I'd say it differently: If you intend to use the features of 5.6.x, then please take 5.8.0 instead. It has all the fixes of 5.6.1 and lots more. It has unicode support which doesn't suck, no, even rocks.

If you ever find yourself being bitten by utf-8 related problems in 5.6.x, then don't go searching for workarounds. Install the real thing. 5.8.0

If you think you need to use threads, then don't take 5.005 or 5.6. But also don't bet the farm on the new ithreads in 5.8.0, because interesting issues keep coming up with them.

Do experiment with them though. The lack of people who started using threads when 5.8.0 was being developed is one of the reasons why they're not stable yet.

bart@smop.org

Hmmm.... I'd actually disagree with that advice... in a hopefully constructive manner :-)

Note: I am not saying "don't use 5.8". I'm happily using 5.8 on my laptop at the moment! It''s fluffy and nice. It's been the most pain-free upgrade of perl ever for me. I have it running on a couple of dev boxes and use it daily.

However....

5.8 will also have bugs (maybe less than 5.6.1 :-). At the moment, most of them are unknown and undocumented.

While perl 5.6.1 has bugs, many (possibly even most) are known and documented.

When it comes to a production server that people depend on I will pick more-known-bugs over less-unknown-bugs any day of the week --- unless the known-bug in question is a significant barrier to developing/maintaining/running the server in question.

Hopefully that makes some vague sort of sense.

I would agree, unless you need better thread support, or some of the new IO features or Unicode support (btw that does really rock) stay with 5.6.1. For the most part in my book, known bugs with known workarounds are much better suited in a production environment than unknown bugs and a few new goodies. =) It may be worth your while to consider waiting for the 5.8 jump until 5.8.1 is out -- Although I would most defiantly code with 5.8 in mind so the upgrade is less freighting when you do it.

-Waswas