The term dignorary is made of the words ignore and dictionary and denotes a dictionary storing names to be ignored during vahunzation.
In this context, ignore means do not vahunz a name - just like if
it would have been marked with a blank (' '). A name can be set to be
ignored by changing the first column of its line to a hyphen (-).
Names marked to be ignored are not stored in the normal dictionary - by default called vahunz.names -, but in another file by default called vahunz.ignore. Still ignored names can be left in the normal dictionary at first, as Vahunz takes care of writing them to the dignorary when it is finished with scanning all names.
Vahunz is already distributed with some often needed dignoraries.
You can find them in the vahunz/dignorary directory. To
use one of them, simply copy it to vahunz.ignore. Assuming you did
not set --base-name to something else.
<assert.h>,
<ctype.h>,
<errno.h>,
<limits.h>,
<locale.h>,
<signal.h>,
<stdarg.h>,
<stddef.h>,
<stdio.h>,
<stdlib.h>,
<string.h> and
<time.h>.
As these dignoraries where created scanning the include files of
Geek Gadget's gcc for Amiga, they may also include some
names you would not expect to be there. For example,
_KERNEL is not really an ANSI symbol. However, this
should normally not cause any problems.
If none of the included dignoraries can satisfy your needs, you will have to create your own ones. This can easily be performed using Vahunz-Tschak. But if you prefer, you can also do it manually. This is how the amiga.ignore was created:
First, a file list for all header files stored on my include directory was created in the temporary t: directory:
list >t:amiga.files lformat=%p%n all pat=#?.h include:After that, Vahunz was told to scan them for names:
vahunz --base-name t:amiga
As a result, all names are now stored in t:amiga.names
and the filenames of the standard libraries are already in
t:amiga.ignore.
Unfortunately, all the names in t:amiga.names have a
blank in the first column, and not a hyphen (-) as we would like it.
However, this is no real problem, if you have filter capable of
replacing these blanks by a hyphen installed. Even the common
tr command can do that:
tr " " "-" <t:amiga.names >>t:amiga.ignore
Note that there are two greater-then signs (>) so
the file will be appended. After that, you should have a
t:amiga.ignore which is similar to the one already
included with Vahunz.
If you do not have such a filter, load t:amiga.ignore into your favourite editor, insert t:amiga.names, use the global Search&Replace function and save the result again as t:amiga.ignore.
Sometimes it can be convenient to use all words stored in two separate dignoraries. For example, you could write an application which is both using Posix-compatible and Amiga-specific functions.
As dignoraries entirely consist of simple lines with ASCII-text, you
can easily concatenate them by means of the join command,
for instance:
join amiga.ignore unix.ignore AS vahunz.ignore
A minor problem might be that several names are defined in both dignoraries, and Vahunz will come up with a warning for every duplicate name. This will only happen once, as such names are automatically removed from dictionaries once they are written again by Vahunz.
Therefore you might want to specify the command line option
--no-warning when creating the normal dictionary while your
dignorary already exists.
Of course Vahunz is smart enough to utilize a dignorary like vahunz.ignore even if no dictionary like vahunz.names exists yet. In such a case, it will only create vahunz.names, but vahunz.ignore will still exist afterwards - with duplicate entries removed. This behavior can be practical if you just want to "clean up" a dignorary without doing a whole vahunzation.