ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran" <stev...@iseran.com>
Subject Re: [DISC] core extensions
Date Tue, 27 Mar 2001 19:21:48 GMT

----- Original Message -----
From: "Stefan Bodewig" <bodewig@apache.org>
To: <ant-dev@jakarta.apache.org>
Sent: Tuesday, March 27, 2001 04:47
Subject: Re: [DISC] core extensions

> > * specify an onfail task or target that runs in case of a build
> >   failure.
>
> And what if that one fails? -1

the onfail target specified in the failure task gets called, and so on:

<target name="deploy" onfail="deployfail">
  <ftp ... />
</target>

<target name="deployfail" onfail="page-steve">
  <telnet .../>
</target>

<target name="page-steve" onfail="do-nothing">
 <mail tolist="${pager.steve}" message="big build failure"/>
</target>

<target name="do-nothing" />

Where the concept gets confusing is in dependencies. Should failure tasks
have all their dependencies executed? It should only be when things start
going wrong that they are called, but then what happens when the dependent
tasks start to fail. Should the 'onfail' task of the target highest up the
tree be the default 'handler' for failures (like VB's ON ERROR GOTO hack)?
It soon gets really complex, in the absence of the 'backtracking model of
prolog'

-steve



Mime
View raw message