[Lldb-commits] [PATCH] D70177: [lldb] Fix that trailing backslashes in source lines break the Clang highlighter

[Raphael Isemann via Phabricator via lldb-commits]

teemperor created this revision.
teemperor added a reviewer: JDevlieghere.
Herald added a project: LLDB.
Herald added a subscriber: lldb-commits.

Clang's raw Lexer doesn't produce any tokens for trailing backslashes in a line. This doesn't work with
LLDB's Clang highlighter which builds the source code to display from the list of tokens the Lexer returns.
This causes that lines with trailing backslashes are lacking the backslash and the following newline when
rendering source code in LLDB.

This patch removes the trailing newline from the current line we are highlighting. This way Clang doesn't
drop the backslash token and we just restore the newline after tokenising.

Fixes rdar://57091487


Repository:
  rLLDB LLDB

https://reviews.llvm.org/D70177

Files:
  lldb/source/Plugins/Language/ClangCommon/ClangHighlighter.cpp
  lldb/unittests/Language/Highlighting/HighlighterTest.cpp


Index: lldb/unittests/Language/Highlighting/HighlighterTest.cpp
===================================================================
--- lldb/unittests/Language/Highlighting/HighlighterTest.cpp
+++ lldb/unittests/Language/Highlighting/HighlighterTest.cpp
@@ -205,6 +205,43 @@
             highlightC("#include \"foo\" //c", s));
 }
 
+TEST_F(HighlighterTest, ClangPreserveNewLine) {
+  HighlightStyle s;
+  s.comment.Set("<cc>", "</cc>");
+
+  EXPECT_EQ("<cc>//</cc>\n", highlightC("//\n", s));
+}
+
+TEST_F(HighlighterTest, ClangTrailingBackslashBeforeNewline) {
+  HighlightStyle s;
+
+  EXPECT_EQ("\\\n", highlightC("\\\n", s));
+  EXPECT_EQ("\\\r\n", highlightC("\\\r\n", s));
+
+  EXPECT_EQ("#define a \\\n", highlightC("#define a \\\n", s));
+  EXPECT_EQ("#define a \\\r\n", highlightC("#define a \\\r\n", s));
+}
+
+TEST_F(HighlighterTest, ClangTrailingBackslashWithWhitespace) {
+  HighlightStyle s;
+
+  EXPECT_EQ("\\  \n", highlightC("\\  \n", s));
+  EXPECT_EQ("\\ \t\n", highlightC("\\ \t\n", s));
+  EXPECT_EQ("\\ \n", highlightC("\\ \n", s));
+  EXPECT_EQ("\\\t\n", highlightC("\\\t\n", s));
+
+  EXPECT_EQ("#define a \\  \n", highlightC("#define a \\  \n", s));
+  EXPECT_EQ("#define a \\ \t\n", highlightC("#define a \\ \t\n", s));
+  EXPECT_EQ("#define a \\ \n", highlightC("#define a \\ \n", s));
+  EXPECT_EQ("#define a \\\t\n", highlightC("#define a \\\t\n", s));
+}
+
+TEST_F(HighlighterTest, ClangTrailingBackslashMissingNewLine) {
+  HighlightStyle s;
+  EXPECT_EQ("\\", highlightC("\\", s));
+  EXPECT_EQ("#define a\\", highlightC("#define a\\", s));
+}
+
 TEST_F(HighlighterTest, ClangComments) {
   HighlightStyle s;
   s.comment.Set("<cc>", "</cc>");
Index: lldb/source/Plugins/Language/ClangCommon/ClangHighlighter.cpp
===================================================================
--- lldb/source/Plugins/Language/ClangCommon/ClangHighlighter.cpp
+++ lldb/source/Plugins/Language/ClangCommon/ClangHighlighter.cpp
@@ -139,6 +139,16 @@
   FileManager file_mgr(file_opts,
                        FileSystem::Instance().GetVirtualFileSystem());
 
+  // The line might end in a backslash which would cause Clang to drop the
+  // backslash and the terminating line feed. This makes sense when parsing C++,
+  // but when highlighting we care about preserving the backslash/newline. To
+  // not lose this information we remove the line feed here so that Clang knows
+  // this is just a single line we are highlighting. We add back the newline
+  // after tokenizing.
+  const bool line_had_cr_lf = line.endswith("\r\n");
+  const bool line_had_lf = line.endswith("\n") && !line_had_cr_lf;
+  line = line.trim("\r\n");
+
   unsigned line_number = previous_lines.count('\n') + 1U;
 
   // Let's build the actual source code Clang needs and setup some utility
@@ -227,6 +237,11 @@
     color.Apply(result, to_print);
   }
 
+  if (line_had_cr_lf)
+    result << "\r\n";
+  else if (line_had_lf)
+    result << "\n";
+
   // If we went over the whole file but couldn't find our own file, then
   // somehow our setup was wrong. When we're in release mode we just give the
   // user the normal line and pretend we don't know how to highlight it. In


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D70177.229080.patch
Type: text/x-patch
Size: 3174 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/lldb-commits/attachments/20191113/89d6e54f/attachment-0001.bin>