Compared to 2017 and earlier, the data at GDC are changing rapidly these days; this makes the current approach of the GDCtools regression test suite something of a burden, because it needs to be updated after each new GDC data update.
We can keep the spirit of the regression tests, but discontinue the exhaustive MD5/UUID cross-checking (against our regression baseline data) of the names and content of each file downloaded. Instead, we'll continue to download the data, but only check things like:
the integrity of the downloaded (i.e. mirrored) data (against its respective MD5 file)
that the downloaded files are indeed the same in number and type as expected
that the dicing, loadfiles & sample reports can be generated ...
Compared to 2017 and earlier, the data at GDC are changing rapidly these days; this makes the current approach of the GDCtools regression test suite something of a burden, because it needs to be updated after each new GDC data update.
We can keep the spirit of the regression tests, but discontinue the exhaustive MD5/UUID cross-checking (against our regression baseline data) of the names and content of each file downloaded. Instead, we'll continue to download the data, but only check things like: