ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Steve Loughran <>
Subject Re: Retry task container
Date Mon, 14 May 2007 14:54:26 GMT
Dominique Devienne wrote:
> On 5/12/07, Kevin Jackson <> wrote:
>> <retry noRetries="3">
> noRetries as in "Number Of Retries" or "No retries"???
> How about retrycount instead? --DD
> PS: I second all comments made so far. <retry> could live in
> Ant-Contrib too, rather than Ant.
> PPS: If <retry> had additionally a condition to satisfy before it
> executed its content, with Steve's internaltime attribute, it would
> have a lot of parallels with <waitfor>, no?

ahh, you are into advanced workflows there. I am thinking of adding for 
smartforg a during component that deploys its children whenever a test 
is in a specific state, undeploys them when the state is not met, and 
redeploys them when the condition is met again. you could do the same 
with some while loops, and I will probably write the component through 

during extends During {

  condition extends NetworkVisible {
    host GmailJabberClient:host;

   interval 1000;
   delay 5000;  //hysteresis effects

  action extends GmailJabberClient {
   user "host_account_1";
   password "secret";


This would let me deploy things that go away when the external stuff 
they depend on (servers, the network itself) go away, but which return 
when the network comes back. I hate programs that complain or never 
recover from network outages.

And yes, we do use Gmail as an informal way of tracking the avaiability 
of remote machines. If my home servers arent listed as available, it 
means the network to the home is down. I think my next step would be to 
support shell commands to the server, which would be stunningly insecure 
but really cool.


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message