[DebugInfo] Simplify string table dumpers.

[Higuoxing]

Summary

This patch adds a helper function DumpStrSection to simplify codes.
Besides, nonprintable chars in debug_str and debug_str.dwo sections
are printed as escaped chars.

[Higuoxing]

After this patch, the dumpers of debug_str, debug_str.dwo and debug_line_str sections are exactly the same one. Can we merge tests into a single file?

[jhenderson]

I would suggest merging debug-str.yaml and debug-str-dwo.yaml. The only difference is the section name. You could even show that both are dumped if in the same input file.

I think keeping .debug_line_str separate is still wise, because it uses a different switch to get dumped. I don't feel strongly about it though, so am happy if others think it should be merged into a single test. I don't have an answer about what to do about the duplicate coverage of the escaping stuff.

llvm/test/tools/llvm-dwarfdump/debug-str.yaml

0–1

My same comments apply here as in D86916.

[Higuoxing]

Address comments.

Thanks!

[Higuoxing]

rm llvm/test/tools/llvm-dwarfdump/debug-str-dwo.yaml

[jhenderson]

LGTM.

llvm/test/tools/llvm-dwarfdump/debug-str.yaml

1–3

[Higuoxing]

Address comments.

llvm/test/tools/llvm-dwarfdump/debug-str.yaml

1–3

Thanks!

llvm/lib/DebugInfo/DWARF/DWARFContext.cpp

while (!Parser.done()) {
       }
       OS << "debug_line[" << format("0x%8.8" PRIx64, Parser.getOffset())
          << "]\n";
       Parser.parseNext(DumpOpts.WarningHandler, DumpOpts.WarningHandler, &OS,
                        DumpOpts.Verbose);
     }
   };
 
+  auto DumpStrSection = [&](StringRef Section) {
+    DataExtractor StrData(Section, isLittleEndian(), 0);
+    uint64_t Offset = 0;
+    uint64_t StrOffset = 0;
+    while (const char *CStr = StrData.getCStr(&Offset)) {
+      OS << format("0x%8.8" PRIx64 ": \"", StrOffset);
+      OS.write_escaped(CStr);
+      OS << "\"\n";
+      StrOffset = Offset;
+    }
+  };
+
   if (const auto *Off = shouldDump(Explicit, ".debug_line", DIDT_ID_DebugLine,
                                    DObj->getLineSection().Data)) {
     DWARFDataExtractor LineData(*DObj, DObj->getLineSection(), isLittleEndian(),
                                 0);
     DWARFDebugLine::SectionParser Parser(LineData, *this, compile_units(),
                                          type_units());
     DumpLineSection(Parser, DumpOpts, *Off);
   }
void DWARFContext::dump(
   }
 
   if (shouldDump(Explicit, ".debug_tu_index", DIDT_ID_DebugTUIndex,
                  DObj->getTUIndexSection())) {
     getTUIndex().dump(OS);
   }
 
   if (shouldDump(Explicit, ".debug_str", DIDT_ID_DebugStr,
-                 DObj->getStrSection())) {
-    DataExtractor strData(DObj->getStrSection(), isLittleEndian(), 0);
-    uint64_t offset = 0;
-    uint64_t strOffset = 0;
-    while (const char *s = strData.getCStr(&offset)) {
-      OS << format("0x%8.8" PRIx64 ": \"%s\"\n", strOffset, s);
-      strOffset = offset;
-    }
-  }
+                 DObj->getStrSection()))
+    DumpStrSection(DObj->getStrSection());
+
   if (shouldDump(ExplicitDWO, ".debug_str.dwo", DIDT_ID_DebugStr,
-                 DObj->getStrDWOSection())) {
-    DataExtractor strDWOData(DObj->getStrDWOSection(), isLittleEndian(), 0);
-    uint64_t offset = 0;
-    uint64_t strDWOOffset = 0;
-    while (const char *s = strDWOData.getCStr(&offset)) {
-      OS << format("0x%8.8" PRIx64 ": \"%s\"\n", strDWOOffset, s);
-      strDWOOffset = offset;
-    }
-  }
+                 DObj->getStrDWOSection()))
+    DumpStrSection(DObj->getStrDWOSection());
+
   if (shouldDump(Explicit, ".debug_line_str", DIDT_ID_DebugLineStr,
-                 DObj->getLineStrSection())) {
-    DataExtractor strData(DObj->getLineStrSection(), isLittleEndian(), 0);
-    uint64_t offset = 0;
-    uint64_t strOffset = 0;
-    while (const char *s = strData.getCStr(&offset)) {
-      OS << format("0x%8.8" PRIx64 ": \"", strOffset);
-      OS.write_escaped(s);
-      OS << "\"\n";
-      strOffset = offset;
-    }
-  }
+                 DObj->getLineStrSection()))
+    DumpStrSection(DObj->getLineStrSection());
 
   if (shouldDump(Explicit, ".debug_addr", DIDT_ID_DebugAddr,
                  DObj->getAddrSection().Data)) {
     DWARFDataExtractor AddrData(*DObj, DObj->getAddrSection(),
                                    isLittleEndian(), 0);
     dumpAddrSection(OS, AddrData, DumpOpts, getMaxVersion(), getCUAddrSize());
   }
 

llvm/test/tools/llvm-dwarfdump/debug-str.yaml

-## Show that llvm-dwarfdump dumps the .debug_str section when --debug-str is
-## specified.
+## a) Show that llvm-dwarfdump dumps the .debug_str and
+## .debug_str.dwo sections when --debug-str is specified.
 
-# RUN: yaml2obj %s -o %t.o
+##                        "some string\0foo\0\0"
+# RUN: yaml2obj -DCONTENT="736f6d6520737472696e6700666f6f0000" %s -o %t.o
 # RUN: llvm-dwarfdump %t.o --debug-str | FileCheck %s
 
-# CHECK:      .debug_str contents:
-# CHECK-NEXT: 0x00000000: "some string"
-# CHECK-NEXT: 0x0000000c: "foo"
-# CHECK-NEXT: 0x00000010: ""
-# CHECK-NOT:  {{.}}
+#  CHECK:      .debug_str contents:
+#  CHECK-NEXT: 0x00000000: "some string"
+#  CHECK-NEXT: 0x0000000c: "foo"
+#  CHECK-NEXT: 0x00000010: ""
+#  CHECK:      .debug_str.dwo contents:
+#  CHECK-NEXT: 0x00000000: "some string"
+#  CHECK-NEXT: 0x0000000c: "foo"
+#  CHECK-NEXT: 0x00000010: ""
+# CHECK-EMPTY:
 
 --- !ELF
 FileHeader:
   Class:   ELFCLASS64
   Data:    ELFDATA2LSB
   Type:    ET_REL
   Machine: EM_X86_64
 Sections:
-  - Name: .debug_str
-    Type: SHT_PROGBITS
-    Content: '736f6d6520737472696e6700666f6f0000' ## "some string\0foo\0\0"
+  - Name:    .debug_str
+    Type:    SHT_PROGBITS
+    Content: [[CONTENT]]
+  - Name:    .debug_str.dwo
+    Type:    SHT_PROGBITS
+    Content: [[CONTENT]]
+
+## b) Test how we dump unprintable chars.
+
+##                        ['\t', '\0', '\001', '\0', '\\', '0', '0', '1', '\0']
+# RUN: yaml2obj -DCONTENT="090001005C30303100" %s -o %t2.o
+# RUN: llvm-dwarfdump --debug-str %t2.o | FileCheck %s --check-prefix=ESCAPED
+
+#       ESCAPED: .debug_str contents:
+#  ESCAPED-NEXT: 0x00000000: "\t"
+#  ESCAPED-NEXT: 0x00000002: "\001"
+#  ESCAPED-NEXT: 0x00000004: "\\001"
+#       ESCAPED: .debug_str.dwo contents:
+#  ESCAPED-NEXT: 0x00000000: "\t"
+#  ESCAPED-NEXT: 0x00000002: "\001"
+#  ESCAPED-NEXT: 0x00000004: "\\001"
+# ESCAPED-EMPTY: