sheesh.

[sev]

fucking spamassassin and their fucking memory-hog code.

SCALAR EACH IS YOUR FRIEND, DIPSHIT.

sheesh. I did *not* need to be optimizing code written by people I've never met. (EDIT: though technically, in my mind, addressing 'out of memory' errors that crop up under normal usage isn't optimizing anymore; it's debugginging.)

Comments

[nmc.livejournal.com]

But I thought if code was reasonably written, it wasn't good Perl?

You wouldn't want the SpamAssassin people to write bad Perl, now, would you?

*innocent*

[sev]

*snarl*

ask me again after I've spent two weeks upgrading software so I can get my installation running a new enough version that the SA folks will speak to me.

[nmc.livejournal.com]

I will not laugh at my friend's misfortune.
I will not laugh at my friend's misfortune.
I will not laugh at my friend's misfortune.

*giggle*

Well, at least I didn't laugh.

it gets worse...

[sev]

...see the next entry, about me trying to upgrade the damn thing.

Do you know what that led to?

I spent today helping Z get an NFS server set up on his windows server.

I don't wanna talk about it.

(*thwwwwwwpt!*)

Re: it gets worse...

[xlerb.livejournal.com]

I don't want to think about that.

Re: it gets worse...

[sev]

It was easier than I thought it'd be.

Still took far too long, but, well, it's NFS, and NFS is scary all by itself!

Re: it gets worse...

[cjcollier.livejournal.com]

Eew. Is that even possible?

[xlerb.livejournal.com]

There's More Than One Way To Do It.

Although it would be nice if Perl could recognize that, in for $x (keys %y) { foobar $x }, the array value of that keys %y is never actually used as such and so doesn't need to be constructed. I can understand if static analysis would be too much, but surely there's already enough runtime magic strewn about that this kind of thing can be tossed in. (For example, having while ($x = each %y) { foobar $x } not halt when the key is "0" or "".)

On the other hand, at what point does Perl stop being a scripting language and start being a DB query planner?

[sev]

And if we can do it for
for $x (keys %y) { ... }

then we should be able to do it for
for $x (grep(/pattern/,keys %y)) { ... }

and, and, and!

(people just like me are why perl is the swiss army chainsaw.)

[xlerb.livejournal.com]

grep (and map) would make it a little trickier; it'd have to make sure the predicate/block is free of side-effects so that the lazy-evaluation thing can be done without possibly changing the code's behavior. I think.

(Bigger chainsaw!!!)

[cjcollier.livejournal.com]

while( defined $x = each %y ) foobar $x doesn't halt on "0" or ""

In a lot of cases, you want to halt on false-evaluating values, and it's relatively simple to add a C<defined> operator.

[cjcollier.livejournal.com]

did someone do a foreach(){ ... } ?

I hate that one, and it seems that *everyone* and their brother does it.