[analyzer] Teach scan-build how to rebuild index.html without analyzing.

[NoQ]

Summary

This is useful for performing custom build system integration that works by appending '--analyze --analyzer-output html' to all clang build commands. In this case there is now still a way to have a fancy index.html with the output.

[NoQ]

clang/test/Analysis/scan-build/rebuild_index/rebuild_index.test

11–12

The script is indeed capable of finding html reports in sub-directories and it's a fairly useful feature because in a custom build system integration it's likely that clang's -o flag will be very messed up.

[Charusso]

Thanks! I would mention in the Summary the necessary flags to perform index-only output. (May write some release notes?)

That option sounds very strange, but I like it. For example to run only necessary custom tests one could write:

-analyzer-checker=core,unix,custom.checkers \
-analyzer-config silence-checkers="core;unix"

clang/tools/scan-build/bin/scan-build

953

May inject Diag "Analysis run complete.\n"; here?

976

If we stick to printing information to the user: Diag "No bugs found.\n"?.

1303

What about --index-only?

1618

I may smash this duplicated error-handling above in one if-stmt, given that the options are correlate with each other.

[NoQ]

clang/tools/scan-build/bin/scan-build

953

Indeed!

976

We already kind of have it when it says "deleting output directory because it contains no reports".

1303

Renamed to --generate-index-only :)

1618

Dunno, i'd rather keep the structure. The function is huge, i'd rather keep it well-structured, and "each option is handled in exactly once if-statement" is good for it.

[NoQ]

Reverted due to a lot of buildbot failures, eg. http://lab.llvm.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/builds/62758/steps/test-check-all/logs/stdio

$ "/home/buildslave/ps4-buildslave4/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm-project/clang/tools/scan-build/bin/scan-build" "--use-analyzer=/home/buildslave/ps4-buildslave4/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/build/bin/clang" "--generate-index-only" "/home/buildslave/ps4-buildslave4/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/build/tools/clang/test/Analysis/scan-build/rebuild_index/Output/rebuild_index.test.tmp.output_dir"
# command output:
scan-build: Analysis run complete.
scan-build: 3 bugs found.
scan-build: Run 'scan-view /home/buildslave/ps4-buildslave4/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/build/tools/clang/test/Analysis/scan-build/rebuild_index/Output/rebuild_index.test.tmp.output_dir' to examine bug reports.

$ ":" "RUN: at line 16"
$ "ls" "/home/buildslave/ps4-buildslave4/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/build/tools/clang/test/Analysis/scan-build/rebuild_index/Output/rebuild_index.test.tmp.output_dir"
$ "/home/buildslave/ps4-buildslave4/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/build/bin/FileCheck" "-check-prefix" "CHECK-FILES" "/home/buildslave/ps4-buildslave4/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm-project/clang/test/Analysis/scan-build/rebuild_index/rebuild_index.test"
# command stderr:
/home/buildslave/ps4-buildslave4/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm-project/clang/test/Analysis/scan-build/rebuild_index/rebuild_index.test:19:19: error: CHECK-FILES-NEXT: expected string not found in input
CHECK-FILES-NEXT: report-1.html
                  ^
<stdin>:2:1: note: scanning from here
report-2.html
^

error: command failed with exit status: 1

Weird, all three bugs found but report-1 is not present, is it just that ls doesn't sort the output or is it about our deduplication removing a different duplicate?..

[Charusso]

That is very unfortunate, but may if you could introduce a bullet-proof ls we could see if the scan-build sub-directory removal is non-alphabetical. I think the latter smells more badly.

clang/test/Analysis/scan-build/rebuild_index/rebuild_index.test

17

If you think that the ls is the problem may we need ls -R to print out every folder. (https://explainshell.com/explain?cmd=ls+-R)

There are more exotic ways to sort the order, like LANG=C ls: https://stackoverflow.com/questions/878249/unixs-ls-sort-by-name/878269

[Charusso]

Could you mention how to use this feature in the Summary please?

cd reports
scan-build --generate-index-only .

---

And something is not right, I have tried it:

Use of uninitialized value $Clang in concatenation (.) or string at /llvm-project/clang/tools/scan-build/bin/scan-build line 1895.
sh: 1: : Permission denied

But it made the index.html and works like a charm.

[NoQ]

In D74467#1905011, @Charusso wrote:

Could you mention how to use this feature in the Summary please?

cd reports
scan-build --generate-index-only .

Yup, this was supposed to work.

And something is not right, I have tried it:

Use of uninitialized value $Clang in concatenation (.) or string at /llvm-project/clang/tools/scan-build/bin/scan-build line 1895.

Uh-oh. Does this still show up if you specify --use-analyzer explicitly? Because %scan_build in tests includes this flag because otherwise it has no knowledge of the clang build directory.

sh: 1: : Permission denied

This looks like a separate problem, dunno.

[Charusso]

In D74467#1905416, @NoQ wrote:

In D74467#1905011, @Charusso wrote:

Could you mention how to use this feature in the Summary please?
And something is not right, I have tried it:

Use of uninitialized value $Clang in concatenation (.) or string at /llvm-project/clang/tools/scan-build/bin/scan-build line 1895.

Uh-oh. Does this still show up if you specify --use-analyzer explicitly? Because %scan_build in tests includes this flag because otherwise it has no knowledge of the clang build directory.

sh: 1: : Permission denied

This looks like a separate problem, dunno.

Well, then please mention the proper usage is:

cd reports
scan-build --use-analyzer clang --generate-index-only .

With that none of the concatenation and sh permission issue shows up. However it would be a nice QoL if that extra flag is not needed to generate HTML.

[NoQ]

Re-landed as rG0c27fd82e1e6.

I sorted the list of HTML files before deduplicating and the buildbots seem silent for now.

Also fxd the uninitialized variable $Clang use when --use-analyzer= is not specified; here's what it was used for:

clang/test/Analysis/scan-build/rebuild_index/rebuild_index.test

+// FIXME: Actually, "perl".
+REQUIRES: shell
+
+// FIXME: Should ideally work on Windows.
+UNSUPPORTED: system-windows
+
+RUN: rm -rf %t.output_dir && mkdir %t.output_dir
+RUN: cp %S/report-1.html %t.output_dir
+RUN: cp %S/report-2.html %t.output_dir
+RUN: cp %S/report-3.html %t.output_dir
+RUN: mkdir %t.output_dir/subdirectory
+RUN: cp %S/subdirectory/report-4.html %t.output_dir/subdirectory
+
+RUN: %scan-build --generate-index %t.output_dir
+
+RUN: ls %t.output_dir | FileCheck -check-prefix CHECK-FILES %s
+
+CHECK-FILES:      index.html
+CHECK-FILES-NEXT: report-1.html
+CHECK-FILES-NEXT: report-2.html
+
+// report-3.html is a duplicate of report-1.html so it's not present.
+CHECK-FILES-NOT:  report-3.html
+CHECK-FILES-NEXT: scanview.css
+CHECK-FILES-NEXT: sorttable.js
+CHECK-FILES-NEXT: subdirectory
+
+RUN: ls %t.output_dir/subdirectory | FileCheck -check-prefix CHECK-SUB %s
+
+CHECK-SUB: report-4.html
+
+RUN: cat %t.output_dir/index.html | FileCheck -check-prefix CHECK-INDEX %s
+
+CHECK-INDEX:      cat1
+CHECK-INDEX-NEXT: bug1
+CHECK-INDEX-NEXT: cat2
+CHECK-INDEX-NEXT: bug2
+CHECK-INDEX-NEXT: cat4
+CHECK-INDEX-NEXT: bug4
+
+CHECK-INDEX:     report-1.html#EndPath
+CHECK-INDEX:     report-2.html#EndPath
+CHECK-INDEX-NOT: report-3.html#EndPath
+CHECK-INDEX:     subdirectory/report-4.html#EndPath

clang/test/Analysis/scan-build/rebuild_index/report-1.html

+<!-- BUGTYPE bug1 -->
+<!-- BUGFILE file1 -->
+<!-- BUGPATHLENGTH 1 -->
+<!-- BUGLINE 1 -->
+<!-- BUGCATEGORY cat1 -->
+<!-- BUGDESC desc1 -->
+<!-- FUNCTIONNAME func1 -->
+<!-- BUGMETAEND -->

clang/test/Analysis/scan-build/rebuild_index/report-2.html

+<!-- BUGTYPE bug2 -->
+<!-- BUGFILE file2 -->
+<!-- BUGPATHLENGTH 2 -->
+<!-- BUGLINE 2 -->
+<!-- BUGCATEGORY cat2 -->
+<!-- BUGDESC desc2 -->
+<!-- FUNCTIONNAME func2 -->
+<!-- BUGMETAEND -->

clang/test/Analysis/scan-build/rebuild_index/report-3.html

+<!-- BUGTYPE bug1 -->
+<!-- BUGFILE file1 -->
+<!-- BUGPATHLENGTH 1 -->
+<!-- BUGLINE 1 -->
+<!-- BUGCATEGORY cat1 -->
+<!-- BUGDESC desc1 -->
+<!-- FUNCTIONNAME func1 -->
+<!-- BUGMETAEND -->

clang/test/Analysis/scan-build/rebuild_index/subdirectory/report-4.html

+<!-- BUGTYPE bug4 -->
+<!-- BUGFILE file4 -->
+<!-- BUGPATHLENGTH 4 -->
+<!-- BUGLINE 4 -->
+<!-- BUGCATEGORY cat4 -->
+<!-- BUGDESC desc4 -->
+<!-- FUNCTIONNAME func4 -->
+<!-- BUGMETAEND -->

clang/tools/scan-build/bin/scan-build

my %Options = (
   InternalStats => undef,
   OutputFormat => "html",
   ConfigOptions => [],       # Options to pass through to the analyzer's -analyzer-config flag.
   ReportFailures => undef,
   AnalyzerStats => 0,
   MaxLoop => 0,
   PluginsToLoad => [],
   AnalyzerDiscoveryMethod => undef,
-  OverrideCompiler => 0,      # The flag corresponding to the --override-compiler command line option.
-  ForceAnalyzeDebugCode => 0
+  OverrideCompiler => 0,     # The flag corresponding to the --override-compiler command line option.
+  ForceAnalyzeDebugCode => 0,
+  GenerateIndex => 0         # Skip the analysis, only generate index.html.
 );
 lock_keys(%Options);
 
 ##----------------------------------------------------------------------------##
 # Diagnostics
 ##----------------------------------------------------------------------------##
 
 sub Diag {
if ($Num > 0 && -r "$Dir/index.html") {
     Diag("Run 'scan-view $Dir' to examine bug reports.\n");
   }
 
   DiagCrashes($Dir) if (scalar @failures || scalar @attributes_ignored);
 
   return $Num;
 }
 
+sub Finalize {
+  my $BaseDir = shift;
+  my $ExitStatus = shift;
+
+  if (defined $Options{OutputFormat}) {
+    if ($Options{OutputFormat} =~ /plist/ ||
+        $Options{OutputFormat} =~ /sarif/) {
+      Diag "Analysis run complete.\n";
+      Diag "Analysis results (" .
+        ($Options{OutputFormat} =~ /plist/ ? "plist" : "sarif") .
+        " files) deposited in '$Options{OutputDir}'\n";
+    }
+    if ($Options{OutputFormat} =~ /html/) {
+      # Postprocess the HTML directory.
+      my $NumBugs = Postprocess($Options{OutputDir}, $BaseDir,
+                                $Options{AnalyzerStats}, $Options{KeepEmpty});
+
+      if ($Options{ViewResults} and -r "$Options{OutputDir}/index.html") {
+        Diag "Analysis run complete.\n";
+        Diag "Viewing analysis results in '$Options{OutputDir}' using scan-view.\n";
+        my $ScanView = Cwd::realpath("$RealBin/scan-view");
+        if (! -x $ScanView) { $ScanView = "scan-view"; }
+        if (! -x $ScanView) { $ScanView = Cwd::realpath("$RealBin/../../scan-view/bin/scan-view"); }
+        exec $ScanView, "$Options{OutputDir}";
+      }
+
+      if ($Options{ExitStatusFoundBugs}) {
+        exit 1 if ($NumBugs > 0);
+        exit $ExitStatus;
+      }
+    }
+  }
+
+  exit $ExitStatus;
+}
+
 ##----------------------------------------------------------------------------##
 # RunBuildCommand - Run the build command.
 ##----------------------------------------------------------------------------##
 
 sub AddIfNotPresent {
   my $Args = shift;
   my $Arg = shift;
   my $found = 0;
-v
    Enable verbose output from scan-build. A second and third '-v' increases
    verbosity.
 
  -V
  --view
 
    View analysis results in a web browser when the build completes.
 
+ --generate-index <output location>
+
+   Do not perform the analysis, but only regenerate the index.html file
+   from existing report.html files. Useful for making a custom Static Analyzer
+   integration into a build system that isn't otherwise supported by scan-build.
+
 ADVANCED OPTIONS:
 
  -no-failure-reports
 
    Do not create a 'failures' subdirectory that includes analyzer crash reports
    and preprocessed source files.
 
  -stats
while (@$Args) {
 
     if ($arg eq '-analyze-headers') {
       shift @$Args;
       $Options{AnalyzeHeaders} = 1;
       next;
     }
 
     if ($arg eq "-o") {
+      if (defined($Options{OutputDir})) {
+        DieDiag("Only one of '-o' or '--generate-index' can be specified.\n");
+      }
+
       shift @$Args;
 
       if (!@$Args) {
         DieDiag("'-o' option requires a target directory name.\n");
       }
 
       # Construct an absolute path.  Uses the current working directory
       # as a base if the original path was not absolute.
       my $OutDir = shift @$Args;
       mkpath($OutDir) unless (-e $OutDir);  # abs_path wants existing dir
       $Options{OutputDir} = abs_path($OutDir);
 
       next;
     }
 
+    if ($arg eq "--generate-index") {
+      if (defined($Options{OutputDir})) {
+        DieDiag("Only one of '-o' or '--generate-index' can be specified.\n");
+      }
+
+      shift @$Args;
+
+      if (!@$Args) {
+        DieDiag("'--generate-index' option requires a target directory name.\n");
+      }
+
+      # Construct an absolute path.  Uses the current working directory
+      # as a base if the original path was not absolute.
+      my $OutDir = shift @$Args;
+      mkpath($OutDir) unless (-e $OutDir);  # abs_path wants existing dir
+      $Options{OutputDir} = abs_path($OutDir);
+      $Options{GenerateIndex} = 1;
+
+      next;
+    }
+
     if ($arg =~ /^--html-title(=(.+))?$/) {
       shift @$Args;
 
       if (!defined $2 || $2 eq '') {
         if (!@$Args) {
           DieDiag("'--html-title' option requires a string.\n");
         }
 

 
 if (!@ARGV) {
   $ForceDisplayHelp = 1
 }
 
 ProcessArgs(\@ARGV);
 # All arguments are now shifted from @ARGV. The rest is a build command, if any.
 
-if (!@ARGV and !$RequestDisplayHelp) {
-  ErrorDiag("No build command specified.\n\n");
-  $ForceDisplayHelp = 1;
-}
-
 my $ClangNotFoundErrMsg = FindClang();
 
 if ($ForceDisplayHelp || $RequestDisplayHelp) {
   DisplayHelp($ClangNotFoundErrMsg);
   exit $ForceDisplayHelp;
 }
 
+$CmdArgs = HtmlEscape(join(' ', map(ShellEscape($_), @ARGV)));
+
+# Make sure to use "" to handle paths with spaces.
+$ClangVersion = HtmlEscape(`"$Clang" --version`);
+
+if ($Options{GenerateIndex}) {
+  Finalize($Options{OutputDir}, 0);
+}
+
+if (!@ARGV and !$RequestDisplayHelp) {
+  ErrorDiag("No build command specified.\n\n");
+  $ForceDisplayHelp = 1;
+}
+
+# Determine the output directory for the HTML reports.
+my $BaseDir = $Options{OutputDir};
+$Options{OutputDir} = GetHTMLRunDir($Options{OutputDir});
+
 DieDiag($ClangNotFoundErrMsg) if (defined $ClangNotFoundErrMsg);
 
 $ClangCXX = $Clang;
 if ($Clang !~ /\+\+(\.exe)?$/) {
   # If $Clang holds the name of the clang++ executable then we leave
   # $ClangCXX and $Clang equal, otherwise construct the name of the clang++
   # executable from the clang executable name.
 
   # Determine operating system under which this copy of Perl was built.
   my $IsWinBuild = ($^O =~/msys|cygwin|MSWin32/);
   if($IsWinBuild) {
     $ClangCXX =~ s/.exe$/++.exe/;
   }
   else {
     $ClangCXX =~ s/\-\d+\.\d+$//;
     $ClangCXX .= "++";
   }
 }
 
-# Make sure to use "" to handle paths with spaces.
-$ClangVersion = HtmlEscape(`"$Clang" --version`);
-
-# Determine where results go.
-$CmdArgs = HtmlEscape(join(' ', map(ShellEscape($_), @ARGV)));
-
-# Determine the output directory for the HTML reports.
-my $BaseDir = $Options{OutputDir};
-$Options{OutputDir} = GetHTMLRunDir($Options{OutputDir});
-
 # Determine the location of ccc-analyzer.
 my $AbsRealBin = Cwd::realpath($RealBin);
 my $Cmd = "$AbsRealBin/../libexec/ccc-analyzer";
 my $CmdCXX = "$AbsRealBin/../libexec/c++-analyzer";
 
 # Portability: use less strict but portable check -e (file exists) instead of
 # non-portable -x (file is executable). On some windows ports -x just checks
 # file extension to determine if a file is executable (see Perl language
my %EnvVars = (
   'CLANG_ANALYZER_TARGET' => $Options{AnalyzerTarget},
   'CCC_ANALYZER_FORCE_ANALYZE_DEBUG_CODE' => $Options{ForceAnalyzeDebugCode}
 );
 
 # Run the build.
 my $ExitStatus = RunBuildCommand(\@ARGV, $Options{IgnoreErrors}, $Options{KeepCC},
                           $Cmd, $CmdCXX, \%EnvVars);
 
-if (defined $Options{OutputFormat}) {
-  if ($Options{OutputFormat} =~ /plist/ ||
-      $Options{OutputFormat} =~ /sarif/) {
-    Diag "Analysis run complete.\n";
-    Diag "Analysis results (" .
-      ($Options{OutputFormat} =~ /plist/ ? "plist" : "sarif") .
-      " files) deposited in '$Options{OutputDir}'\n";
-  }
-  if ($Options{OutputFormat} =~ /html/) {
-    # Postprocess the HTML directory.
-    my $NumBugs = Postprocess($Options{OutputDir}, $BaseDir,
-                              $Options{AnalyzerStats}, $Options{KeepEmpty});
-
-    if ($Options{ViewResults} and -r "$Options{OutputDir}/index.html") {
-      Diag "Analysis run complete.\n";
-      Diag "Viewing analysis results in '$Options{OutputDir}' using scan-view.\n";
-      my $ScanView = Cwd::realpath("$RealBin/scan-view");
-      if (! -x $ScanView) { $ScanView = "scan-view"; }
-      if (! -x $ScanView) { $ScanView = Cwd::realpath("$RealBin/../../scan-view/bin/scan-view"); }
-      exec $ScanView, "$Options{OutputDir}";
-    }
-
-    if ($Options{ExitStatusFoundBugs}) {
-      exit 1 if ($NumBugs > 0);
-      exit $ExitStatus;
-    }
-  }
-}
-
-exit $ExitStatus;
+Finalize($BaseDir, $ExitStatus);