Upgrading Perl... Somebody just shoot me and put me out of my misery!

[Ronald F. Guilmette]

In message <845A2E7E540E58EFD7F05B6B at atuin.in.mat.cc>, 
Mathieu Arnold <mat at FreeBSD.org> wrote:

>rfg wrote:
>| Why would _anything_ that is in any way dependent upon the Perl
>| interpreter need to be rebuilt?  In this switch to threads=on, has the
>| language itself changed?  And if not, shouldn't the change to
>| multi-threading capability within the interpreter be utterly transparent
>| to (and a non-event for) any and all pre-existing Perl code?
>| 
>| Obviously, there's something that I'm missing, but I have no idea what it
>| might be.
>
>Because, hum, quite a few things change when you enable threads, some
>headers bits change, some calls that are noop without become real call
>with, things like that.
>
>Now, it obviously is a non issue with ports that only use perl to run
>scripts, or p5- ports that are only scripts, but for ports that have XS
>files that get compiled into .so, they need to get recompiled, and the same
>goes for every bit of software that includes the interpreter.
>
>As there is no simple way to differentiate between those two categories of
>dependencies, I ask people to rebuild (or reinstall, if you're using binary
>packages) everything.

OK.  It is all clear now.  Thank you for taking the time to explain.  It
_does_ all make sense now.

>I assure you, it does not make me happy at all to have people rebuild
>everything depending on Perl every two weeks (like it feels I've been doing
>that for a few months...)

Well, I apologize if I cam off as being a bit... um... testy.  I'm actually
one of the lucky people, I guess, since I only update my ports very
infrequently... only once in every several months... so I've managed to
miss most of the excitement. :-)

(As I mentioned in my original post in this thread, it was late and I was
tired when I first posted about all this.  Please do forgive me if I seemed
at all unappreciative of your hard work, which is clearly of great value,
both to me personally and also to countless others.)

>| I'm *not* claiming that the maintainer didn't have a good reason for
>| suggesting these rebuilds.  I'm only saying that *I* personally still
>| don't have a good understanding of what the need for this is/was.
>
>As the maintainer, I hope my previous bit did explain that a bit better, if
>things are not that clear, do feel free to point them out and I'll try
>better.

No no.  You have now explained the resons for the rebuilds clearly and
admirably.  It all makes sense.

>The thing is that all those explanations can't go into UPDATING, we try to
>keep it short not to confuse people.

Yes.

I'm glad that we have these mailing lists, and their associated archives,
so that people like me with an interest in such arcana can ask and get
answers... at least from the subset of port maintainers who, like you,
are nicely responsive.

Keep up the good work!  And thanks again, for everthing.


Regards,
rfg