lmi
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[lmi] wx_test_expiry_dates.cpp


From: Greg Chicares
Subject: [lmi] wx_test_expiry_dates.cpp
Date: Sat, 06 Dec 2014 14:53:55 +0000
User-agent: Mozilla/5.0 (Windows NT 5.1; rv:24.0) Gecko/20100101 Thunderbird/24.6.0

http://svn.savannah.nongnu.org/viewvc/lmi/trunk/wx_test_expiry_dates.cpp?root=lmi&r1=6040&r2=6047

/// Validate dates in the 'expiry' file.
///
/// Write the begin and end dates to stdout, as JDN and as YYYYMMDD,
/// all on a single line, e.g.:
///   begin: 2457024 2015-01-01  end: 2457055 2015-02-01
/// (Those are the dates that would normally be expected for a
/// distribution prepared in December 2014.)
///
/// Validate dates only when the '--distribution' option is given.
///  - The begin date should be the first day of the next month.
///  - The end date should be the first day of the month after next.
/// Binaries are normally distributed toward the end of a month for
/// use throughout the following month. Consequently, these tests will
/// pass if run when a regular distribution is prepared, but they will
/// fail if rerun on the begin date--and that's exactly as desired:
/// such "failure" is not an error. (Incidentally, this illuminates
/// our decision to write all GUI-test output to stdout, not stderr.)
///
/// (When '--distribution' is not given, we had considered writing a
/// line to stdout saying that the tests had been skipped. But that is
/// not necessary, because in our normal workflow we'll compare each
/// '--distribution' run to a previously-saved '--distribution' run,
/// and diffs will make it obvious that the tests have been skipped.)
///
/// Occasionally "interim" distributions are issued, e.g., to add an
/// urgently-needed feature or to fix a critical mistake. They are
/// to be tested in the same way as regular distributions. All regular
/// distributions resemble each other; each "interim" distribution is
/// irregular in its own way, and its validation "failures" are not
/// errors, but may indeed convey useful information.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]