cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <stef...@apache.org>
Subject Re: [GUMP] Build Failure - cocoon-block-web3
Date Fri, 04 Apr 2003 19:37:45 GMT
Sam Ruby wrote:
> Stefano Mazzocchi wrote:
> 
>>>
>>> compile-scratchpad:
>>>     [javac] Compiling 39 source files to 
>>> /home/rubys/jakarta/cocoon-2.1/build/cocoon-20030404/scratchpad/dest
>>>     [javac] 
>>> /home/rubys/jakarta/cocoon-2.1/build/cocoon-20030404/scratchpad/src/org/apache/cocoon/transformation/JexlTransformer.java:75:

>>> package org.apache.commons.jexl does not exist
>>>     [javac] import org.apache.commons.jexl.Expression;
>>
>>
>> I'm starting to hate gump!
>>
>> GRRRRRR.
>>
>> Stefano.
> 
> 
> Care to provide a bit more details about what you don't care for?

Oh, no, the above was meant to be a joke. i forgot to add a smiley 
somewhere :)

> I certainly would have eased the introduction of Gump to cocoon a bit 
> more gently than the way that you dove in, but I'm confident that your 
> way will work out too.

yes, "gently" is not my kind of word recently, but since I'm the one 
cleaning up my own mess, I don't mind that much.

I love gump. I love it so much that I'm spending my time to make sure it 
works. and it works reliably.

what is driving me nuts is the 24-hours try/fail cycle, but now that you 
and Steven are helping me out on cocoondev.org, things will be much less 
frustrating.

> Summary of what I've seen so far:
> 
> 1) Gump has already found a potential future backwards compatibility 
> problem that was introduced into Avalon but never made it to release.

yep

> 2) Gump is getting cocoon to think more seriously about what is core and 
> what is extensions.  For my tastes, the current core of cocoon still has 
> too many dependencies - I tend to prefer a minimal core myself.  But in 
> any case, this is not for Gump to decide as to which dependencies are OK 
> and which ones are not.

believe me, we'll be getting there. the proposal about 'modules' and 
'blocks' will make the cocoon look more like a microkernel.

> If you would like a suggestion: I would remove all nags from builds that 
> you don't expect to be succeeding yet on a nightly basis, and then only 
> add them in when a failure would be a surprise.  Then simply check the 
> builds every few days, and if you can bring another block or two online 
> per week then that is fine.

that's a good suggestion.

Even better, I'll have it nag only me as long as I'm not confident about 
them, how does that sound?

-- 
Stefano.



Mime
View raw message