This way the writing of the JSON file is handled automatically if the update does not fail.
It pave the way to further global improvements, such as a better error handling.
It may not be the best place for that (gha.py would have been better), but it's the shorter / faster way to do it for now.
Moreover it now uses logging for writing the group. The logger format has been updated for this to work. This was done to fix issues on GitHub Action logs, where groups were declared after the logs.
Make the script more readable, mostly by:
- using the endoflife.Product class,
- removing the use of functions when unnecessary,
- a little bit of renaming and documentation.
Create common functions parse_date, parse_month_year_date and parse_datetime.
Those functions support trying multiple formats, and come with default formats lists that support most of the date format encountered so far.
Notable change: year-month dates are now set to the end of month (impacted couchbase-server and ibm-aix).
Move the parallel URL fetching from firefox.py to endoflife.py to make available parallel URL fetching for all scripts.
Also a a fix found on https://stackoverflow.com/a/44511691/374236 to avoid ChunkedEncodingError.
Create a common function to write resulting JSON files to the releases directory.
It makes this task simpler to read and maintain, while making it modifiable at a central point in the future.
One example of such modification could be the sorting of the versions in a uniform way for all the scripts.
This creates a common function to fetch HTTP URLs, with enhanced capabilities (retry, use of a known User-Agent).
It makes scripts that need those capabilities simpler, while improving other scripts.
This commit also fixes some scripts that did not log properly (cos.py, eks.py, haproxy.py, palo-alto-networks.py, rhel.py, ros.py, unrealircd.py).
