ant-ivy-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ross Clewley <>
Subject Re: Validate setting not used for implicit resolve.
Date Thu, 16 Aug 2007 17:18:21 GMT

That was also my first thought but I think it is getting the rest of my
settings because it's correctly using my resolvers defined in the
settings file during the implicit resolve.

Normally when we're running 1.4.1 we have the usual configure target
call ivy:configure. I'm trying 2.0-alpha2 out on a branch in production
build system and when I started seeing this problem changed it to use
the new ivy:settings type to see what would happen. So now it's
something like this

  <target name="configure" depends="ivy-init,version">
    <ivy:settings file="${wpbuild.common.dir}/ivyconf.xml"

  <target name="resolve" depends="configure, ohome">
    <!--ivy:resolve conf="compile" useorigin="true"/-->
    <ivy:cachepath pathid="compile.classpath" conf="compile"

Where the settings file ${wpbuild.common.dir}/ivyconf.xml is

     <settings defaultCache="${ivy.cache.dir}" defaultResolver="default"

If I uncomment the explicit resolve it works. With the resolve
commented, I get validation errors for because of our extended
attributes. I also tried explicitly setting the settingsref on the
cachepath but that did not help.



Xavier Hanin wrote:
> On 8/16/07, Ross Clewley <> wrote:
>> Hi
>> Playing with 2.0-alpha2 a bit I notice that if my project does an
>> implicit resolve because I do a post resolve task involving a conf that
>> I haven't explicitly resolve, then it doesn't seem to pick up the
>> default "validate" setting from the settings file. If the resolve is
>> explicit, then the default "validate" setting from the settings file is
>> used.
>> I assume this isn't what's desired? If not I will log a jira issue.
> No, this is not desired. But are you sure your settings are loaded in both
> cases? How do you specify the settings? Could you give a small extract of
> your build.xml?
> Xavier
> thanks
>> Ross

View raw message