If I understand Stefan right - thank's for that history lesson ;-) - you
don't store anything in /PaXHeaders.
You just store an 'illegal' file, e.g. with a file path longer than 250
characters.

Then your "newer" untar tool will work as usual and just extracts these
long-path-files.
An "older" untar tool can't interpret the content of the tar in the "newer"
way.

Jan

-----Ursprüngliche Nachricht-----
Von: S Abirami <s.abir...@ericsson.com.INVALID> 
Gesendet: Freitag, 29. Januar 2021 09:52
An: Ant Developers List <dev@ant.apache.org>
Betreff: RE: PaxHeaders folder is missing upgrade from 1.8.2 to 1.10.9

Hi Stefan,

Our untar task will just untar *.so file and  *.dll file inside gzip and
tar.gz.
We didn't store anything inside a PaxHeader folder.

I aware the usage of PaxHeader to determine the higher size of file names
available in tar. In our application, we don't have big file name which has
more than 100 characters.
Just to confirm the PaxHeader missing is expected in Ant build of 1.10.9
version or we are missing anything.

Regards,
Abirami.S
-----Original Message-----
From: Stefan Bodewig <bode...@apache.org>
Sent: Friday, January 29, 2021 1:40 PM
To: dev@ant.apache.org
Subject: Re: PaxHeaders folder is missing upgrade from 1.8.2 to 1.10.9

On 2021-01-28, S Abirami wrote:

> In our build.xml, we are using the untar task to untar gzip and tar.gz
file.
> When we tar, always we saw a PaxHeaders.* folder in the untar location.
> After upgrading from 1.8.2 to 1.10.9 version of Ant build.

> We are not seeing this folder. We want to know whether missing of that 
> folder will have any impact on the untar files.

1.8.2 is a looong time ago and a lot has changed. :-)

When tar became a POSIX standard a few ways to break out of the limits of
the original format have been been invented. One of them was the
introduction of an additional type of entry - a so called PAX header - that
could contain (meta-)data not compatible with the original format. A common
example would be non-ASCII filenames or paths longer than 250 characters.

A tar implementation that doesn't understand these extensions will see them
as files inside a directory called PaXHeaders - an tar implementation that
does understand them will parse them and not treat the headers as actual
files at all.

The Java classes implementing tar initially didn't understand the POSIX
extensions and have ben gradually improved. Looking through Ant's changelogs
I believe with 1.9.0 they started to understand PAX headers.

So what your untar task does now is actually applying the data contained
inside the PAX headers where possible instead of ignoring them.

Stefan

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


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



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

Reply via email to