HTML report and LCOV report discrepancy

Hi I am having a strange issue when generating code coverage with karma-coverage. My source files are written in typescript where as the test files are written in javascript with qunit being the test framework. When I run the karma start, all my tests are getting passed and the code coverage report being generated. But the istanbul report and the lcov.info which I am using sonarqube to view the code coverage are showing different reports.

For example, in html report by istanbul even though the source file is in ts, the code coverage shown in transpiled js file with all of the code is covered except the __awaiter and __generator functions which will be generated when we use Promises in typescript, since we are using the target as **es-5**.

The Istanbul report is showing as 65% lines coverage, but in Sonar the coverage is showing as 17%, please note that with the below coverageReporter configuration, it generates the lcov.info report which I am using it to publish it to sonar.

The Sonar dashboard is displaying the ts file as-is not in the transpiled/javascript code.

Here is the karma-config.

```
preprocessors: {
'src/js/**/*.ts': ['typescript'],
'src/js/**/*.scss': ['scss'],
},

typescriptPreprocessor: {
// options passed to the typescript compiler
options: {
sourceMap: false, // (optional) Generates corresponding .map file.
target: 'ES5', // (optional) Specify ECMAScript target version: 'ES3' (default), or 'ES5'
module: 'amd', // (optional) Specify module code generation: 'commonjs' or 'amd'
noImplicitAny: true, // (optional) Warn on expressions and declarations with an implied 'any' type.
noResolve: true, // (optional) Skip resolution and preprocessing.
removeComments: true, // (optional) Do not emit comments to output.
concatenateOutput: false, // (optional) Concatenate and emit output to single file.
// By default true if module option is omited, otherwise false.
},

// transforming the filenames
transformPath: function(_path) {
return _path.replace(/\.ts$/, '.js');
},
},

scssPreprocessor: {
options: {
sourceMap: true,
},
},

reporters: ['progress', 'coverage'],

coverageReporter: {
dir: 'reports/unit-tests/coverage',
subdir: 'report',
instrumenterOptions: {
istanbul: { noCompact: true },
},
type: 'lcov',
watermarks: {
statements: [50, 90],
branches: [1, 2],
functions: [1, 2],
lines: [1, 2],
},
check: {
global: {
statements: 60,
lines: 65,
functions: 60,
branches: 45,
},
},
},
```

sonar-project.properties

```
sonar.host.url=http://localhost:9000
sonar.projectKey=hello
sonar.projectName=hello
sonar.projectVersion=0.0.1

sonar.modules=hola

hola.sonar.projectBaseDir=hola/
hola.sonar.sources=src/
hola.sonar.exclusions=src/js/**/@types/*,src/js/**/**/resources/**/*
hola.sonar.tests=tests/
hola.sonar.language=ts
hola.sonar.typescript.lcov.reportPaths=reports/unit-tests/lcov.info
```

Start by reproducing the discrepancy using the shown Karma coverage configuration and sonar-project.properties. Compare the generated Istanbul HTML and lcov.info inputs with the TypeScript files Sonar displays, and check how the disabled source maps affect file attribution. Done means the reports agree or the supported configuration and limitation are documented.

Written by the indexing model from the issue text.