For specific regex functions, you might want to take a look at http://www.regexr.com/ It has a live example to help code on the fly.

Thanks rknichols

I had completely missed that the "." was still in the command. After removing it grep is returning what it did before.

Also I want "only" the words that have the antiquated word endings "est" or "eth" as in becamest or accepteth, so adding the \w at the end gets me words I don't want and so I removed it from the example I started from.

Thanks joec@home I will take a good look at the regexr site. Looks to be very useful.

So there is not a grep bug after all. And what's more the command as originally posted works as it should as near as I can tell. Of coarse I do run Debian Testing and I don't always restart after updates so maybe grep was using mismatched libraries or something.

With more advanced commands with regex and so on, I tend to experiment switching between grep and egrep to see which works better in what situation.

Then, you might want "\b" following the "est" to require the word to end at that point and not match partial words that have "est" somewhere in the middle. Or, perhaps you aren't seeing enough of those false matches to matter.