    - The group must either be at the end of the message or be the last non-whitespace
+      lines before a line that starts with "---" (the "divider").
+
+    Args:
+        message: The commit message to parse.
+
+    Returns:
+        A tuple containing:
+            - subject: The first line of the message
+            - body: The body of the message (may be empty)
+            - trailers: The trailer block as a raw string (may be empty)
+    """
+    if not message:
+        return "", "", ""
+
+    # Split into lines and get the subject (first line)
+    lines = message.splitlines()
+    subject = lines[0] if lines else ""
+
+    if len(lines) <= 1:
+        return subject, "", ""
+
+    # Remove subject
+    message_lines = lines[1:]
+
+    if not message_lines:
+        return subject, "", ""
+
+    # Find where the trailer block starts
+    trailer_start = find_trailer_block_start(message_lines)
+
+    if trailer_start == -1:
+        # No trailer block found, everything after subject is body
+        body = "\n".join(message_lines).strip()
+        return subject, body, ""
+
+    # Body is everything between subject and trailers (with empty lines trimmed)
+    body = "\n".join(message_lines[:trailer_start]).strip()
+
+    # Keep trailers as a raw string
+    trailers = "\n".join(message_lines[trailer_start:]).strip()
+
+    return subject, body, trailers
+
+
+def find_trailer_block_start(lines: List[str]) -> int:
+    """
+    Find the start index of the trailer block in a list of lines.
+
+    Args:
+        lines: List of message lines (without subject and divider).
+
+    Returns:
+        Index of the first line of the trailer block, or -1 if no trailer block is found.
+    """
+    # Remove trailing empty lines
+    trimmed_lines = list(reversed([line for line in reversed(lines) if line.strip()]))
+
+    if not trimmed_lines:
+        return -1
+
+    # Find the last non-empty block
+    block_indices = [-1] + [i for i, line in enumerate(lines) if not line.strip()]
+
+    # Try blocks from last to first
+    for i in range(len(block_indices) - 1, -1, -1):
+        start_idx = block_indices[i] + 1
+        # If we're at the beginning or checking the whole message
+        if i == 0 or start_idx == 0:
+            # Check if the whole remaining content is a trailer block
+            if is_trailer_block(lines[start_idx:]):
+                return start_idx
+            # No more blocks to check
+            return -1
+
+        # Check if the block after this blank line is a trailer block
+        end_idx = block_indices[i + 1] if i + 1 < len(block_indices) else len(lines)
+        if is_trailer_block(lines[start_idx:end_idx]):
+            return start_idx
+
+    return -1
+
+
+def is_trailer_block(lines: List[str]) -> bool:
+    """
+    Determine if the given lines form a trailer block.
+
+    A block is a trailer block if:
+    1. All lines are trailers, or
+    2. At least one Git-generated trailer exists and at least 25% of lines are trailers
+
+    Args:
+        lines: List of lines to check.
+
+    Returns:
+        True if the lines form a trailer block, False otherwise.
+    """
+    # Filter out empty lines
+    content_lines = [line for line in lines if line.strip()]
+
+    if not content_lines:
+        return False
+
+    trailer_lines = 0
+    non_trailer_lines = 0
+    has_git_generated_trailer = False
+
+    i = 0
+    while i < len(content_lines):
+        line = content_lines[i]
+
+        # Skip continuation lines (they belong to the previous trailer)
+        if CONTINUATION_RE.match(line):
+            i += 1
+            continue
+
+        # Check if it's a git-generated trailer
+        if any(line.startswith(prefix) for prefix in GIT_GENERATED_PREFIXES):
+            has_git_generated_trailer = True
+            trailer_lines += 1
+        elif TRAILER_RE.match(line):
+            # Regular trailer
+            trailer_lines += 1
+        else:
+            # Not a trailer line
+            non_trailer_lines += 1
+
+        i += 1
+
+    # A block is a trailer block if all lines are trailers OR
+    # it has at least one git-generated trailer and >= 25% of lines are trailers
+    return (trailer_lines > 0 and non_trailer_lines == 0) or (
+        has_git_generated_trailer and trailer_lines * 3 >= non_trailer_lines
+    )
+
+
+def interpret_trailers(message: str, trailers_to_add: List[str]) -> str:
+    """
+    Add trailers to a commit message, mimicking git interpret-trailers.
+
+    Args:
+        message: The commit message to add trailers to
+        trailers_to_add: List of trailers to add in the format "Key: Value"
+
+    Returns:
+        The commit message with trailers added
+    """
+    subject, body, existing_trailers = parse_message(message)
+
+    # Create a new list with all trailers (existing + new)
+    all_trailers = []
+    if existing_trailers:
+        all_trailers.append(existing_trailers)
+
+    all_trailers.extend(trailers_to_add)
+
+    # Build the new message
+    new_message = subject
+
+    if body:
+        new_message += "\n\n" + body
+
+    if all_trailers:
+        if body or (not body and existing_trailers):
+            new_message += "\n"
+        if not existing_trailers:
+            new_message += "\n"
+        new_message += "\n" + "\n".join(all_trailers)
+
+    return new_message