STRANGE BEHAVIOR

[Mezzenilium]

PATTERN : ".*bb.star" matches with "aabbccdd" (I had to replace * by star to have the message accepted)
but
PATTERN : ".*bb" does not match with "aabbccdd"

It would be fine if ada-regex could at least avoid imposing to end any pattern by ".*" , even if this can be done automatically in the user's code.

Is it something imposed by the code in Compiler Dragon book ?

[skordal]

This is the expected behaviour for the Regex.Matchers.Matches functions. They check for a match between a regular expression and the (complete) input string.

In your example, .*bb.* means "match all strings that has bb somewhere inside them". In your second example, .*bb means "match all strings that end with bb.

[Mezzenilium]

Thank you for your reply. I understand the pedagogical reasoning behind this approach.

I would like to draw your attention to this piece of Regex code of only 35 lines , originally proposed in C by Rob Pike:

https://benhoyt.com/writings/rob-pike-regex

In this article, Ben Hoyt proposes an elegant version of this algorithm in Go, based on string indexing and string slicing. This is exactly what the ability to specify a lower bound for unconstrained arrays, provided by the GNAT extension of ADA 2022 (-gnatX compilation option), allows now in Ada.

https://blog.adacore.com/going-beyond-ada-2022