[File] [PATCH] Magdir/ FIG image *.fig; ASCII text versus data

Mon Feb 14 22:04:46 UTC 2022

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

some days ago i recompiled lilo boot loader source. In doc sub
directory are files with FIG name extension.

When running file command version 5.41 on such examples and related
files i get an output like:

COM-0-Modem-Stecker.fig:     data
NCC1701.FIG:                 data
bootloader.fig:              data
gnupg-card-architecture.fig: FIG image text, version 3.2, ASCII text
image.fig:                   data
map.fig:                     FIG image text, version 3.1, ASCII text
other.fig:                   data
thermostat.fig:              data

Furthermore with -i option only generic application/octet-stream or
text/plain is shown. With option --extension only 3 byte sequence ???
is shown.

The example thermostat.fig is part of xfig package. The example
COM-0-Modem-Stecker.fig was created by myself long time in the past.
So i know these examples are valid. So the non detection of lilo
example bootloader.fig is not an accident. This can be verified by
opening examples by xfig graphic tool. Or on the command line this
can be verified by command sequence like:
fig2mpdf -v bootloader.fig && file bootloader.pdf

For comparison reason i run the file format identification utility
TrID ( See https://mark0.net/soft-trid-e.html). Most examples,
especially also the "data" examples are described as "FIG vector
drawing" by fig.trid.xml (See appended fig-trid-v.txt.gz).

Luckily TrID tool displays correct file name extension FIG for
inspected samples. This list with -v option the URL pointing to
some information sites. That informations are expressed now by
comment lines inside Magdir/images like:
# URL: 		http://fileformats.archiveteam.org/wiki/Fig
#		https://en.wikipedia.org/wiki/Xfig
# Reference:	http://mark0.net/download/triddefs_xml.7z
#		defs/f/fig.trid.xml

The description happen by lines like:
0	search/1	#FIG		FIG image text
This only applies to text files and is equivalent to line like:
0	search/1/t	#FIG		FIG image text
Unfortunately a few of hundreds examples like COM-0-Modem-Stecker.fig
,
thermostat.fig and bootloader.fig contain Non ASCII text characters
like CTRL-A or phrase "°C" like in thermostat.fig. So these
examples are not detected by the magic line. These becomes vice
versa if i use a line with binary search option like:
 0	search/1/b	#FIG		FIG image text
When i changed this to another variant both are recognised but with
keep-going option i get text messages twice. I do not know if it a
BUG or feature. This is done by line like:
 0	search/1/tb	#FIG		FIG image text
So i finally use 2 different start magic lines. That look like
 0	search/1/ts	#FIG
 >&0	use		image-xfig
 0	search/1/bs	#FIG
 >&0	use		image-xfig
The displaying part is now done by new sub routine image-xfig. This
starts like:
 0	name		image-xfig
 >8	ubyte		x		FIG image text
 !:mime	image/x-xfig
 !:ext	fig
According to FreeDesktop.org shared MIME database instead of
generic mime type like text/plain these get their own user defined
one image/x-xfig. That information can be found for example on web
site
reposcope.com. After the start magic the version string (like 1.4
2.1 3.1 3.2) is shown by line like:
 >5	string		x		\b, version %.3s
Afterwards some times on first line comes text like: "Produced by
xfig version 3.2.5-alpha5". That information is now also shown by
line like:
 >8	ubyte		>0x0D
 >>8	string		x		"%s"
The TrID command also checks for point character (2Eh) of version
string. Probably no such exception examples exist. So maybe show
unexpected examples by line like:
 >6	ubyte		!0x2E		\b, at 6 %#x
Furthermore TrID assume that 4-byte magic pattern always start at
the beginning, but according to current file command this magic
could occur a little bit later. I my self do not found such
examples, but i keep the current behaviour.

Some "data" examples like CASTLE.FIG or NCC1701.FIG are described
as "DeskMate Draw drawing" by dm-fig.trid.xml (See appended
fig-trid-v.txt.gz).
Luckily TrID tool displays correct file name extension FIG for such
inspected samples. This list with -v option the URL pointing to
some information site:
That informations are expressed inside Magdir/images after XFIG
section by by comment lines like:
# URL: 		http://fileformats.archiveteam.org/wiki/DeskMate_Draw
#		http://en.wikipedia.org/wiki/Deskmate
# Reference:	http://mark0.net/download/triddefs_xml.7z
#		defs/d/dm-fig.trid.xml
According to that information such examples are now described by
lines like:
0	string		\x14FIG		DeskMate Drawing
!:mime	image/x-deskmate-fig
!:ext	fig
Instead of generic mime type application/octet-stream i display a
user defined on.

After applying the above mentioned modifications by patch
file-5.41-images-fig.diff then all my inspected examples are now
described as before, but the data examples are now also identified.
This now looks like:

COM-0-Modem-Stecker.fig:     FIG image text, version 2.1
NCC1701.FIG:                 DeskMate Drawing
bootloader.fig:              FIG image text, version 2.1
gnupg-card-architecture.fig: FIG image text, version 3.2
			     "  Produced by xfig
			     version 3.2.5-alpha5", ASCII text
image.fig:                   FIG image text, version 2.1
map.fig:                     FIG image text, version 3.1, ASCII text
other.fig:                   FIG image text, version 2.1
thermostat.fig:              FIG image text, version 3.2

I hope my diff file can be applied in future version of file
utility.

According to TrID there exist some other image files with fig file
name extension (fig-cabri.trid.xml fig-playmation.trid.xml), but i
do not find such examples.

With best wishes
Jörg Jenderek
- --
Jörg Jenderek
-----BEGIN PGP SIGNATURE-----
Comment: Using GnuPG with Thunderbird - https://www.enigmail.net/

iF0EARECAB0WIQS5/qNWKD4ASGOJGL+v8rHJQhrU1gUCYgrR/QAKCRCv8rHJQhrU
1qqKAJ9Y9aA075rV8SURZ10X+Btqmqe4NwCaAnP6oBZNdaVNEH5pk1mNrP7vvMI=
=bVd7
-----END PGP SIGNATURE-----
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fig-trid-v.txt.gz
Type: application/x-gzip
Size: 447 bytes
Desc: not available
URL: <https://mailman.astron.com/pipermail/file/attachments/20220214/7009a04b/attachment.bin>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: file-5.41-images-fig.diff
URL: <https://mailman.astron.com/pipermail/file/attachments/20220214/7009a04b/attachment.ksh>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: file-5.41-images-fig.diff.sig
Type: application/octet-stream
Size: 1296 bytes
Desc: not available
URL: <https://mailman.astron.com/pipermail/file/attachments/20220214/7009a04b/attachment.obj>