This release represents a huge amount of refactoring and under-the-hood improvements. Most users should not experience any regressions, although there are lots of smaller incompatible changes.

Highlights:

:star: Gcovr now runs perfectly on Python 3 (Python 3.5+ recommended). :star: Filters are no longer restricted on Windows. The filter syntax was changed to use forward slashes everywhere, so a filter like -f 'src/.*\.cpp' will now behave identically on all platforms.

You can pip install gcovr the most recent release from PyPI.

Breaking changes:

• This release drops support for Python 2.6. (#250)
• PIP is the only supported installation method.
• No longer encoding-agnostic under Python 2.7. If your source files do not use the system encoding (probably UTF-8), you will have to specify a --source-encoding. (#148, #156, #256)
• Filters now use forward slashes as path separators, even on Windows. (#191, #257)
• Filters are no longer normalized into pseudo-paths. This could change the interpretation of filters in some edge cases.

Improvements and new features:

• Improved --help output. (#236)
• Parse the GCC 8 gcov format. (#226, #228)
• New --source-encoding option, which fixes decoding under Python 3. (#256)
• New --gcov-ignore-parse-errors flag. By default, gcovr will now abort upon parse errors. (#228)
• Detect the error when gcov cannot create its output files (#243, #244)
• Add -j flag to run gcov processes in parallel. (#3, #36, #239)
• The --html-details flag now implies --html. (#93, #211)
• The --html output can now be used without an --output filename (#223)
• The docs are now managed with Sphinx. (#235, #248, #249, #252, #253)
• New --html-title option to change the title of the HTML report. (#261, #263)
• New options --html-medium-threshold and --html-high-threshold to customize the color legend. (#261, #264)

Internal changes:

• Huge refactoring. (#214, #215, #221 #225, #228, #237, #246)
• Various testing improvements. (#213, #214, #216, #217, #218, #222, #223, #224, #227, #240, #241, #245)
• HTML reports are now rendered with Jinja2 templates. (#234)
• New contributing guide. (#253)