|
in reply to 5.10 imminent?
BrowserUk,
Six months ago, a regular commiter to p5p suggested Easter was a likely date. In a conversation I had with demerphq and avar the other day, it seems like there is no big rush to finish developing. I tried to explain my perspective:
For those using and contributing to bleed, it probably doesn't seem like a big deal. There are folks though that are in a position that their shop won't adopt a new version of perl until it has been out a year or more. These are the folks that are suffering. It doesn't matter how simple a new feature is or how little risk it has, if you don't draw a line in the sand and say "that's it", you will never stop developing.
I want it yesterday. I explained that this doesn't mean I am interested in rushing a product that isn't ready. It means that I am willing to miss out on whatever great ideas developers have until the next release and would like a code freeze for the purposes of producing a stable and well tested release now.
|
Re^2: 5.10 imminent?
by grinder (Bishop) on Apr 12, 2007 at 14:54 UTC
|
I [...] would like a code freeze for the purposes of producing a stable and well tested release now
So does rgs, the current pumpking. There was a mammoth thread on p5p last month discussing some brokenness in pack, unpack and Unicode. At the end of it (if I got it straight) Rafael said that he didn't want to hold up 5.10 by fixing it now, because it would be too destabilising. 5.10 is in a freeze.
He followed with a post asking for what open issues remained with blead.
At the risk of repeating myself, if people want to see 5.10 released any time soon, the easiest way to help is to install a copy of blead (not 5.9.4, but rsync'ed from the public ActiveState repository) and draw up a list of every single module installed on your systems (perllocal will help with that). Then install CPAN::Reporter or CPANPLUS and start smoke testing them. If it fails on blead, but passes on 5.8.8, perlbug it.
Get those errors out into the open. The more XS-ive, the more Unicodish, the more magical the module, the better. Acme modules push the parser and runtime in strange and peculiar ways. So do Damian Conway's. Smoke 'em all. And when all the lights on the control panel are green, we'll be ready for launch.
Back to the original question: no, 5.9.4 was not a release candidate, but 5.9.5 will be. The plan, assuming sufficient energy and tuits from the volunteers who develop the core, is for 5.9.5 to be "it", with a few Release Candidates following on to smooth down any remaining rough edges. And then we have 5.10.0.
• another intruder with the mooring in the heart of the Perl
| [reply] |
|
|
Then install CPAN::Reporter or CPANPLUS and start smoke testing
As of two days ago, CPANPLUS is included in blead.
| [reply] |
|
|
(not 5.9.4, but rsync'ed ...
Oh well...
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
| [reply] |
|
|
| [reply] |
|
|
|
|
|
Re^2: 5.10 imminent?
by BrowserUk (Patriarch) on Apr 12, 2007 at 13:36 UTC
|
Thanks Limbic~Region. That's the kind of stuff I was looking to hear. I had thought I'd heard that 5.9.4 was a release candidate, but it seems not.
I've been playing with some stuff, on and off for a couple of years now, that really needs some of the new features in 5.10 to be viable. I just encountered something on the web that brought it to the top of the stack and the thought again crossed my mind.
People have been talking about some of the new features for a very long time now, the earliest reference I found to defined-or was 49 months ago, but it was a half-hearted search and I'd be very surprised if it didn't go back a year or two before that. Seems almost as if they are waiting for something?
Still, I guess the onion will be cooked (5 minutes to fry; 15 if you want to caramelise without burning; 20 to boil; 25 to roast;) when it's cooked.
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
| [reply] |
|
|
| [reply] |
|
|
Thanks for addressing my OP. The whole smoke testing thing is really a side issue. I was not trying to hurry the process of releasing 5.10, just enquiring as to when it was likely to be.
Patches are great for those who build their own binaries. From my best guess, that is about 10-15% of perl's installed base.
But getting back to my OP, defined-or was just an example, possibly the oldest, of the new features that I would like to use. In particular, the new mechanism that allows the addition of new, scoped keywords is of particular interest. Specifically, I would like to continue my experiments, starting with the sources of a new release that incorporates those new features, and all the (I believe fairly extensive), source code changes that underlie them. The critical part for my purposes is starting from a known, fixed, meaningful point. A new release would fit the bill nicely.
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
| [reply] |
Re^2: 5.10 imminent?
by demerphq (Chancellor) on Apr 12, 2007 at 21:30 UTC
|
I dont think its a matter of "no big rush". I think we would all like to see Perl 5.10 released soon. However releasing is a big responsibility, and nobody wants to see Perl 5.10 go out the door followed by huge panic and a new release soon after. In addition we dont want to put it out the door half baked. Given probable release dates for a 5.12 certain things we do in Perl 5.10 have to be close enough to right to last for some time (possibly forever). (In this area im speaking of the regex engine interface).
Also do remember that we are volunteers, nobody is getting paid for hacking on Perl right now to the best of my knowledge, so any rush that is going to happen will have to coexist with the rest of our lives.
So rgs will release it when its in a state that he is comfortable with and that suits him. And I dont blame him, its a huge responsibility. :-)
---
$world=~s/war/peace/g
| [reply] |
Re^2: 5.10 imminent?
by Herkum (Parson) on Apr 12, 2007 at 13:39 UTC
|
It is a shame that having a perfect product leads to Opportunity Costs from not having a product at all.
I understand the desire to get right, there needs to be a balancing need of getting it done. I cannot force anyone to do anything, nor would I try.
I ask "Is it better to have a perfect product that no one uses, or a product that is flawed but others are able to enjoy?"
| [reply] |
|
|
