Time to upgrade to 5.10.0?

[redhotpenguin]

Greetings Monks,

5.10.0 has been out for a while now. I am an early adopter with some software, such as mod_perl and apache, but I am a laggard with others, such as PostgreSQL and the Perl core. 5.10 has a lot of cool new stuff, and I'd like to get into it but the investment is fairly significant for me.

So being Lazy, I'd like to ask the Monks here if they have upgraded to 5.10.0 yet. I'm still running 5.8.8, but would like to start using 5.10 and get it onto my production systems, but am wondering if I should wait for 5.10.1. Tell me your success stories of upgrading to 5.10.0, or your stories about trying to do that but there was something you needed to wait for 5.10.1 to push it to production.

UPDATE - one of the reasons that this has come to mind now for me is the recent release of mod_perl 2.0.4, the 'works with 5.10' release.

[elmex]

I and the software company I work for completely switched to perl 5.10. While there were some problems with some modules from CPAN, which have been resolved mostly, it has been a stable experience. There are still some bugs though, so if you really need the stablest of the stablest perl, wait a bit.

However, I am very happy with 5.10.

[telemachus]

I used the release to learn more about building my own Perl. It's pretty easy to compile your own 5.10.0 in a distinct directory (/usr/local or /opt or even $HOME/.perltests) and then make a symlink to the new Perl as /usr/bin/perl-5.10.0 (and don't forget perdoc-5.10.0).

This way, the base of my system can still count on /usr/bin/perl being the tried and true, while I can begin testing scripts on 5.10.0. Download the source (say at CPAN) and check out the INSTALL file. It's as easy as ./Compile -des -Dprefix=/path/to/mylocalperl

[grinder]

Dave Mitchell asked people to make a list of 5.10 issues that need to be fixed in 5.10.1.

Surprisingly, most of the issues are quite minor. Perhaps the worst is the @_ slowdown (which has already been fixed). Which means that 5.10.0 is looking pretty good at the moment.

The most important thing that will be fixed in 5.10.1 is to nail down some of the edge cases in given/when and smart matching that have come up over the past few months. My advise would be to not go crazy with this part of the language at the moment, in case 5.10.1 changes behaviour that you have begun to rely upon.

The other issue to keep in mind is that the porters are working on getting 5.8.9 out the door first, and that should happen in the next couple of months, as the codebase is more or less in a code freeze for that. 5.10.1 will not get seriously underway until 5.8.9 is out the door.

(That's all speculation on my part, not the summary of any official announcement).

[ambrus]

The innovations of 5.10 has ruined three of my obfus so far (Japh algebra revisited because reading $* now gives a warning; Fun with duff's device and AUTOLOAD but that doesn't matter because it didn't really work in the first place; and One-liner japh with a completely unnecessary incompatibility). If they change smart match semantics and ruin JAPH to commemorate the release of Perl 5.10.0 too, I'll be very angry. (In reality, I think perl5's smart match is unusable, which is why I wrote that parody. Why couldn't they just copy a sane smart match semantics from ruby or perl6 or something?)

[hipowls]

According to this page 5.10.0 is a testing release. It probably means exactly that, this is the release to test your software on in preparation for the 5.10.1 maintenance release.

If you find any bugs or performance problems in perl then report them to perlbug.

[jethro]

I upgraded to 5.10 because of a bug (or bugs) in 5.8.8 in relation to utf8 and Parse::RecDescent. The upgrade solved my problems

I didn't try out any of the new things although they look like my dreams come true (except for the smart matching which is probably too error prone in my bumbling hands). But will do soon