geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From David Jencks <>
Subject Re: Trunk or Branch?
Date Tue, 06 Jun 2006 05:40:12 GMT

On Jun 5, 2006, at 9:29 PM, Neal Sanche wrote:

> David Jencks wrote:
>> On Jun 5, 2006, at 9:10 PM, Neal Sanche wrote:
>>> Hey guys,
>>> Could you be so kind as to let me know what I should be building  
>>> today? I currently have branches\1.1 and am building that. If I  
>>> want a 1.1 server to play with, do I build from there or from  
>>> trunk\ ?
>> you have 1.1[-SNAPSHOT].  Trunk is 1.2-SNAPSHOT
>> thanks and good luck :-) with all the outages
> Thanks David,
> Yeah, I see a distinct lack of snappiness in the build process this  
> evening. Many long, drawn out, pauses punctuated by shorter pauses.  
> These outages, are they easily explainable? :)

I guess it depends on what you mean by "easily" and "explainable".  I  
think that all the repos we depend on are underpowered and  
overwhelmed.  I was actually referring more to the recent sequence of  
hardware failures at all the open source repos I know of -  
sourceforge, apache, and codehaus.  However I'm afraid that the  
current maven model of trying to download artifacts from a list of  
single remote http servers during a build has reached its limit and  
is proving to be essentially broken.  There are a lot of obvious  
changes most of which would make it considerably more reliable:

-- asynchronous downloads, a process sits there trying to check for  
newer snapshots all day and night, so when you actually try to build  
you don't look at remote repos
-- bittorrent wagon, so all people who have an artifact participate  
in making it available
-- parallel checking of all repos
-- someone paying for a reliable geronimo-only repository.

However I'm not in a position to implement any of these.  Hopefully  
m2 will be better, maybe they already have (1) with the proxying  

david jencks

> -Neal

View raw message