Hello trid users,
some days ago i do some retro stuff. I run trid on GEM GDOS fonts with file
extension fnt or gft. Some are not recognized by fnt-gdos.trid.xml ( see
appended output/trid-old.txt).
The newest version of file utility { see Wikipedia
https://en.wikipedia.org/wiki/File_(command) } identifies the unknown
files correctly as GEM GDOS font ( see appended output/file-new.txt).
So i run tridscan to refine trid definition file. Afterwards i look at the
differences. According to documentation many values are stored as short
value. If value is below 256 then upper byte is 0. So in old definition often
found such pattern expressed by XML construct:
<Bytes>00</Bytes>
<Pos>39</Pos>
According to documentation at offset 38 highest character index in face is
stored. For many fonts this values is below 256. But when inspecting more
fonts i found example with value above 256. Then in definition file the
pattern vanish.
Font files begin with a header 88 bytes long. Afterwards tables and font
data is stored. When inspecting many examples tables and data should be
different. So all patterns with offset about 88 vanish like
<Pattern>
<Bytes>00</Bytes>
<Pos>95</Pos>
</Pattern>
With the updated trid definition examples are now recognized ( see
appended output/trid-v-new.txt).
Furthermore i also found font examples with eps file name extension like in
example AA070GEP.EPS or AI360GEP.EPS. Do not know if this is an accident,
because normally eps is used for encapsulated postscript.
But such files are also identified as GEM GDOS fonts ( See
eps/output/trid-old.txt) .
According to documentation for many GEM GDOS fonts ASCII string "UUUU"
usually appears at offset 62. But this not always true. So i find some fonts
with lightening mask and skewing mask values 0 instead 0x5555 ( ASCII UU)
So i generate variant fnt-gdos0mask.trid.xml where 0 masks values
appear. This is express by changed XML construct:
<Bytes>0000000000</Bytes>
<Pos>61</Pos>
With this second trid definition file now null mask GEM fonts are identified (
see 0mask/output/trid-new.txt), but also other files like BACKM212.COM or
many CD images like plpbt.iso are misidentified.
Now it depends on detection philosophy. If you think bad identification is
better than no identification. Then use fnt-gdos0mask.trid.xml in your trid
definitions.
If you do not like false hits, when leave it. Then also remove
abr.trid.xml for Adobe PhotoShop Brush, sybase-ianywhere-dbf.trid.xml for
Advantage Database Systems, because these XML files give also unprecise
identifications.
TrID definitions and output are stored in archive fnt_.zip. I hope that my
XML file can be used in future version of triddefs.
With best wishes
J?rg Jenderek
