Go to the previous, next chapter.

Tell me more about NeXTstep fonts

Outline files should be in PFA or hexadecimal ASCII format. The font name should be taken either from the outline (font) file or the AFM file. In both case the name is given after the word ``FontName'' at the beginning of the file)

As a matter of fact, fonts don't strictly HAVE to be in all hexadecimal ASCII format. The eexec portion of the font can be in binary if you wish, and in fact some Mac->NeXTstep or PFB->NeXTstep font converters simply strip off the Mac/PFB storage format headers and leave the binary sections as binary.

However, if you wish to send such a font across a serial channel to a PostScript printer, you will need some way to unpack the binary eexec portion to seven-bit ASCII before sending the font.

Converted Fonts After Conversion

After conversion they are just like any other freeware or shareware font that you can get in NeXTstep-format from the archives. That's just outline and AFM files but no bitmapped screen fonts. So small point size means poor resolution on screen but they most of should print OK if they are any good ( = usually made with Fontographer).

About Conversion Utilities

NeXTstep utilities

unfont
You can find a package, named something like pcATMfont2NeXT.tar.Z, from NeXT archives (cs.orst.edu) that converts PC fonts to NeXT format (PFB -> PFA).
The most useful tool for me has been ``unfont'' which converts the .pfb (binary outline) font to ASCII outline font.
I usually use it like this
```
         $ unfont new_font.pfb >NewFont
```
If the conversion was successful all I have to after that is maybe to rename the font correctly and move the outline file in the correct .font folder.
Opener.app
Opener seems to be a very useful application since it can open several kinds file packages on NeXTstep that are common on other platforms. E.g. ``.sit'', ``.hqx'', ``.zoo'', ``.zip'', ``.z'', etc.
I haven't used it a lot but looks very promising.
T1utils-1.1
This is collection of command-line programs that manipulate PS Type 1 fonts and one of them can also do the PFB->PFA conversion (t1ascii?).

Basic unarchiving of Mac and PC files.

On your Unix machine:

xbin

Converts .hqx to:

.data: Rename and transfer to PC (or use opener.app on NeXT?)
.info: Discard
.rsrc: Discard

unzip Converts .zip to:

.inf: Discard
.wfn: Discard
.pfm: Discard (unless it can generate a better AFM file)
everything else: Transfer to NeXT

On a PC:

xbin

Converts .hqx to:

.data: Rename and transfer to PC (or use opener.app on NeXT?)
.info: Discard
.rsrc: Discard

extract -f ... Converts .cpt to:

file with no extension: This is usually the outline font. Refont and transfer to NeXT.
.afm: Transfer to NeXT.
.pfm: Discard (unless it can generate a better AFM file)
.bma: Discard if you have an AFM file.

unsit30 -eb ... Converts .sit to:

file with no extension: This is usually the outline font. Refont and transfer to NeXT.
.afm: Transfer to NeXT.
.pfm: Discard (unless it can generate a better AFM file)
.bma: Discard if you have an AFM file.

refont Converts outline formats from Mac to NeXT format (PFA).

pkunzip Converts .zip to:

.inf: Discard
.wfn: Discard
.pfm: Discard (unless it can generate a better AFM file)
everything else: Transfer to NeXT

On a NeXT

Opener.app: Converts archive formats (.sit, .hqx, .zip) to NeXT format.
unfont: Converts PFB files to NeXT format.
afm: Converts AFM files to NeXT format AFM files (CR/LF hackery)

Installation

There are scripts (installfont) available that can handle the installation process but here is how you do it manually.

.font
After all that you have to create the .font folder, move the outline and .afm files there and start fighting with the strangely formatted .afm file. The most common problems are font name mismatch between outline and AFM files (family name is incorrect or too long, etc) and missing fields (ex. no ItalicAngle entry) in the AFM file.
buildafmdir AND cacheAFMData
buildafmdir puts its complains to Console but cacheAFMData put them on stdout or stderr (ie. Terminal Window).
PARSE ERRORS ------------ ``Parse error 10000011 ...'' comes from mismatch between of CharMetrics declared in the .afm and actually found. I haven't been able to figure out the other strange parse errors.
buildafmdir in the 3.0 release has the limitation of not being able to install more that 255 fonts in any one font folder. This is supposed to be fixed in 3.1.
The Dreaded > Message
> appears in the font panel when you have run buildafmdir and it finds things it thinks are wrong with the AFM file. Errors can also be generated by parsing routines inside the PasteBoard Services.
> almost NEVER has anything to do with the font itself, since buildafmdir doesn't actually look inside the font. Errors in the font due to faulty conversion will likely show up at the time the PostScript server actually attempts to define the font or render characters from the font.
The only error I have ever seen from a converted font was the results of a naive Macintosh to PFA converter, which didn't understand that the POST resources in a Macintosh format Type 1 font do not have to be in order, nor do the POST resources all have to be contiguous -- they can be interspersed with other resources. The results were that a comment resource ended up in the middle of the eexec section of the font and the PostScript interpreter threw out lots of errors.
> almost ALWAYS occurs because there is something wrong with the AFM file you installed. Here is a partial list of problems that can occur with AFM files:
- Carriage-return characters (^M) at ends of lines.
 This happens when you get incomplete translations from PC files, which use carriage-return-line-feed combinations at ends of lines.
 Solution: edit away the carriage returns. Make sure the lines are terminated only by line-feed characters.
- Spaces or tabs at ends of lines.
 Fixed in NeXTstep 3.1.
- Missing fields.
 NeXTstep DEMANDS that certain fields be present in the AFM file. Required fields are: FontName, FullName, FamilyName, Weight, EncodingScheme, and ItalicAngle. If any of these fields are missing, you will get the > message.
 Solution: fill in the required fields.
- Incorrect Weight field.
 buildafmdir accepts only a certain set of values for the Weight field. Acceptable values are: ``Ultra Light'', ``Thin'', ``Light'', ``Extra Light'', ``Book'', ``Regular'', ``Plain'', ``Roman'', ``Medium'', ``Demi'', ``Demi-Bold'', ``Semi-Bold'', ``Bold'', ``Extra Bold'', ``Heavy'', ``Heavyface'', ``Black'', ``Ultra'', ``UltraBlack'', ``Fat'', ``ExtraBlack'', and ``Obese''.
- Character information count mismatches.
 AFM files contain several sets of information which are introduced by a ``Startxxxxx nnn'' line where the xxxxx is the name of the section (such as StartCharMetrics) and nnn is the purported number of lines of information of this type to follow. Sad to say, many many AFM files supplied by vendors and others are such that the actual number of lines of data do not match the number stated on the Startxxxxx line. When this error occurs in the AFM file, buildafmdir emits a Parse Error message to the console and the font will be marked unusable. The parse error messages from buildafmdir is of the form:
```
 Parse error nnnnnnnn xx yy 
```
 where nnnnnnnn is the error number, xx is the number of lines of information claimed to exist, and yy is the number of lines actually read. The nnnnnnnn are are:
```
 10000011 mismatch in the StartCharMetrics section
 10000012 mismatch in the StartTrackKern section
 10000013 mismatch in the StartKernPairs section
 10000014 mismatch in the StartComposites section
 10000015 mismatch in a composite character specification
```
 I have converted many fonts from the Berkeley Macintosh User Group CD ROM and fully half of the supplied AFM files are incorrect.
Other AFM file errors.
Parse error numbers 10000001 through 10000010 means some kinds of syntax errors in the AFM data file. Any of these errors mean that the AFM file is truly hosed.