| Ticket UUID: | d5ddbc7f49bc0a8820c50da00db83fd4f2629e28 | ||
| Title: | Inconsistent terminology in error messages: should be vs. must be, unknown/ambiguous vs. bad | ||
| Type: | Bug | Created on: | 2014-06-17 21:06:30 |
| Submitter: | andy | Assigned to: | dkf |
| Subsystem: | 45. Parsing and Eval | Severity: | Minor |
| Priority: | 4 | Last modified: | 2014-09-14 11:15:00 |
| Status: | Open | Closed by: | nobody |
| Resolution: | None | Closed on: | |
| Version: | fffc281b07873e1d3f381fe3dce5dec29df9aa47 | ||
| Description: | ||||
% puts wrong # args: should be "puts ?-nonewline? ?channelId? string" % binary asdf unknown or ambiguous subcommand "asdf": must be decode, encode, format, or scan % fileevent asdf fdsa bad event name "fdsa": must be readable or writable Sometimes we get "should be", sometimes "must be". Sometimes it's "unknown or ambiguous", sometimes just "bad". [fileevent] accepts unambiguous prefixes of readable and writable, so why doesn't it say "unknown or ambiguous event name"? | ||||
| User Comments: | ||||
dkf added on 2014-06-23 08:40:26:
Just FYI, “unknown or ambiguous” is a mark of something hand-written; the proper language is “ambiguous” when there are too many possible matches and “bad” when there are too few. dkf added on 2014-07-12 15:59:01:
Will be “should be” as that results in a smaller change to the test suite than “must be”. | ||||