Control: severity -1 serious Re: Thorsten Glaser > >Why isn't Etc/UTC an alternative to UTC? > > It’s not as portable, it binds to the Olson database whereas > an otherwise unqualified UTC is pretty standard. Worse, if > Etc/UTC is not available, the fallback makes it assume Etc, > not UTC, as timezone name.
We can't know how many people out there have "TZ=UTC" in their shell scripts, and we shouldn't break that. Perhaps Etc/UTC is more correct, but I don't see why insisting on correctness needs to annoy users, when the cost is just maintaining one symlink that we have shipped for decades. For a concrete data point: This change is now causing PostgreSQL to FTBFS since UTC is used in the regression tests: SET TimeZone to 'UTC'; +ERROR: invalid value for parameter "TimeZone": "UTC" I'm hence raising to serious. Christoph