Problems with release checking

[sebb-2-2]

It was quite tedious checking the 4.0.1 release because of spurious
variations between the tgz and zip archives.

Apart from the LF/CRLF line ending differences, which are easy to
allow for, there are quite a lot of files in the ZIP archive that have
an additional CRLF at the end. This can be fixed by adding

fixlast="false"

to the Ant script which repackages the zip file.

BTW, the Ant script is not suitable for use on Windows, because it
only fixes the zip file.
It ought to be changed to detect the EOL type and fix the appropriate archive.

The $HeadURL:$ SVN markers can cause problems when comparing the
source archive with the tag unless one uses the same URL as the
builder used.

HeadURL causes major problems when comparing SVN checkouts.
I think the markers should be removed (or disabled).

Similarly, the $Date:$ SVN marker causes problems when comparing
files, because the date is expressed as a local date. This can be
allowed for in automated comparisons, but is still a nuisance. IMO,
the benefits of having the Date outweigh the disadvantages. If one
really needs the date, then consider using $Id which uses UTC, for
example:

$Id: calc.c 148 2006-07-28 21:30:43Z sally $

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

[olegk]

sebb wrote:
> It was quite tedious checking the 4.0.1 release because of spurious
> variations between the tgz and zip archives.
>
> Apart from the LF/CRLF line ending differences, which are easy to
> allow for, there are quite a lot of files in the ZIP archive that have
> an additional CRLF at the end. This can be fixed by adding
>
> fixlast="false"
>

Sebastian,

Just go ahead and fix obvious issues without too much of a protocol.

> to the Ant script which repackages the zip file.
>
> BTW, the Ant script is not suitable for use on Windows, because it
> only fixes the zip file.
> It ought to be changed to detect the EOL type and fix the appropriate archive.
>

How can one reliably detect the EOL type of an archive with ant or
maven? Can you elaborate?

> The $HeadURL:$ SVN markers can cause problems when comparing the
> source archive with the tag unless one uses the same URL as the
> builder used.
>
> HeadURL causes major problems when comparing SVN checkouts.
> I think the markers should be removed (or disabled).
>
> Similarly, the $Date:$ SVN marker causes problems when comparing
> files, because the date is expressed as a local date. This can be
> allowed for in automated comparisons, but is still a nuisance. IMO,
> the benefits of having the Date outweigh the disadvantages. If one
> really needs the date, then consider using $Id which uses UTC, for
> example:
>
> $Id: calc.c 148 2006-07-28 21:30:43Z sally $
>

Let's just get rid of those in trunk. Hopefully there will be no need
for another release off the 4.0.x branch.

Oleg


> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@...
> For additional commands, e-mail: dev-help@...
>


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

[sebb-2-2]

On 21/06/2009, Oleg Kalnichevski <olegk@...> wrote:

> sebb wrote:
>
> > It was quite tedious checking the 4.0.1 release because of spurious
> > variations between the tgz and zip archives.
> >
> > Apart from the LF/CRLF line ending differences, which are easy to
> > allow for, there are quite a lot of files in the ZIP archive that have
> > an additional CRLF at the end. This can be fixed by adding
> >
> > fixlast="false"
> >
> >
>
>  Sebastian,
>
>  Just go ahead and fix obvious issues without too much of a protocol.
>

Thanks!
Sometimes an obvious issue for me is not obvious to others, so I
thought I better raise it.

>
> > to the Ant script which repackages the zip file.
> >
> > BTW, the Ant script is not suitable for use on Windows, because it
> > only fixes the zip file.
> > It ought to be changed to detect the EOL type and fix the appropriate
> archive.
> >
> >
>
>  How can one reliably detect the EOL type of an archive with ant or maven?
> Can you elaborate?

I meant check the EOL type of the OS, and fix whichever archive needs
a different EOL.

I use the following check for JMeter:

        <condition property="native.lf">
            <os family="unix"/>
        </condition>
        <condition property="native.crlf">
            <os family="dos"/>
        </condition>

and then use the condition to decide what to do.

>
> > The $HeadURL:$ SVN markers can cause problems when comparing the
> > source archive with the tag unless one uses the same URL as the
> > builder used.
> >
> > HeadURL causes major problems when comparing SVN checkouts.
> > I think the markers should be removed (or disabled).
> >
> > Similarly, the $Date:$ SVN marker causes problems when comparing
> > files, because the date is expressed as a local date. This can be
> > allowed for in automated comparisons, but is still a nuisance. IMO,
> > the benefits of having the Date outweigh the disadvantages. If one
> > really needs the date, then consider using $Id which uses UTC, for
> > example:
> >
> > $Id: calc.c 148 2006-07-28 21:30:43Z sally $
> >
> >
>
>  Let's just get rid of those in trunk. Hopefully there will be no need for
> another release off the 4.0.x branch.

OK.

>  Oleg
>
>
>
> >
> ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@...
> > For additional commands, e-mail: dev-help@...
> >
> >
>
>
> ---------------------------------------------------------------------
>  To unsubscribe, e-mail: dev-unsubscribe@...
>  For additional commands, e-mail: dev-help@...
>
>

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

[sebb-2-2]

On 21/06/2009, sebb <sebbaz@...> wrote:

> On 21/06/2009, Oleg Kalnichevski <olegk@...> wrote:
>  > sebb wrote:
>  >
>  > > It was quite tedious checking the 4.0.1 release because of spurious
>  > > variations between the tgz and zip archives.
>  > >
>  > > Apart from the LF/CRLF line ending differences, which are easy to
>  > > allow for, there are quite a lot of files in the ZIP archive that have
>  > > an additional CRLF at the end. This can be fixed by adding
>  > >
>  > > fixlast="false"
>  > >
>  > >
>  >
>  >  Sebastian,
>  >
>  >  Just go ahead and fix obvious issues without too much of a protocol.
>  >

I've updated the build.xml file and am trying to test using

mvn package assembly:assembly -DskipTests=true

However that fails with the following error:

[INFO] [bundle:bundle]
[ERROR] Error building bundle
org.apache.httpcomponents:httpcore-osgi:bundle:4.1-SNAPSHOT : Input
file does not exist: target/maven-shared-archive-res
ources/META-INF/DEPENDENCIES~
[ERROR] Error building bundle
org.apache.httpcomponents:httpcore-osgi:bundle:4.1-SNAPSHOT : Input
file does not exist: target/maven-shared-archive-res
ources/META-INF/LICENSE~
[ERROR] Error building bundle
org.apache.httpcomponents:httpcore-osgi:bundle:4.1-SNAPSHOT : Input
file does not exist: target/maven-shared-archive-res
ources/META-INF/NOTICE~
[ERROR] Error(s) found in bundle configuration

So I commented out the httpcore-osgi module and the build then failed
because it could not find index.pdf. Running

mvn docbkx:generate-pdf

fixed that problem, and the archives are created OK. However the
build.xml file is not called to process the archives.

Is there some special Maven flag needed to invoke this phase?

>
> Thanks!
>  Sometimes an obvious issue for me is not obvious to others, so I
>  thought I better raise it.
>
>
>  >
>  > > to the Ant script which repackages the zip file.
>  > >
>  > > BTW, the Ant script is not suitable for use on Windows, because it
>  > > only fixes the zip file.
>  > > It ought to be changed to detect the EOL type and fix the appropriate
>  > archive.
>  > >
>  > >
>  >
>  >  How can one reliably detect the EOL type of an archive with ant or maven?
>  > Can you elaborate?
>
>
> I meant check the EOL type of the OS, and fix whichever archive needs
>  a different EOL.
>
>  I use the following check for JMeter:
>
>         <condition property="native.lf">
>             <os family="unix"/>
>         </condition>
>         <condition property="native.crlf">
>             <os family="dos"/>
>         </condition>
>
>  and then use the condition to decide what to do.
>
>
>  >
>  > > The $HeadURL:$ SVN markers can cause problems when comparing the
>  > > source archive with the tag unless one uses the same URL as the
>  > > builder used.
>  > >
>  > > HeadURL causes major problems when comparing SVN checkouts.
>  > > I think the markers should be removed (or disabled).
>  > >
>  > > Similarly, the $Date:$ SVN marker causes problems when comparing
>  > > files, because the date is expressed as a local date. This can be
>  > > allowed for in automated comparisons, but is still a nuisance. IMO,
>  > > the benefits of having the Date outweigh the disadvantages. If one
>  > > really needs the date, then consider using $Id which uses UTC, for
>  > > example:
>  > >
>  > > $Id: calc.c 148 2006-07-28 21:30:43Z sally $
>  > >
>  > >
>  >
>  >  Let's just get rid of those in trunk. Hopefully there will be no need for
>  > another release off the 4.0.x branch.
>
>
> OK.
>
>
>  >  Oleg
>  >
>  >
>  >
>  > >
>  > ---------------------------------------------------------------------
>  > > To unsubscribe, e-mail: dev-unsubscribe@...
>  > > For additional commands, e-mail: dev-help@...
>  > >
>  > >
>  >
>  >
>  > ---------------------------------------------------------------------
>  >  To unsubscribe, e-mail: dev-unsubscribe@...
>  >  For additional commands, e-mail: dev-help@...
>  >
>  >
>

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

[olegk]

sebb wrote:

> On 21/06/2009, sebb <sebbaz@...> wrote:
>> On 21/06/2009, Oleg Kalnichevski <olegk@...> wrote:
>>  > sebb wrote:
>>  >
>>  > > It was quite tedious checking the 4.0.1 release because of spurious
>>  > > variations between the tgz and zip archives.
>>  > >
>>  > > Apart from the LF/CRLF line ending differences, which are easy to
>>  > > allow for, there are quite a lot of files in the ZIP archive that have
>>  > > an additional CRLF at the end. This can be fixed by adding
>>  > >
>>  > > fixlast="false"
>>  > >
>>  > >
>>  >
>>  >  Sebastian,
>>  >
>>  >  Just go ahead and fix obvious issues without too much of a protocol.
>>  >
>
> I've updated the build.xml file and am trying to test using
>
> mvn package assembly:assembly -DskipTests=true
>
> However that fails with the following error:
>
> [INFO] [bundle:bundle]
> [ERROR] Error building bundle
> org.apache.httpcomponents:httpcore-osgi:bundle:4.1-SNAPSHOT : Input
> file does not exist: target/maven-shared-archive-res
> ources/META-INF/DEPENDENCIES~
> [ERROR] Error building bundle
> org.apache.httpcomponents:httpcore-osgi:bundle:4.1-SNAPSHOT : Input
> file does not exist: target/maven-shared-archive-res
> ources/META-INF/LICENSE~
> [ERROR] Error building bundle
> org.apache.httpcomponents:httpcore-osgi:bundle:4.1-SNAPSHOT : Input
> file does not exist: target/maven-shared-archive-res
> ources/META-INF/NOTICE~
> [ERROR] Error(s) found in bundle configuration
>
> So I commented out the httpcore-osgi module and the build then failed
> because it could not find index.pdf. Running
>
> mvn docbkx:generate-pdf
>
> fixed that problem, and the archives are created OK. However the
> build.xml file is not called to process the archives.
>
> Is there some special Maven flag needed to invoke this phase?
>

There appears to be some kind of issue with the latest maven bundle
plugin (don't we all love maven?). I raised a JIRA with Apache Felix [1]
and downgraded the plugin to previous stable version for the time being.

Please pull the latest snapshot and try again.

Cheers

Oleg


[1] https://issues.apache.org/jira/browse/FELIX-1262

>> Thanks!
>>  Sometimes an obvious issue for me is not obvious to others, so I
>>  thought I better raise it.
>>
>>
>>  >
>>  > > to the Ant script which repackages the zip file.
>>  > >
>>  > > BTW, the Ant script is not suitable for use on Windows, because it
>>  > > only fixes the zip file.
>>  > > It ought to be changed to detect the EOL type and fix the appropriate
>>  > archive.
>>  > >
>>  > >
>>  >
>>  >  How can one reliably detect the EOL type of an archive with ant or maven?
>>  > Can you elaborate?
>>
>>
>> I meant check the EOL type of the OS, and fix whichever archive needs
>>  a different EOL.
>>
>>  I use the following check for JMeter:
>>
>>         <condition property="native.lf">
>>             <os family="unix"/>
>>         </condition>
>>         <condition property="native.crlf">
>>             <os family="dos"/>
>>         </condition>
>>
>>  and then use the condition to decide what to do.
>>
>>
>>  >
>>  > > The $HeadURL:$ SVN markers can cause problems when comparing the
>>  > > source archive with the tag unless one uses the same URL as the
>>  > > builder used.
>>  > >
>>  > > HeadURL causes major problems when comparing SVN checkouts.
>>  > > I think the markers should be removed (or disabled).
>>  > >
>>  > > Similarly, the $Date:$ SVN marker causes problems when comparing
>>  > > files, because the date is expressed as a local date. This can be
>>  > > allowed for in automated comparisons, but is still a nuisance. IMO,
>>  > > the benefits of having the Date outweigh the disadvantages. If one
>>  > > really needs the date, then consider using $Id which uses UTC, for
>>  > > example:
>>  > >
>>  > > $Id: calc.c 148 2006-07-28 21:30:43Z sally $
>>  > >
>>  > >
>>  >
>>  >  Let's just get rid of those in trunk. Hopefully there will be no need for
>>  > another release off the 4.0.x branch.
>>
>>
>> OK.
>>
>>
>>  >  Oleg
>>  >
>>  >
>>  >
>>  > >
>>  > ---------------------------------------------------------------------
>>  > > To unsubscribe, e-mail: dev-unsubscribe@...
>>  > > For additional commands, e-mail: dev-help@...
>>  > >
>>  > >
>>  >
>>  >
>>  > ---------------------------------------------------------------------
>>  >  To unsubscribe, e-mail: dev-unsubscribe@...
>>  >  For additional commands, e-mail: dev-help@...
>>  >
>>  >
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@...
> For additional commands, e-mail: dev-help@...
>

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

[sebb-2-2]

On 21/06/2009, Oleg Kalnichevski <olegk@...> wrote:

> sebb wrote:
>
> > On 21/06/2009, sebb <sebbaz@...> wrote:
> >
> > > On 21/06/2009, Oleg Kalnichevski <olegk@...> wrote:
> > >  > sebb wrote:
> > >  >
> > >  > > It was quite tedious checking the 4.0.1 release because of spurious
> > >  > > variations between the tgz and zip archives.
> > >  > >
> > >  > > Apart from the LF/CRLF line ending differences, which are easy to
> > >  > > allow for, there are quite a lot of files in the ZIP archive that
> have
> > >  > > an additional CRLF at the end. This can be fixed by adding
> > >  > >
> > >  > > fixlast="false"
> > >  > >
> > >  > >
> > >  >
> > >  >  Sebastian,
> > >  >
> > >  >  Just go ahead and fix obvious issues without too much of a protocol.
> > >  >
> > >
> >
> > I've updated the build.xml file and am trying to test using
> >
> > mvn package assembly:assembly -DskipTests=true
> >
> > However that fails with the following error:
> >
> > [INFO] [bundle:bundle]
> > [ERROR] Error building bundle
> >
> org.apache.httpcomponents:httpcore-osgi:bundle:4.1-SNAPSHOT
> : Input
> > file does not exist: target/maven-shared-archive-res
> > ources/META-INF/DEPENDENCIES~
> > [ERROR] Error building bundle
> >
> org.apache.httpcomponents:httpcore-osgi:bundle:4.1-SNAPSHOT
> : Input
> > file does not exist: target/maven-shared-archive-res
> > ources/META-INF/LICENSE~
> > [ERROR] Error building bundle
> >
> org.apache.httpcomponents:httpcore-osgi:bundle:4.1-SNAPSHOT
> : Input
> > file does not exist: target/maven-shared-archive-res
> > ources/META-INF/NOTICE~
> > [ERROR] Error(s) found in bundle configuration
> >
> > So I commented out the httpcore-osgi module and the build then failed
> > because it could not find index.pdf. Running
> >
> > mvn docbkx:generate-pdf
> >
> > fixed that problem, and the archives are created OK. However the
> > build.xml file is not called to process the archives.
> >
> > Is there some special Maven flag needed to invoke this phase?
> >
> >
>
>  There appears to be some kind of issue with the latest maven bundle plugin
> (don't we all love maven?). I raised a JIRA with Apache Felix [1] and
> downgraded the plugin to previous stable version for the time being.
>
>  Please pull the latest snapshot and try again.

That has fixed the problem with the bundle processing, but I still
don't see the execution of build.xml.

>  Cheers
>
>  Oleg
>
>
>  [1] https://issues.apache.org/jira/browse/FELIX-1262
>
>
>
> >
> > > Thanks!
> > >  Sometimes an obvious issue for me is not obvious to others, so I
> > >  thought I better raise it.
> > >
> > >
> > >  >
> > >  > > to the Ant script which repackages the zip file.
> > >  > >
> > >  > > BTW, the Ant script is not suitable for use on Windows, because it
> > >  > > only fixes the zip file.
> > >  > > It ought to be changed to detect the EOL type and fix the
> appropriate
> > >  > archive.
> > >  > >
> > >  > >
> > >  >
> > >  >  How can one reliably detect the EOL type of an archive with ant or
> maven?
> > >  > Can you elaborate?
> > >
> > >
> > > I meant check the EOL type of the OS, and fix whichever archive needs
> > >  a different EOL.
> > >
> > >  I use the following check for JMeter:
> > >
> > >        <condition property="native.lf">
> > >            <os family="unix"/>
> > >        </condition>
> > >        <condition property="native.crlf">
> > >            <os family="dos"/>
> > >        </condition>
> > >
> > >  and then use the condition to decide what to do.
> > >
> > >
> > >  >
> > >  > > The $HeadURL:$ SVN markers can cause problems when comparing the
> > >  > > source archive with the tag unless one uses the same URL as the
> > >  > > builder used.
> > >  > >
> > >  > > HeadURL causes major problems when comparing SVN checkouts.
> > >  > > I think the markers should be removed (or disabled).
> > >  > >
> > >  > > Similarly, the $Date:$ SVN marker causes problems when comparing
> > >  > > files, because the date is expressed as a local date. This can be
> > >  > > allowed for in automated comparisons, but is still a nuisance. IMO,
> > >  > > the benefits of having the Date outweigh the disadvantages. If one
> > >  > > really needs the date, then consider using $Id which uses UTC, for
> > >  > > example:
> > >  > >
> > >  > > $Id: calc.c 148 2006-07-28 21:30:43Z sally $
> > >  > >
> > >  > >
> > >  >
> > >  >  Let's just get rid of those in trunk. Hopefully there will be no
> need for
> > >  > another release off the 4.0.x branch.
> > >
> > >
> > > OK.
> > >
> > >
> > >  >  Oleg
> > >  >
> > >  >
> > >  >
> > >  > >
> > >  >
> ---------------------------------------------------------------------
> > >  > > To unsubscribe, e-mail: dev-unsubscribe@...
> > >  > > For additional commands, e-mail: dev-help@...
> > >  > >
> > >  > >
> > >  >
> > >  >
> > >  >
> ---------------------------------------------------------------------
> > >  >  To unsubscribe, e-mail: dev-unsubscribe@...
> > >  >  For additional commands, e-mail: dev-help@...
> > >  >
> > >  >
> > >
> > >
> >
> >
> ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@...
> > For additional commands, e-mail: dev-help@...
> >
> >
>
>
> ---------------------------------------------------------------------
>  To unsubscribe, e-mail: dev-unsubscribe@...
>  For additional commands, e-mail: dev-help@...
>
>

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

[olegk]

sebb wrote:

>>  There appears to be some kind of issue with the latest maven bundle plugin
>> (don't we all love maven?). I raised a JIRA with Apache Felix [1] and
>> downgraded the plugin to previous stable version for the time being.
>>
>>  Please pull the latest snapshot and try again.
>
> That has fixed the problem with the bundle processing, but I still
> don't see the execution of build.xml.
>

The script does not execute automatically. It needs to be executed
manually by running mvn antrun:run

The complete maven command to build release assemblies should look like:

mvn javadoc:javadoc docbkx:generate-pdf docbkx:generate-html package
assembly:assembly antrun:run

Ain't maven fun?

Oleg

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