WILL
----- Original Message ----- From: "specialist33" <[EMAIL PROTECTED]>
To: "Velocity Developers List" <[EMAIL PROTECTED]>
Sent: Tuesday, December 14, 2004 10:59 AM
Subject: Re: Exception
Will Glass-Husain wrote: > Thanks for the email. > > I'll note that backwards compatibility is an important philosophy for > this project - believe it or not there are still JDK 1.3 users out > there. (and it seems arrogant to demand they upgrade merely to use one > library).
I certainly do believe it :), but the *standard* for people who are being sensibly cautious is to stay one release behind Sun: which means people should already be migrating to 1.4.
To be two versions behind is moving towards shooting yourself in the head. Especially when 1.3 is so vastly inferior to 1.4 (for a start, you can't write serious servers without 1.4, which is presumably most of the V audience?)
But, as I said, it's not particularly hard to support both at once, especially in this case, so I don't think it really matters what I think :).
> > But it's an interesting idea to bring in useful JDK 1.4 features in a > separate block of code. Seems a little messy, but it's worth
I was afraid the first time I did something like this, but in practice, it's usually not messy at all, because the JDK-specific code all gets contained in a single class :). It looks neat.
Until it's no longer needed, at which point it is still neat, but also strange - you wonder why someone bothered having an extra class?
> considering. Please submit a bugzilla patch so that it's on record.
I don't do bugzilla unless I can just type in the bug and hit OK. c.f. java.sun.com's bug-logging: it hasn't hurt them! This is especially true of a project that hasn't released an update in so many months. The promised release for last weekend? I've heard that so many times before (and I've even been responsible for broken promises myslef when maintaining oss projects; for which I'm still a bit ashamed and very regretful. But, that's how the world is). If the project is very lively, I might just try it. At the moment, I'm more likely to just give up and go and find another templating engine that is actually production-ready and much more active, or else make a new one of my own that supports NIO and is a lot faster than V. I prefer to use V because I prefer to support and promote oss projects, esepcially when - like V - they are fundamentally rather nicely put together. But...when the going gets too tough, it's not worth the cost, and it's cheaper to jump ship.
PS: I recently heard from one of our users that the manual's documentation on instantiating hashmaps completely fails to work *at all* in the current version (1.4) despite being copied and pasted directly as-is. I looked and it appeared to be so, and I noticed that by the looks of things the parser is broken - the parser error messages are actually contradictory if you fiddle with it a little. That doesn't give people much hope, given how often the unified collection iteration is touted as a good feature of V.
Most people have BZ setup to force new people to jump through hoops; if you used as much software as I did you'd realise why I adopted a policy of not submitting anything to systems that require passwords.
It's pointless, it's annoying. There's no security (it doesn't stop me from using a fake one-time email address e.g. @ mailinator.com) and all it really does is irritate people who don't ahve the time nor really give a [insert something here] about someone else's poxy little club of usernames and passwords. (please note: that is *not* a dig at the V mailing list or bug DB; I'm pointing out the general problem of every person + dog thinking that their website is important enough to warrant password access. Innocent projects like V get the fallout of people being hyper-sensitive to unnecessary and pointless passwords, just like genuine retailers sometimes get mistaken for spam and get a grossly unfair reception).
If it had a box where you could *optionally* type in your email address - no questions asked - to be CC'd on changes to the email, then no problem (and, for those that wanted to, to *optionally* create an accoutn and login). But it doesn't do that.
PPS: if you read this and think "arrogant idiot" then fair enough. However, I'm working flat-out to get a major system rolled out ASAP (and yes, I *will* be working on xmas day, it's that desperate), and I really don't have the time to mess about with basic yet fundamental errors in one of the 3rd party libs. I wish I had more time, really. But I don't. I have ensured we stuck with V even when others on the project were demanding we duimp it because of the problems, because I'd like to promote V in my own small way, but I have to be realistic about where and how much time I can throw away on it :(.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
