Re: Fedora ARM is so strong
On 24.02.2014 23:52, Ed Greshko wrote:
On 02/25/14 01:02, poma wrote:
> $ file *
> autocorr-en-4.2.1.1-1.fc20.noarch.rpm: RPM v3.0 bin ARM
> autocorr-en-1:4.2.1.1-1.fc20
> libcmis-0.4.1-2.fc20.x86_64.rpm: RPM v3.0 bin
> i386/x86_64 libcmis-0.4.1-2.fc20
> libreoffice-4.2.1.1-1.fc20.x86_64.rpm: RPM v3.0 bin
> i386/x86_64 libreoffice-1:4.2.1.1-1.fc20
> libreoffice-appdata-4.2.1.1-1.fc20.noarch.rpm: RPM v3.0 bin ARM
> libreoffice-appdata-1:4.2.1.1-1.fc20
>
> Has anyone noticed something oddly in this list? :)
>
>
> poma
So, file a bz against "file".
Actually this is not the 'file' bug.
The 'file' correctly "reads" rpms.
$ wc -l noarch.update-7416sum\=1965noarch+5451ARM.txt
7416 noarch.update-7416sum=1965noarch+5451ARM.txt
$ grep ARM noarch.update-7416sum\=1965noarch+5451ARM.txt | wc -l
5451
$ grep -v ARM noarch.update-7416sum\=1965noarch+5451ARM.txt | wc -l
1965
'noarch.update-7416sum\=1965noarch+5451ARM.txt'
http://goo.gl/O2T79S
e.g. "RPM v3.0 bin"
- hexedit:
00000000 ED AB EE DB 03 00 00 00 ...
- file/magic/Magdir/rpm:
- ED AB EE DB == RPM
0 belong 0xedabeedb RPM
- 03 == v3
4 byte x v%d
- 00 == .0
5 byte x \b.%d
- 00 == bin
6 beshort 0 bin
e.g. 'noarch' -> OK!
$ file -b kdenetwork-devel-4.12.2-1.fc20.noarch.rpm
... noarch ...
- hexedit:
... 00 FF ...
- file/magic/Magdir/rpm:
>8 beshort 255 noarch
i.e. dec 255 == hex FF
e.g. 'ARM' -> WTF!?
$ file -b mdds-devel-0.10.2-1.fc20.noarch.rpm
... ARM ...
- hexedit:
... 00 0C ...
- file/magic/Magdir/rpm:
>8 beshort 12 ARM
i.e. dec 12 == hex C
'file' - rpm
https://raw.github.com/glensc/file/master/magic/Magdir/rpm
Now to whom/where to send this?
poma