ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From peter reilly <peter.rei...@corvil.com>
Subject RE: [Patch] keep-going feature
Date Tue, 08 Jul 2003 16:08:45 GMT
Opps, sorry that should be keep-going (anyway).

Yep, you are right the build does fail, so it
is not equivalent to an implicit fail-on-error attribute
on each target (which I think would be extremely confusing).

Peter.
ps: perhaps we could put the link in the main ant page
    just after the "make is evil" paragraph ;-)

On Tue, 2003-07-08 at 16:52, Alexey Solofnenko wrote:
> Actually the reason to implement keep-alive (originally keep-going) is to
> find as many problems as possible. It is done by executing all targets that
> do not depend directly or indirectly on failed targets. It is not
> fail-on-error flag - the build will still fail.
> 
> There is a link that explaining the feature in make:
> http://www.delorie.com/gnu/docs/make/make_52.html#IDX188
> 
> - Alexey.
> 
> -----Original Message-----
> From: peter reilly [mailto:peter.reilly@corvil.com] 
> Sent: Tuesday, July 08, 2003 7:18 AM
> To: Ant Developers List
> Subject: Re: [Patch] keep-alive feature
> 
> The keep-alive feature is not quite the same
> as a fail-on-error on each task, it is more
> like a fail-on-error for each target.
> 
> I have test-driven it in my build env where
> I have a large number of c++ programs to compile.
> 
> It is nice to able to change a header file and
> then compile all the programs again, and use next-error
> to hop tru all the errors.
> 
> For the nightly build and for normal use however, I want to
> fail on the first error.
> 
> Peter
> On Tue, 2003-07-08 at 08:41, Stefan Bodewig wrote:
> > On 07 Jul 2003, peter reilly <peter.reilly@corvil.com> wrote:
> > 
> > > I am thinking of committing the keep-alive
> > > feature:
> > > 
> > > http://issues.apache.org/bugzilla/show_bug.cgi?id=21144
> > > 
> > > Do any of the ant commiters have a problem with
> > > this feature?
> > 
> > Well, we've already introduced one magic attribute with the
> > polymorphism patch (will introduce).  I'd rather turn the various
> > failonerror attributes into a single magic fail-on-error, which would
> > be handled by Task.
> > 
> > The command line option could then be used to provide a default value
> > for the new attribute, if it really is necessary.  I'm more on Steve's
> > side that I wouldn't want to keep going no matter where the error
> > occurs.
> > 
> > Stefan
> > 
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@ant.apache.org
> > For additional commands, e-mail: dev-help@ant.apache.org
> > 
> > 
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@ant.apache.org
> For additional commands, e-mail: dev-help@ant.apache.org
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@ant.apache.org
> For additional commands, e-mail: dev-help@ant.apache.org
> 
> 


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@ant.apache.org
For additional commands, e-mail: dev-help@ant.apache.org


Mime
View raw message