From e36874cb5478dbb96ef149d700cf7bcfb50ad6c3 Mon Sep 17 00:00:00 2001
From: Michael Hohn <hohn@github.com>
Date: Mon, 15 Nov 2021 14:16:23 -0800
Subject: [PATCH] sarif-results-summary: underline affected code region

Using
    sarif-results-summary -s data/linux-small data/torvalds_linux__2021-10-21_10_07_00__export.sarif |less
now underscores the indicated regions, e.g.

tools/cgroup/iocost_monitor.py:64:5:64:27: Normal methods should have 'self', rather than 'blkcg', as their first parameter.

    def blkcg_name(blkcg):
    ^^^^^^^^^^^^^^^^^^^^^^
---
 bin/sarif-results-summary | 14 ++++++++++++--
 sarif_cli/__init__.py     | 39 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 51 insertions(+), 2 deletions(-)

diff --git a/bin/sarif-results-summary b/bin/sarif-results-summary
index 6bb4c2c..c4e6677 100755
--- a/bin/sarif-results-summary
+++ b/bin/sarif-results-summary
@@ -2,6 +2,7 @@
 import argparse
 import json
 import sarif_cli as S
+import re
 import sys
 import collections
 
@@ -20,7 +21,7 @@ for runi in S.indices(sarif_struct, 'runs'):
                      'semmle.sourceLanguage')
     S.msg("Found %d results for %s\n" %  (num_results, language))
     if num_results == 0: continue
-
+    # 
     for resi in S.indices(sarif_struct, 'runs', runi, 'results'):
         message = S.get(sarif_struct, 'runs', runi, 'results', resi, 'message', 'text')
         artifact = S.get(sarif_struct, 'runs', runi, 'results', resi, 'locations', 0,
@@ -32,7 +33,16 @@ for runi in S.indices(sarif_struct, 'runs'):
         S.msg("%s: %s\n\n" % (filepath, message))
         if args.list_source:
             lines = S.load_lines(args.list_source, artifact['uri'], l1, l2)
-            for line in lines:
+            for line, line_num in zip(lines, range(l1, l2+1)):
+                # Display the line
                 S.msg("%s" % (line))
                 S.msg("\n")
+                # Print the underline
+                underline = S.underline_for_result(l1, c1, l2, c2, line, line_num)
+                S.msg(underline)
+                # Next result
+                S.msg("\n")
         S.msg("\n")
+
+
+
diff --git a/sarif_cli/__init__.py b/sarif_cli/__init__.py
index 511407d..9ba930a 100644
--- a/sarif_cli/__init__.py
+++ b/sarif_cli/__init__.py
@@ -1,10 +1,49 @@
 import sys
 import os
+import re
 
 MIN_PYTHON = (3, 7)
 if sys.version_info < MIN_PYTHON:
     sys.exit("Python %s.%s or later is required.\n" % MIN_PYTHON)
 
+def underline_for_result(first_line, first_column, last_line, last_column, line, line_num):
+    """Provide the underline for a result line.
+
+    first_line, first_column, last_line, last_column : 
+        the region from S.lineinfo(region)
+    line: 
+        the line of source
+    line_num:  
+        the index of line, must satisfy first_line <= line_num <= last_line
+    """
+    # Underline the affected region
+    # col_* use the [start, end) indexing
+    #   From the first non-whitespace char
+    match = re.search("([^\s])+", line)
+    if match:
+        col_from = match.span()[0]
+    else:
+        col_from = 0
+    #   To the last non-whitespace char
+    match = re.search("(\s)+$", line)
+    if match:
+        col_to = match.span()[0]
+    else:
+        col_to = len(line)
+    #   Use 1-indexing
+    col_from += 1 ; col_to += 1
+    #   Adjust first line
+    if line_num == first_line:
+        col_from = max(col_from, first_column)
+    #   Adjust last line
+    if line_num == last_line:
+        col_to = min(col_to, last_column)
+    #   Use 0-indexing
+    col_from -= 1 ; col_to -= 1
+    # Return the underline
+    return " " * col_from + "^" * (col_to - col_from)
+    
+
 def load_lines(root, path, line_from, line_to):
     """Load the line range [line_from, line_to], including both, 
     from the file at root/path.