Bug References

Bug references are usually a three letter abreviation (TLA) for a tracker instance (e.g. bugzilla.suse.com -> bsc) followed by a # and then an id number, so for example bsc#12345.

Test failures can be 'tagged' with a bug reference. This usually looks something like test01:bsc#12345. There are also anti-tags which signal that a failure should no longer be associated with a given bug. These look like test01:!bsc#12345.

This module provides methods for processing bug references and tags which are typically included in a comment on a test failure, but can be taken from any text.

Parse some text for Bug tags and add them to the given tags index

A reference to a bug on a particular tracker

Parser for the Bug references (tags) DSL

See parse_comment's docs for the format.

parse_comment(text)

Try to extract the test-name:bug-ref pairs from a comment

Below is the approximate syntax in EBNF. Assume letter ∈ [a-Z] and digit ∈ [0-9] and whitespace is allowed between testnames, bugrefs, ':' and ','. However there should be no gap between ':' and '!'.

testname = letter | digit { letter | digit | '_' | '-' }
tracker = letter { letter }
id = letter | digit { letter | digit }
bugref = tracker '#' id
tagging = testname {',' testname} ':' ['!'] bugref {',' bugref}
taggings = tagging { tagging }

A tagging can assign many bug references to many testnames, which means you can have something like: test1, test2: bsc#1234, git#a33f4. Which tags tests 1 and 2 with both bug references.

Comments many contain many taggings along with other random text. If the algorithm finds an error it discards the current tagging and starts trying to parse a new tagging from the point where it failed.

tokval(tok)

Get the textual value of any AbstractToken.

Implement using SubString and not [n:m] because the former is zero-copy.

Some kind of symbol or even an expression; so long as it can be represented by SubString