[ 
https://issues.apache.org/jira/browse/IVY-1272?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Oliver Jowett updated IVY-1272:
-------------------------------

    Description: 
If I declare a <cachefileset> for a configuration that is present but empty, I 
get a fileset that (as expected) matches no files.
The problem is that when used, that fileset decides to match no files by 
recursively scanning all files starting from the current working directory 
(checked via strace and via a JVM thread dump)
In our case, often the CWD is the top of our build tree, so this takes a very 
long time to work out that the fileset is in fact empty.

Specifically, we're doing this (inside a macrodef)

            <ivy:cachefileset setid="@{module}.docdeps" conf="doc" 
type="javadoc.zip"/>
            <unzip dest="${target}/javadoc">
                <patternset>
                    <include name="*/package-list"/>
                </patternset>
                <fileset refid="@{module}.docdeps"/>
            </unzip>

and the unzip step takes a very, very, long time to do nothing when the "doc" 
conf is empty.
If I replace the use of cachefileset with an explicit fileset that matches 
nothing, it runs fast as expected.

  was:
If I declare a <cachefileset> for a configuration that is present but empty, I 
get a fileset that (as expected) matches no files.
The problem is that when used, that fileset decides to match no files by 
recursively scanning all files starting from the current working directly 
(checked via strace and via a JVM thread dump)
In our case, often the CWD is the top of our build tree, so this takes a very 
long time to work out that the fileset is in fact empty.

Specifically, we're doing this (inside a macrodef)

            <ivy:cachefileset setid="@{module}.docdeps" conf="doc" 
type="javadoc.zip"/>
            <unzip dest="${target}/javadoc">
                <patternset>
                    <include name="*/package-list"/>
                </patternset>
                <fileset refid="@{module}.docdeps"/>
            </unzip>

and the unzip step takes a very, very, long time to do nothing when the "doc" 
conf is empty.
If I replace the use of cachefileset with an explicit fileset that matches 
nothing, it runs fast as expected.


> <cachefileset> on an empty configuration produces a very slow-to-evaluate 
> fileset
> ---------------------------------------------------------------------------------
>
>                 Key: IVY-1272
>                 URL: https://issues.apache.org/jira/browse/IVY-1272
>             Project: Ivy
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 2.2.0-RC1
>         Environment: Ant 1.8.1
> Ivy 2.2.0-rc1
>            Reporter: Oliver Jowett
>            Priority: Minor
>
> If I declare a <cachefileset> for a configuration that is present but empty, 
> I get a fileset that (as expected) matches no files.
> The problem is that when used, that fileset decides to match no files by 
> recursively scanning all files starting from the current working directory 
> (checked via strace and via a JVM thread dump)
> In our case, often the CWD is the top of our build tree, so this takes a very 
> long time to work out that the fileset is in fact empty.
> Specifically, we're doing this (inside a macrodef)
>             <ivy:cachefileset setid="@{module}.docdeps" conf="doc" 
> type="javadoc.zip"/>
>             <unzip dest="${target}/javadoc">
>                 <patternset>
>                     <include name="*/package-list"/>
>                 </patternset>
>                 <fileset refid="@{module}.docdeps"/>
>             </unzip>
> and the unzip step takes a very, very, long time to do nothing when the "doc" 
> conf is empty.
> If I replace the use of cachefileset with an explicit fileset that matches 
> nothing, it runs fast as expected.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

Reply via email to