davidflanagan/covermonkey

Code coverage analysis for SpiderMonkey

covermonkey's Issues

Line erroniously reported as partially covered while running dom.js tests

In the dom.js wrappers, the CustomEvent constructor has a conditional which CoverMonkey reports as partially covered, even though the offending bytecode "stop" is immediately after an unconditional "return" bytecode and should be regarded as unimportant.

domcore.js:

return (o === undefined) ? undefined : EventInit(o);
Function @  PC  #   Instruction
src/domcore.js:110  0   2   getarg 0
src/domcore.js:110  3   2   name "undefined"
src/domcore.js:110  6   2   stricteq
src/domcore.js:110  7   2   ifeq 16 (+9)
src/domcore.js:110  10  1   name "undefined"
src/domcore.js:110  13  1   goto 25 (+12)
src/domcore.js:110  16  1   callname "EventInit"
src/domcore.js:110  19  1   getarg 0
src/domcore.js:110  22  1   call 1
src/domcore.js:110  25  2   return
src/domcore.js:110  26  0   stop

Sort -c output by percentage covered

It would be nice to see the files sorted by amount of coverage. It might also be nice to leave off any files with 100% coverage (perhaps mentioning that the files were covered outside the table but not giving details)

sanitize <,>, and & in HTML output

HTML chars in the JS source code can mess up the browser output. I had "" in my code and CoverMonkey couldn't display anything past that point.

-q for quiet should be quiet

Small patch which does not pass through output from parsed scripts if the -q quiet option is on.

diff --git a/CoverMonkey b/CoverMonkey
index 998735c..3ddaf16 100755
--- a/CoverMonkey
+++ b/CoverMonkey
@@ -704,7 +704,7 @@ function parseScripts(stream, callback) {
                 inscript = true;
                 scriptlines = [ dataline ];
             }
-            else if (stream === process.stdin) {
+            else if (stream === process.stdin && !options.quiet) {
                 console.log(dataline);
             }
         }

Implement overall coverage percentage

The individual file coverage percentages are great, but it would be extremely useful to know the overall percentage of lines that are covered so there is one number that can easily be observed to be going up and down.