path: root/Master/texmf-dist/source/fontinst/base/CHANGES

--- CHANGES ---

This is the old changes log for the fontinst font installation package. 
As of v1.9, changes in the core are recorded using the \changes 
mechanism of the doc package, and appear as a section in the typeset form 
of the source (fisource.dvi). Changes to supporting files are typically 
recorded in each file, at the end.

Feb 4 1993, v0.01: Created the file, wrote an afm and enc parser,
   started on a vpl writer.

Feb 10 1993, v0.02: Rewrote the enc parser, generally hacked around,
   imported trig.tex to convert Adobe's italic slope (expressed in
   angles!) into a font dimension (expressed as a ratio!) grrrrrr...

Feb 12 1993, v0.03: Tidied up the comments a bit, moved the
   \fontlist out of \readafm, added \writepl (as well as \writevpl).
   Replaced \fake{CHARACTER} by having the .fd file give default
   definitions for \char-CHARACTER.  Added the \map and
   \ligtable entries to \char-CHARACTER.  Redefined \ligtable,
   \map and \codingscheme as token lists.  Moved the \kernpair's
   into \ligtable.

Feb 16 1993, v0.04: Added FontName, FullName, UnderlineThickness and
   Notice.  Added the source font encodings, and MAPFONT.
   Discovered a seriousish bug---if a character exists in the afm
   file but is not in the source encoding, then the fake character
   is still overwritten, since we don't know what the source
   encoding is going to be until later.  Fixed this by reading in
   the source encoding first, and looking to see if there is a
   character in the source encoding for the one in the afm file.
   Added \topaccent.

17 Feb 1993, v0.05: Added the visual centre parameter to \topaccent,
   and added \botaccent and \doublechar.  Fixed a bug where the
   default \missingchar had a horizontal rather than a vertical
   rule.  Added \settoken.

19 Feb 1993, v0.06: Fixed a bug whereby a character which was in the
   source encoding but not in the afm file overwrote it's faked
   version.  Changed the default \missingchar to be independent of
   the values parsed from the afm file (so it doesn't matter when
   you set a \missingchar).  Added \parameter and \ligatures.  Added
   \kernalikes and \copykern, which meant I had to split \ligtable
   back into \ligtable and \kerntable, since when you say \copykern
   you only want the \kerntable copied, not the \ligtable.
   Redefined \map so that it's more portable---the contents of \map
   are allowed to assume they start in font \startfont, should move
   and should finish having moved right by \charwd and in font
   \stopfont.  This makes it possible to have composite characters
   within composite characters---for example to make <ffi> out of
   <f><fi> and <fi> is in turn made out of <f><i>.

23 Feb 1993, v0.07: Introduced the \character token list (previously
   all of the variables had been set by \char-FOO).  \botaccent,
   \topaccent, etc are now commands to be used inside \defchar, so
   you can say something like \defchar{FOO}{\ifblah ... \botaccent
   ... \else ... \topaccent ... \fi}.  Added \cramchar.  Added
   the shift-left argument to \doublechar.

24 Feb 1993, v0.08: Replaced \charic by \leftbearing and
   \rightbearing.  Made the assingments in a \char-FOO command
   global, so we can protect local variables.  Fixed a bug with
   \getchar{FOO} which used to end up calling \the\character.
   However, \the\character may contain other \getchar{BAZ} calls,
   which will override the values of \kernalike, \source, \target,
   etc.  So \getchar can't call \the\character.  Added \shrinkfont.
   Made all the AFM variables counts rather than dimens.  Added
   \afmunits, and did all the conversion between AFM units and TeX
   units as late as possible (which means we can alter \afmunits
   locally and perform scaling).  Added \getfont so fonts are now
   loaded dynamically, and can be scaled.  Added \shrinkchar for
   producing c&sc fonts.

3 Mar 1993, v0.09: Added \fractionchar, \inferiorchar and
   \superiorchar.  No character is produced for characters called
   .notdef in the target encoding.  Produced a first bash at a
   complement encoding to Cork, containing the Adobe standard and
   Expert characters missing from Cork.

9 Mar 1993, v0.10: A major spring clean.  Introduced ATX and ETX
   files, so the parser for AFM files now reads an AFM file and writes
   out the same information in a TeX-friendly fashion as an ATX file.
   Similarly, an ETX file contains the same information as an ENC file.
   Introduced \inputatx, \inputetx and \inputenc to input an ATX, ETX
   or ENC file if it already exists and to make it otherwise.

14 Mar 1993, v0.11: A continuation of the spring clean.  Redefined
   \getchars to define everything locally.  Redefined \writechars to
   read the encoding from a .etx file.  Discovered I'd used the command
   \afmchar twice(!) so I redefined one of them to be \rawchar.
   Rewrote the LIGTABLE section.

15 Mar 1993, v0.12: Finished the spring clean of the file generation.
   Redid \writevpl and \writepl.  Read the FUD file in after the ATX
   files, so the FUD file can use the parameters read in by the ATX
   file.  Fixed a bug with \afmchar that wrote incorrect values
   to \charactersize in the ATX file.  Rewrote \fontdimens to write the
   PARAMETERs out immediately (since we are now writing the VPL file
   when reading the FUD file).  Finished the LIGTABLE section, redid
   \copykern.

22 Mar 1993, v0.13: Rewrote the generation of ATX files, so kern table
   information is gathered together as in a LIGTABLE.  Renamed
   \defchars to \readafm.  Renamed \writevpl to \makevpl ... \endvpl,
   and rewrote it to cope with the new syntax of ATX files.

24 Mar 1993, v0.14: Added \lkern and \rkern entries to ATX files,
   replacing \copykern.  Added kern copying to LIGTABLEs.

29 Mar 1993, v0.15: Added \boundarychar, \nextlarger and \varchar.
   Fixed a bug with \lrkern.

30 Mar 1993, v0.16: Fixed a bug whereby FONTDIMENS was being output
   rather than FONTDIMEN.  Also \writepl was generating a VPL file.
   And I discovered a bug with dvips that means that fonts generated
   with designsize and designunits R 9.9945 cause a rounding error
   and dvips complains about a design size mismatch.  Grr...  Fixed
   a bug with outputting ligatures for characters without
   kernings---previously I output a LABEL without checking that
   there were actually going to be any LIG instructions (which there
   might not be if, for example, it was a monoweight font).

4 Apr 1993, v0.17: Added the interactive prompt.  Generated the .fd
   file.  \makevpl and \makepl don't overwrite files any
   longer---they give a message saying `FILENAME.(v)pl already
   exists.'

17 Apr 1993, v0.18: Added \makefamily, \maketexfont and \makerawfont.

27 Apr 1993, v0.19: Fixed a bug with \finishligs, which caused
   multiple LABEL instructions to be put down (I had a \labelfalse
   where I should have had a \labeltrue).

4 Aug 1993, v1.00: Ripped it up and started again... v1.00 is going
   to be a fairly complete rewrite from the ground up.  I can keep:

    * the AFM parser
    * the ENC parser
    * much of the VPL and PL writer

   but the fud engine is going to be replaced by the notion of a
   glyphbase.  The idea will be that all of the information kept in
   the ATX and FUD files will be put into one file called a
   `glyphbase'.  Then we'll have two utilities:

    * \makerawfont{ENCODING}{AFMFILE}{FONTNAME}
      will convert an AFM file into a glyphbase and a property list.

    * \maketexfont{ENCODING}{GLYPHBASES}{FONTNAME}
      will convert a collection of glyphbases into a virtual font.

   For example:

      \makerawfont{adobe}{Times-Roman}{ptmr0}
      \maketexfont{T1ulc}{ptmr0,cork}{ptmrq}

   makes the raw font ptmr0.pl, the glyphbase ptmr0.gly, and then
   converts the glyphbases ptmr0 and cork into the font ptmrq.  This
   notion of glyphbase will also allow users to modify fonts easily,
   for example you could write a little ptmhax glyphbase, which
   would override the default cork glyphbase by saying:

      \maketexfont{T1ulc}{ptmr0,ptmhax,cork}{ptmrq}

   This will make it much easier to tweak fonts.  I'll also specify
   the glyphbase syntax, and do my best to make everything upwardly
   compatible.

   On the 4th of August I started on the above.  I wrote
   \makeglyphbase, which converts an AFM file into a glyphbase, and
   I wrote the glyphbase commands to store the glyphbase information
   in TeX.  The main thing that's missing is the PL and VPL writers,
   and the ligature, nextlarger and varchar entries.

10 Aug 1993, v1.01: Edited trig.tex to remove uses of \then.  Separated
   \maketexfont into \inputglyphbase and \makevirtualfont.  Added the
   ability to write CHARACTER instructions, (OK, but not the MAPFONTs
   yet).

11 Aug 1993, v1.02: Only output CHARDP, CHARHT and CHARIC when they
   are positive.  Added the MAPFONT and \mapfont instructions.

12 Aug 1993, v1.03: Corrected a bug where I was calling \inputetx 
   *after* an \openout\outfile, which caused the \outfile to be closed, 
   if the etx file had to be generated.

22 Aug 1993, v1.04: Added the concept of variable and number expression.  
   Included trig.tex as part of this file.  Changed design size to an 
   integer.  Renamed `glyphbase' to `metric file', and changed the 
   syntax of ETX files.  Only write out SELECTFONT instructions when 
   necessary.

23 Aug 1993, v1.05: Added the font dimensions.

24 Aug 1993, v1.06: Made the slot of each glyph an int variable like 
   any other.  Added dim variables.  Made designsize a dim variable, 
   which should cure a bug that was causing design sizes not to match 
   up.  Added the LIGTABLE.  Fixed a bug with \inputetx and \inputmtx 
   which caused the redefinition of \relax to survive past the 
   \inputFOO command.  Fixed a bug with \vplrule, since I'd got the 
   arguments to SETRULE the wrong way round (again!).

25 Aug 1993, v1.07: Added the VARCHAR and NEXTLARGER entries.  Added 
   the encoding documentation commands.  Renamed the file to 
   fontinst.sty. 

26 Aug 1993, v1.08: Added the documentation of MTX files.

27 Aug 1993, v1.09: Corrected a bug with \comment.  Added \tfmtomtx.  
   Removed the SCALE argument to \setrawglyph.  Added \convertafm to 
   convert integer AFM units to a dimen, trying to minimize rounding 
   errors whilst avoiding rounding errors.  Renamed \leftkerning and 
   \rightkerning to \setleftkerning and \setrightkerning.

30 Aug 1993, v1.10: Added \resetwidth, \resetheight, \resetdepth and 
   \resetitalic.  Added the ability to use \width{GLYPH} etc. inside 
   \setglyph{GLYPH}.  Added everyglyph and everyendglyph.

10 Sep 1993, v1.11: Discovered a bug with the VPL glyphs which was
   causing it never to write out a SELECTFONT instruction.

13 Sep 1993, v1.12: Added the \familyinst...\endfamilyinst commands.
   Added commands to zero the fonts between calls of \makefont, so
   one document can make a large number of fonts.

14 Sep 1993, v1.13: FIRST ALPHA-RELEASE OF V1.
   Changed the syntax of \useafm and \useetx to use
   the same parameters as LaTeX NFSS.  Removed the prefix f0 from
   the beginning of the font names, in a pathetic attempt to fit 
   MS-DOS 8+3 filenames.  Added the \tempfile command.

4 Oct 1993, v1.14: Removed a spurious \let in \docnextslot.  Thanks
   to Damian Cugley for spotting that one.  Added \usetfm to \familyinst. 

5 Oct 1993, v1.15: Fixed a bug with \dotfm, which was causing it to
   generate a PL file, even though the font was already in TFM
   format!  Thanks again to Damian.  Also changed
   \let\Input=\@@input to a catcode hack, just in case fontinst.sty
   is read in by LaTeX when @ isn't a letter.  That one was spotted
   by Paul Thompson.

6 Oct 1993, v1.16: Replaced Phil Taylor's trig macros with David
   Carlisle's.  Added the PL parser.  Fixed a bug with the bug fix
   for \dotfm, where I'd removed an \inputmtx command that was
   needed...  And I fixed a bug which was causing the VPL files to
   be generated once for each font that contributed to the VPL file.
   Damian got those ones as well.

11 Oct 1993, v1.17: Added the SIZE argument to \useafm and \usepl,
   and redid the FD generation to use the size info.  This should
   allow fonts like Malvern to be generated at different design
   sizes and still have a sensible FD file.  NOTE: This means old
   fontinst files will no longer work!  I also made \pltomtx create
   a \setdim{designsize} command.

13 Oct 1993, v1.18: Fixed a bug with \calculateitslant which caused
   it to bomb if the italic angle didn't have a decimal... this was
   caused by allowing extraneous spaces at the end of the angle.
   These have now been stripped away.  Thanks to Paul Thompson for
   spotting that one!

15 Oct 1993, v1.19: Discovered a bug with \familyinst, which caused
   only one ETX file to be loaded.  This one was introduced when I
   added the SIZE arguments, since I reorganized the main
   \endfamilyinst loop.  Serves me right!  This is now fixed, but
   I'm left with the problem that the sensible order to generate VPL
   files is not the sensible order to generate FD lines in.  FD
   files need the SIZE parameter inside the SERIES and SHAPE,
   whereas if a font is contributing to a number of SERIES and
   SHAPEs but only one SIZE, then it's sensible to have it outside
   the SERIES and SHAPE parameters.  Oh dear...

19 Oct 1993, v1.20: Added the glyph `missingglyph' which is produced
   whenever a glyph is missing.

3 Nov 1993, v1.21: Corrected a bug with \kerning, which was causing
   \kerning{A}{V} to give the kerning between V and A!  And I added
   the PL names for the math font parameters NUM1--AXISHEIGHT.
   Damian spotted those ones.  Rewrote the main loop as I said I
   would have to in the note on v1.19, so there's no useless
   regeneration of fonts now.

17 Nov 1993, v1.22: Restored the final parameter to
   \DeclareFontShape which had disappeared somehow!  Rob Hutchings
   spotted this one.  And I stopped trying to be clever about
   checking to see if a glyph has been defined before using it.
   This means you get a horrible error message if you try to use an
   undeclared glyph, but hey ho.  Trying to trap this has caused so
   many problems with other things breaking that I think I'll give
   up :-)  Added the \glyphwarning command, and removed the
   `missingglyph' glyph.  Docstripped David's trig.doc into
   trig.sty, so I don't have to worry about trig.doc generating the
   trig.sty file.  Saved the catcodes of ` ', `@', `~' `_' and NL so I
   so I can use @ and _ for internal macro names, space and NL
   for ignored white space, and ~ for `real' white space.  Renamed
   all the internal commands (apart from the documentation ones) to
   have _ in their names.

29 Nov 1993, v1.23: Added a check that a glyph exists before writing
   it to a (V)PL file... this was causing fontinst to crash if there
   was a glyph missing from an AFM file.  Yannis Haralambous and Paul 
   Thompson spotted this one.

23 Dec 1993, v1.24: Removed the \familyinst...\endfamilyinst interface, 
   and replaced it by \installfonts ... \endinstallfonts, which makes 
   it easier to generate FD files that look like standard LaTeX2e FD 
   files.  Removed (I'm not sure if this is a great idea) the ability 
   to specify an encoding for an AFM file, since I'm not sure every 
   dvi-to-ps converter can re-encode raw fonts.  I added the 
   \declarefamily command to specify which encoding string corresponds 
   to which ETX file.

24 Dec 1993, v1.25: Redid the \declare... commands, and replaced 
   everyglyh and everyendglyph by sidebearings.

25 Dec 1993, v1.26: Drank a lot.

26 Dec 1993, v1.27: Added \setcommand, \resetcommand and \ifiscommand.

27 Dec 1993, v1.28: Bug-fixes in \setcommand and \resetcommand.  Redid 
   \pltomtx so that it can read the encoding from the CODINGSCHEME line 
   in the file.  Moved the documentation and tables into fontdoc and 
   fonttab.tex.

28 Dec 1993, v1.29: Removed the fd-commands argument from \makefont, 
   since the FD syntax is to allow FD-commands dependent on
   encoding/family/series/shape but not on size.  This means we can't 
   generate font-specific FD-commands, but I've never seen those used, 
   so I'm not too worried...  Fixed a bug with the CC commands, which 
   was producing a glyph with a different width than its horizontal 
   escapement, and I'm not sure what every DVI driver would do with 
   that.  Replaced \etxtopl with \mtxtopl.  Replaced the GLYPH COMMAND 
   \rawglyph with the MTX COMMAND \setrawglyph.

29 Dec 1993, v1.30: Fixed a bug with \setrawglyph, which wasn't 
   checking to see if the glyph was already defined.  Added 
   \unsetglyph.  Corrected the spelling of `\immedaite' (sic) in 
   \vpl_nextlarger.

3 Jan 1994, v1.301: The glyph warning message is now only produced when 
   the glyph is used, rather than when it is defined.  And I've moved 
   over to three-digit version numbers, since I was going through the 
   two-digit ones quite rapidly.  Added \bye, so fontinst documents can 
   run under LaTeX as well as plain TeX.  Added \unsetint, \unsetstr, 
   \unsetdim and \unsetcommand.  Renamed \makefont and \makefamily to 
   \installfont and \installfamily.

4 Jan 1994, v1.302: Fixed a bug with the definition of \bye.

6 Jan 1994, v1.303: Added the FONTNAME argument to \installfont.
   Actually got rid of everyglyph.

8 Jan 1994, v1.304: Added minimumkern and \resetglyph.  Fixed a bug 
   with \vpl_kerning, which was swapping the left- and right-kerns.  
   Surprising I never spotted this before...

9 Jan 1994, v1.305: Worked on the MTX and ETX files.

10 Jan 1994, v1.306: Worked on the MTX and ETX files.

12 Jan 1994, v1.307: Edited the documentation.

20 Jan 1994, v1.308: Added the rawscale integer variable, and the 
  \setnotglyph command.  Redid the way that CC instructions are 
  processed, to use the dimensions stored in the glyph-not variables.  
  Set all of the saved VPL commands to \relax so that they can be 
  \edef'ed on without having to protect them.  Replaced some \def\foo{} 
  by \let\foo\empty_command's.  Replaced a large definition that was 
  being made by \CHARACTER (when parsing PL files) by a \let.  Redid 
  the datatype used to store kerning so it uses control sequences as a 
  compact representation of glyph names.  All in all, fontinst now runs 
  in 79% of the time and 77% of the space that it did last week.

21 Jan 1994, v1.309: Saved some more time and space on the kern table.

22 Jan 1994, v1.310: Added the `FONT scaled AMOUNT', made \setkern take 
   account of \int{rawscale}.  Moved the change log into a separate 
   CHANGES file.  Redid \samesize and \resetwidth so that the width is 
   adjusted by a MOVERIGHT, so the glyph will have the correct horizontal 
   escapement.  Corrected a typo which had renamed \resetint to \setint 
   (oops!).  Corrected a typo of `itslant' rather than `italicslant'.  
   Added the \transformfont command and \mtxtomtx.  Changed the 
   definition of \unfakable{foo} in latin.mtx to look to see if there's a 
   glyph <foo-not> to get the dimensions for.  Added a fake <space> and a 
   fake <dotlessj> which is a missing glyph rule with the dimensions 
   combined from <j> and <dotlessi>.  Changed the default interword 
   stretch to be 1/3 of interword space, rather than zero.  Removed the 
   checks for the existence of <space> and <x> in OT1.etx and T1.etx, 
   since these are now faked in latin.mtx.  Renamed the PS/CM math fonts, 
   after a discussion with Karl Berry, added the OMS and OMX math fonts.  
   Added the Greek letters, digits and other humanists to unsetalf.mtx.  
   Added the DEFAULTRULETHICKNESS and BIGOPSPACING commands to the PL 
   parser.  Fixed a bug I'd introduced by giving \vartop, \varmid, 
   \varrep and \varbot no parameters.  Oops.  Created timesmat.sty and 
   testmath.tex.  All in all, a reasonable day's work...
   
23 Jan 1994, v1.311: Added some more tests to testmath.tex.  Added the 
   new math sizes to timesmat.sty.  Fixed a bug with fontstnd.tex which 
   was causing it to generate the wrong FD entry for italic fonts.  Fixed 
   a bug with CHARACTER parsing in \pltomtx which wasn't resetting the 
   values of WIDTH, HEIGHT, DEPTH and CHARIC between glyphs.
   
24 Jan 1994, v1.312: Added the SEVENBITSSAFE command to the PL parser.

25 Jan 1994, v1.313: Restored the catcode of ~, which I'd left as 10.
   Made /, < and > in zptmcmrm come from cmmi rather than ptmri.  Added 
   zptmcmr to fontptcm.tex, which is Adobe Times with +, =, (, ), [ and ] 
   taken from CMR.  This means that [...] and (...) in text is different 
   from in math, but means that the delimiters are the right size for 
   math setting with CM symbols, and match the growing delimiters in cmex. 
   Moved the files ptmcmhax.mtx and psycmhax.mtx to zrmhax.mtx and 
   zrvhax.mtx.

27 Jan 1994, v1.314: Took the textstyle and displaystyle big operators 
   in zpsycmrv from cmex8 rather than cmex10, since they are more 
   suitably sized for the <summation> and <product> from Adobe Symbol.  
   Made all the accents in latin.mtx zero depth (which TeX's math accent 
   placement requires).  Discovered that cmex8 doesn't have a 
   CODINGSCHEME, so I added one by hand to the PL file.

30 Jan 1994, v1.315: Replaced cmex8 in zpsycmrv by cmex9.  Added
   <epsilon1>, <pi1> and <rho1> to unsethum.mtx.  Added the \skewchar
   declaration for zptmcmrm.  Fixed a bug which was causing negative
   mathchardefs.  Finished off the skewchar kerns for zptcmrm.
   Renamed timesmat.sty to mathptm.sty, since there's already a
   timesmat.sty. 

7 Feb 1994, v1.316: Fixed the \ProvidesPackage line in mathptm.sty.

8 Feb 1994, v1.317: Fixed a bug with fontptcm.tex where I'd put the
   \skewchar command in the wrong argument!  Made the `size' argument
   to \installfont an FD size specification.  Added fakes for <lslash>
   and <Lslash> built from <lslashslash> to latin.mtx.  Added a check
   that a ligature glyph exists before writing out the LIGTABLE entry.
   Edited the documentation a bit---thanks to Karl Berry for comments.
   Added the fontcmro example, inspired by Norman Walsh.  Fixed a bug
   with .fd file generation which caused multiple \DeclareFontShape
   entries to be produced.  Added \unfakable{endash} and
   \unfakable{emdash} to latin.mtx.  Sebastian spotted that one.

10 Feb 1994, v1.318: Added \unfakable{dotlessi} to latin.mtx.  Karl
   Berry pointed that one out.

1 Mar 1994, v1.319: Added the HEADER.  Karl Berry pointed that one out
   too. 

6 Apr 1994, v1.320: Write out a LABEL BOUNDARYCHAR ligable entry where
   necessary.  Alan Hoenig pointed that one out.  Rewrote mathptm.sty
   to remove some of the nastier hacks.

7 Apr 1994, v1.321: Added the ability to generate font substitutions
   in the FD files.

8 Apr 1994, v1.322: Fixed a bug with font substitution which was
   causing the font substitutions to be written out before the real
   fonts.  Moved the mathptm package to standing on its own.

17 Apr 1994, v1.323: Added the top-level \latinfamily commands, and
   updated the example files and documentation to use them.  Added
   OT1ci.etx.  Made the list of files in an \installfont command
   expanded before running along the list, so you can have a macro
   which expands out to a list of files now.  Added substitutions for
   semibold and demibold, for upright italic, and for sc to nc.  Added
   OT19.etx, OT1c9.etx, OT1ci9.etx, OT1i9.etx, T19.etx and T1c9.etx.
   Removed OT1o.etx.

19 Apr 1994, v1.324: Removed italic and oblique small caps.  Fixed a
   bug which was causing the latin.mtx file to be read in inside a
   group if a latin family without an expert font was created.

25 Apr 1994, v1.325: Removed a bug which caused sc fonts to have the
   same names as u&lc fonts.  Fixed a bug which was causing oblique
   fonts not to be generated properly.  Fixed a bug with FD
   generation which was producing too few arguments for
   DeclareFontShape for substitutions.

27 May 1994, v1.326: Made fontinst.tex a 2e document.  Fixed a bug
   with generating slant which was generating slant in design size per
   point rather than points per point.  Sigh.  Ciar\'an \'O Duibh\'in
   spotted that one.  Declared this to be the first full release.

1 Jun 1994, v1.327: Fixed bug with filename parsing which was losing
   variants.  Rewrote the encoding files to use the OT1 tt encoding
   whenever monoweight is set.  Removed the f-liagatures from the T1
   encoding for monowidth fonts.  Sebastian spotted that one.

2 Jun 1994, v1.328: Fixed a bug with FD generation: I'd added the
   extra argument for series substitutions but not shape
   substitutions.  Ta Sebastian again.

3 Jun 1994, v1.329: Made Upsilon1 unfakable.

15 Jun 1994, v1.330: Fixed a bug with font substitution which was
   ignoring chains of substitutions such as bx -> b -> db.  Added the
   m->l substitution.  Renamed the `mn' and `bn' series to `mc' and
   'bc'.  Added the  `k' and `m' series.

16 Jun 1994, v1.331: Sebastian spotted another problem with
   substitutions.  If a font had both a series and a shape
   substitution, the series was sometimes getting lost.
   Steve Grathwohl pointed out that I'd forgotten to switch off
   kerning before loading the math italic fonts in fontcmro.

27 Jun 1994, v1.332: Added braces to \needsfontinstversion, which was
   clobbering \a_dimen and \b_dimen.  Fixed a bug with compwordmark
   which was causing it to get letterspacing added to it.  Added the
   GLYPH-spacing variable to cope with this.  These were
   spotted by Constantin Kahn.

29 Jun 1994, v1.333: Added \ProvidesFile line to the .fd files.

16 Aug 1994, v1.334: Added \frompl, fixed bugs with \resetdim and
   <ccedilla>---all reported by Constantin Kahn.

9 Sep 1994, v1.335: Fixed a bug with NEXTLARGER and VARCHAR generation
   which was causing the entries to be printed out for nonexistent
   glyphs---spotted by Alan Hoenig.  Updated the example installation
   files for Times, Helvetica and Courier.  Updated to a more recent
   trig.sty, and added the hacks to allow trig.sty to be used outside
   LaTeX2e.   Made fontinst.sty initex processable, so you can save
   memory by building your own fontinst format without plain.  Fixed
   the names of `narrow' fonts, which used to clobber the non-narrow
   fonts---spotted by Peter Busk Laursen.

12 Sep 1994, v1.336: Fixed a bug with fontcmro, which was listing the
   9pt font as being 10pt in the fd file---spotted by Steve Grathwohl.

22 Sep 1994, v1.337: Corrected spelling mistake in \declareencoding in
   error message of \CODINGSCHEME---spotted by Alan Hoenig.

8 Oct 1994, v1.338: Corrected definition of \space in
   fontinst.ini---spotted by Rob Hutchings.

18 Nov 1994, v1.339: Added 8r.etx and 8r.enc from Sebastian.
   Distributed a patched fontinst.rc file containing a reimplementation
   of \latinfamily that uses the 8r encoding.

21 Nov 1994, v1.400: Added missing \fi to 8r.etx---spotted by Steve
   Grahthwohl.  Added \mtxtopl for expert fonts in
   \latinfamily---spotted by Rob Hutchings.

29 Nov 1994, v1.401: Added hyphen+hyphenchar = hyphenchar lig to
   T1.etx (so that foo-bar words can be hyphenated as foo<hyphenchar>
   bar)---spotted by Frank Mittelbach.  Made default stretch and
   shrink 50% and 16.7% of <space> respectively.  Added \everyjob to
   fontinst.ini to load the fontinst.rc file---spotted by Rob
   Hutchings.

7 Dec 1994, v1.402: Corrected typos with 8r.etx---spotted by
   Constantin Kahn.

12 May 1995, v1.403: Mentioned the pl files in fontinst.tex, and added
   the encoding TEX TEXT WITHOUT F-LIGATURES.  Both spotted by John
   Wells. 

16 Sep 1995, v1.500: Incorporated changes by SPQR and CK in their
   fontinst.rc file.   In detail...  Added the \NOFILES debugging
   option.  Used SPQR and CK's complete reimplementation of
   \latinfamily.  Added the TEXBASE1ENCODING.  Added \etxtopl and
   \installrawfont.  Added the setting of \errorcontextlines.  Grabbed
   T1i.etx from CTAN.  Added the setting of \newlinechar (pointed out
   by Constantin).  Fixed bug with \resetcommand which stopped braces
   around the csname from working (Contantin again).  Removed the
   contrib/zd directory at Constantin's request.  Rewrote the
   documentation.  Added fontinst.cls.

[16 Sep 1995, v1.500 on CTAN with fonts/psfonts/tools]

21 Sep 1995, v1.501: Corrected some typos in the documentation.
   Added kerns for ng, Ng and hyphenchar.   Fixed spacing in faked
   <FFsmall> and friends.  Fixed kerning of <dotlessJ>.  Raised the
   <macron> in <dbar> to half-way between the xheight and the height of
   <d>.  Added fake <ng> -> <eng>.  Made \cscbotaccent and
   \csctopaccent look to see if both of the glyphs exist rather than just
   the first.  Thanks to Hilmar Schlegel for all these points.   

29 Sep 1995, v1.502: Renamed all the distributed afm files from
   *0*.afm to *8a*.afm.  Updated the file header on 8r.etx.  
   Renamed Karl's fontinst.rc to karl-fontinst.rc, to avoid it being
   picked up by every punter.
   All pointed out by Sebastian.

24 Nov 1995, v1.503: Included Alan Hoenig's vfinst in contrib/vfinst.

23 Jan 1996, v1.504: Included some of Sebastian's changes to OT1.etx: 
   upped stretch and shrink to .6 and .24 of space; reduced extra
   space after . to .24 of space; removed <endash> + <hyphen> =
   <emdash> ligature from OT1 monowidth; made \setslot{grave}
   etc. into \setslot{\lc{Grave}{grave}} etc; fixed a bug with T1c
   `monoweight' -> `monowidth' (!); implemented Rob's changes to 
   small caps and small accented caps; included fake kerning for
   Udieresis; added hackery from spqr.rc for faking small caps and
   faking widths.

[28 Jan 1996, v1.504 on CTAN with:fonts/psfonts/tools]

24 Jan 1996, v1.505: Updated fontptcm.tex to use ptmr8a rather than ptmr0.

25 Jan 1996, v1.506: Added extra braces to \latinfamily to stop
   setting of \digits in one run from affecting it in subsequent
   runs.   Spotted by Richard Walker.

23 Feb 1996, v1.507: Replaced vpltovf by vptovf.  Spotted by Sebastian.

17 Jun 1996, v1.508: Added an \ifdim check to stop lots of MOVERIGHT R 0.0 
   instructions being generated.  Spotted by Sebastian.

18 Jun 1996, v1.509: Allow real numbers in .afm files.  Spotted by
   Gintautas Grigelionis.  Made spaces before ; optional, in order to
   parse v2.0 afm files (spaces before ; were made compulsory in v3.0).

20 Jun 1996, v1.510: Added 9d and 9o encodings for fonts with
   old-style digits. 

21 Jun 1996, v1.511: Removed \endinput from the end of .fd files, so
   that additions can be made more easily.

24 Jul 1996, v1.512: Updated version of trig.sty.

7 Aug 1996, v1.513: Moved unfakable glyphs eariler in latin.mtx so
   that all-caps fonts will work.

6 Feb 1997, v1.514: Added a zero-width rule to cwm in latin.mtx (to
   stop the MAP entry from being empty).  Made cwm 1ex high.

--- Copyright 1993, 1994, 1995, 1996 Alan Jeffrey ---
 
11 Feb 1997, unofficial v1.6 SPQR on CTAN:fonts/psfonts/tools
   Derived from v1.504 with minor hacks and fixes (not documented).  

29 May 1998, unofficial v1.7 SPQR on CTAN:fonts/psfonts/tools
   Derived from v1.6 with fixes by UV.  Corrected mapping of weights.
   Added code for 8r-encoded AFM files (intended to cope with ttf2afm
   limitations, but taken out again after ttf2afm was updated).
   
22 Jun 1998, unofficial v1.8 UV
   Derived from v1.7 and v1.511.  Many .etx/.mtx files touched up.
   fontinst.sty and fontdoc.sty now generated from fontinst.dtx.
   \textcompfamily integrated into \latinfamily.  Reintroduced
   oldstyle encodings (from v1.511).  8y.etx/8y.mtx added as an
   alternative to 8r.etx/8r.mtx. ts1.etx and textcomp.mtx revised.

30 Jun 1998, unofficial v1.800 UV on CTAN:fonts/utilities/fontinst
   Several changes to examples/mathptm and examples/mathptmx.

07 Jun 1998, v1.801: Fixed a bug when oldstyle small-caps fonts
   are generated from <font>8r + <font>8x instead of <font>c8r.