lint: Convert lint-whitespace.sh to Python

2 files changed, 135 insertions, 115 deletions

diff --git a/test/lint/lint-whitespace.py b/test/lint/lint-whitespace.py
new file mode 100755
index 0000000000..d98fc8d9a2
--- /dev/null
+++ b/test/lint/lint-whitespace.py
@@ -0,0 +1,135 @@
+#!/usr/bin/env python3
+#
+# Copyright (c) 2017-2022 The Bitcoin Core developers
+# Distributed under the MIT software license, see the accompanying
+# file COPYING or http://www.opensource.org/licenses/mit-license.php.
+#
+# Check for new lines in diff that introduce trailing whitespace or
+# tab characters instead of spaces.
+
+# We can't run this check unless we know the commit range for the PR.
+
+import argparse
+import os
+import re
+import sys
+
+from subprocess import check_output
+
+EXCLUDED_DIRS = ["depends/patches/",
+                 "contrib/guix/patches/",
+                 "src/leveldb/",
+                 "src/crc32c/",
+                 "src/secp256k1/",
+                 "src/minisketch/",
+                 "src/univalue/",
+                 "doc/release-notes/",
+                 "src/qt/locale"]
+
+def parse_args():
+    """Parse command line arguments."""
+    parser = argparse.ArgumentParser(
+        description="""
+            Check for new lines in diff that introduce trailing whitespace
+            or tab characters instead of spaces in unstaged changes, the
+            previous n commits, or a commit-range.
+        """,
+        epilog=f"""
+            You can manually set the commit-range with the COMMIT_RANGE
+            environment variable (e.g. "COMMIT_RANGE='47ba2c3...ee50c9e'
+            {sys.argv[0]}"). Defaults to current merge base when neither
+            prev-commits nor the environment variable is set.
+        """)
+
+    parser.add_argument("--prev-commits", "-p", required=False, help="The previous n commits to check")
+
+    return parser.parse_args()
+
+
+def report_diff(selection):
+    filename = ""
+    seen = False
+    seenln = False
+
+    print("The following changes were suspected:")
+
+    for line in selection:
+        if re.match(r"^diff", line):
+            filename = line
+            seen = False
+        elif re.match(r"^@@", line):
+            linenumber = line
+            seenln = False
+        else:
+            if not seen:
+                # The first time a file is seen with trailing whitespace or a tab character, we print the
+                # filename (preceded by a newline).
+                print("")
+                print(filename)
+                seen = True
+            if not seenln:
+                print(linenumber)
+                seenln = True
+            print(line)
+
+
+def get_diff(commit_range, check_only_code):
+    exclude_args = [":(exclude)" + dir for dir in EXCLUDED_DIRS]
+
+    if check_only_code:
+        what_files = ["*.cpp", "*.h", "*.md", "*.py", "*.sh"]
+    else:
+        what_files = ["."]
+
+    diff = check_output(["git", "diff", "-U0", commit_range, "--"] + what_files + exclude_args, universal_newlines=True, encoding="utf8")
+
+    return diff
+
+
+def main():
+    args = parse_args()
+
+    if not os.getenv("COMMIT_RANGE"):
+        if args.prev_commits:
+            commit_range = "HEAD~" + args.prev_commits + "...HEAD"
+        else:
+            # This assumes that the target branch of the pull request will be master.
+            merge_base = check_output(["git", "merge-base", "HEAD", "master"], universal_newlines=True, encoding="utf8").rstrip("\n")
+            commit_range = merge_base + "..HEAD"
+    else:
+        commit_range = os.getenv("COMMIT_RANGE")
+
+    whitespace_selection = []
+    tab_selection = []
+
+    # Check if trailing whitespace was found in the diff.
+    for line in get_diff(commit_range, check_only_code=False).splitlines():
+        if re.match(r"^(diff --git|\@@|^\+.*\s+$)", line):
+            whitespace_selection.append(line)
+
+    whitespace_additions = [i for i in whitespace_selection if i.startswith("+")]
+
+    # Check if tab characters were found in the diff.
+    for line in get_diff(commit_range, check_only_code=True).splitlines():
+        if re.match(r"^(diff --git|\@@|^\+.*\t)", line):
+            tab_selection.append(line)
+
+    tab_additions = [i for i in tab_selection if i.startswith("+")]
+
+    ret = 0
+
+    if len(whitespace_additions) > 0:
+        print("This diff appears to have added new lines with trailing whitespace.")
+        report_diff(whitespace_selection)
+        ret = 1
+
+    if len(tab_additions) > 0:
+        print("This diff appears to have added new lines with tab characters instead of spaces.")
+        report_diff(tab_selection)
+        ret = 1
+
+    sys.exit(ret)
+
+
+if __name__ == "__main__":
+    main()
diff --git a/test/lint/lint-whitespace.sh b/test/lint/lint-whitespace.sh
deleted file mode 100755
index 9d55c71eb5..0000000000
--- a/test/lint/lint-whitespace.sh
+++ /dev/null
@@ -1,115 +0,0 @@
-#!/usr/bin/env bash
-#
-# Copyright (c) 2017-2021 The Bitcoin Core developers
-# Distributed under the MIT software license, see the accompanying
-# file COPYING or http://www.opensource.org/licenses/mit-license.php.
-#
-# Check for new lines in diff that introduce trailing whitespace.
-
-# We can't run this check unless we know the commit range for the PR.
-
-export LC_ALL=C
-while getopts "?" opt; do
-  case $opt in
-    ?)
-      echo "Usage: $0 [N]"
-      echo "       COMMIT_RANGE='<commit range>' $0"
-      echo "       $0 -?"
-      echo "Checks unstaged changes, the previous N commits, or a commit range."
-      echo "COMMIT_RANGE='47ba2c3...ee50c9e' $0"
-      exit 0
-    ;;
-  esac
-done
-
-if [ -z "${COMMIT_RANGE}" ]; then
-  if [ -n "$1" ]; then
-    COMMIT_RANGE="HEAD~$1...HEAD"
-  else
-    # This assumes that the target branch of the pull request will be master.
-    MERGE_BASE=$(git merge-base HEAD master)
-    COMMIT_RANGE="$MERGE_BASE..HEAD"
-  fi
-fi
-
-showdiff() {
-  if ! git diff -U0 "${COMMIT_RANGE}" -- "." ":(exclude)depends/patches/" ":(exclude)contrib/guix/patches/" ":(exclude)src/leveldb/" ":(exclude)src/crc32c/" ":(exclude)src/secp256k1/" ":(exclude)src/minisketch/" ":(exclude)src/univalue/" ":(exclude)doc/release-notes/" ":(exclude)src/qt/locale/"; then
-    echo "Failed to get a diff"
-    exit 1
-  fi
-}
-
-showcodediff() {
-  if ! git diff -U0 "${COMMIT_RANGE}" -- *.cpp *.h *.md *.py *.sh ":(exclude)src/leveldb/" ":(exclude)src/crc32c/" ":(exclude)src/secp256k1/" ":(exclude)src/minisketch/" ":(exclude)src/univalue/" ":(exclude)doc/release-notes/" ":(exclude)src/qt/locale/"; then
-    echo "Failed to get a diff"
-    exit 1
-  fi
-}
-
-RET=0
-
-# Check if trailing whitespace was found in the diff.
-if showdiff | grep -E -q '^\+.*\s+$'; then
-  echo "This diff appears to have added new lines with trailing whitespace."
-  echo "The following changes were suspected:"
-  FILENAME=""
-  SEEN=0
-  SEENLN=0
-  while read -r line; do
-    if [[ "$line" =~ ^diff ]]; then
-      FILENAME="$line"
-      SEEN=0
-    elif [[ "$line" =~ ^@@ ]]; then
-      LINENUMBER="$line"
-      SEENLN=0
-    else
-      if [ "$SEEN" -eq 0 ]; then
-        # The first time a file is seen with trailing whitespace, we print the
-        # filename (preceded by a newline).
-        echo
-        echo "$FILENAME"
-        SEEN=1
-      fi
-      if [ "$SEENLN" -eq 0 ]; then
-        echo "$LINENUMBER"
-        SEENLN=1
-      fi
-      echo "$line"
-    fi
-  done < <(showdiff | grep -E '^(diff --git |@@|\+.*\s+$)')
-  RET=1
-fi
-
-# Check if tab characters were found in the diff.
-if showcodediff | perl -nle '$MATCH++ if m{^\+.*\t}; END{exit 1 unless $MATCH>0}' > /dev/null; then
-  echo "This diff appears to have added new lines with tab characters instead of spaces."
-  echo "The following changes were suspected:"
-  FILENAME=""
-  SEEN=0
-  SEENLN=0
-  while read -r line; do
-    if [[ "$line" =~ ^diff ]]; then
-      FILENAME="$line"
-      SEEN=0
-    elif [[ "$line" =~ ^@@ ]]; then
-      LINENUMBER="$line"
-      SEENLN=0
-    else
-      if [ "$SEEN" -eq 0 ]; then
-        # The first time a file is seen with a tab character, we print the
-        # filename (preceded by a newline).
-        echo
-        echo "$FILENAME"
-        SEEN=1
-      fi
-      if [ "$SEENLN" -eq 0 ]; then
-        echo "$LINENUMBER"
-        SEENLN=1
-      fi
-      echo "$line"
-    fi
-  done < <(showcodediff | perl -nle 'print if m{^(diff --git |@@|\+.*\t)}')
-  RET=1
-fi
-
-exit $RET