svn commit: r333919 - in head/contrib/file: . doc magic magic/Magdir python src tests

Andreas Tobler andreast at FreeBSD.org
Mon May 21 20:22:17 UTC 2018


Hi Eitan,

On 20.05.18 07:06, Eitan Adler wrote:
> Author: eadler
> Date: Sun May 20 05:06:42 2018
> New Revision: 333919
> URL: https://svnweb.freebsd.org/changeset/base/333919
> 
> Log:
>    MFV: file 5.33
>    
>    Merge the latest file(1) in.
>    
>    Relevent Changelog:
>    - extend the support for ${x?:} expansions for magic descriptions
>    - add support for ${x?:} in mime types to handle pie binaries.
>    - add support for negative offsets (offsets from the end of file)
>    - close the file on error when writing magic
>    
>    Relnotes:	yes

I bissected this commit as the one which breaks my ports build.
The one before, 333916 lets me build ports w/o problems.
This and the following one (333922) which fixed compilation leaves me 
with a weird situation.

For example, when I try to build x11-servers/xorg-server I do not find 
the libpciaccess.so lib, then the build tries to build the missing 
library and complains about the library is already there. Force 
installing it doesn't help, the configure step still complains about a 
libpciaccess.so which was not found.

Another example is a gccX build where it complains about libgmp.so not 
found.

This happens on amd64, two different machines, and also on armv7. I 
didn't try on other archs.

Stepping between the two revisions solves/exposes the issue.

Do you have an idea what happens here?

TIA,
Andreas


More information about the svn-src-head mailing list