summaryrefslogtreecommitdiff
path: root/src/fontloader/misc/fontloader-font-tfm.lua
blob: 6584190ce380d31e04305f3e9ca2a3c9b500df81 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
if not modules then modules = { } end modules ['font-tfm'] = {
    version   = 1.001,
    comment   = "companion to font-ini.mkiv",
    author    = "Hans Hagen, PRAGMA-ADE, Hasselt NL",
    copyright = "PRAGMA ADE / ConTeXt Development Team",
    license   = "see context related readme files"
}

local next, type = next, type
local match, format = string.match, string.format
local concat, sortedhash = table.concat, table.sortedhash

local trace_defining           = false  trackers.register("fonts.defining", function(v) trace_defining = v end)
local trace_features           = false  trackers.register("tfm.features",   function(v) trace_features = v end)

local report_defining          = logs.reporter("fonts","defining")
local report_tfm               = logs.reporter("fonts","tfm loading")

local findbinfile              = resolvers.findbinfile
local setmetatableindex        = table.setmetatableindex

local fonts                    = fonts
local handlers                 = fonts.handlers
local readers                  = fonts.readers
local constructors             = fonts.constructors
local encodings                = fonts.encodings

local tfm                      = constructors.handlers.tfm
tfm.version                    = 1.000
tfm.maxnestingdepth            = 5
tfm.maxnestingsize             = 65536*1024

local otf                      = fonts.handlers.otf
local otfenhancers             = otf.enhancers

local tfmfeatures              = constructors.features.tfm
local registertfmfeature       = tfmfeatures.register

local tfmenhancers             = constructors.enhancers.tfm
local registertfmenhancer      = tfmenhancers.register

constructors.resolvevirtualtoo = false -- wil be set in font-ctx.lua

fonts.formats.tfm              = "type1" -- we need to have at least a value here
fonts.formats.ofm              = "type1" -- we need to have at least a value here

--[[ldx--
<p>The next function encapsulates the standard <l n='tfm'/> loader as
supplied by <l n='luatex'/>.</p>
--ldx]]--

-- this might change: not scaling and then apply features and do scaling in the
-- usual way with dummy descriptions but on the other hand .. we no longer use
-- tfm so why bother

-- ofm directive blocks local path search unless set; btw, in context we
-- don't support ofm files anyway as this format is obsolete

-- we need to deal with nested virtual fonts, but because we load in the
-- frontend we also need to make sure we don't nest too deep (esp when sizes
-- get large)
--
-- (VTITLE Example of a recursion)
-- (MAPFONT D 0 (FONTNAME recurse)(FONTAT D 2))
-- (CHARACTER C A (CHARWD D 1)(CHARHT D 1)(MAP (SETRULE D 1 D 1)))
-- (CHARACTER C B (CHARWD D 2)(CHARHT D 2)(MAP (SETCHAR C A)))
-- (CHARACTER C C (CHARWD D 4)(CHARHT D 4)(MAP (SETCHAR C B)))
--
-- we added the same checks as below to the luatex engine

function tfm.setfeatures(tfmdata,features)
    local okay = constructors.initializefeatures("tfm",tfmdata,features,trace_features,report_tfm)
    if okay then
        return constructors.collectprocessors("tfm",tfmdata,features,trace_features,report_tfm)
    else
        return { } -- will become false
    end
end

local depth = { } -- table.setmetatableindex("number")

-- Normally we just load the tfm data and go on. However there was some demand for
-- loading good old tfm /pfb files where afm files were lacking and even enc files
-- of dubious quality so we now support loading such (often messy) setups too.
--
-- Because such fonts also use (ugly) tweaks achieve some purpose (like swapping
-- accents) we need to delay the unicoding actions till after the features have been
-- applied.
--
-- It must be noted that in ConTeXt we don't expect this to be used at all. Here is
-- example:
--
--   tfm metrics + pfb vector for index + pfb file for shapes
--
--   \font\foo=file:csr10.tfm:reencode=auto;mode=node;liga=yes;kern=yes
--
--   tfm metrics + pfb vector for index + enc file for tfm mapping + pfb file for shapes
--
--   \font\foo=file:csr10.tfm:reencode=csr.enc;mode=node;liga=yes;kern=yes
--
--   tfm metrics + enc file for mapping to tfm + bitmaps shapes
--
--   \font\foo=file:csr10.tfm:reencode=csr.enc;bitmap=yes;mode=node;liga=yes;kern=yes
--
-- One can add features:
--
--  fonts.handlers.otf.addfeature {
--      name = "czechdqcheat",
--      type = "substitution",
--      data = {
--          quotedblright = "csquotedblright",
--      },
--  }
--
-- So "czechdqcheat=yes" is then a valid feature. And yes, it's a cheat.


local function read_from_tfm(specification)
    local filename  = specification.filename
    local size      = specification.size
    depth[filename] = (depth[filename] or 0) + 1
    if trace_defining then
        report_defining("loading tfm file %a at size %s",filename,size)
    end
    local tfmdata = font.read_tfm(filename,size) -- not cached, fast enough
    if tfmdata then

        local features = specification.features and specification.features.normal or { }
        local features = constructors.checkedfeatures("tfm",features)
        specification.features.normal = features

        -- If reencode returns a new table, we assume that we're doing something
        -- special. An 'auto' reencode pickt up its vector from the pfb file.

        local newtfmdata = (depth[filename] == 1) and tfm.reencode(tfmdata,specification)
        if newtfmdata then
             tfmdata = newtfmdata
        end

        local resources     = tfmdata.resources  or { }
        local properties    = tfmdata.properties or { }
        local parameters    = tfmdata.parameters or { }
        local shared        = tfmdata.shared     or { }
        --
        shared.features     = features
        shared.resources    = resources
        --
        properties.name     = tfmdata.name           -- todo: fallback
        properties.fontname = tfmdata.fontname       -- todo: fallback
        properties.psname   = tfmdata.psname         -- todo: fallback
        properties.fullname = tfmdata.fullname       -- todo: fallback
        properties.filename = specification.filename -- todo: fallback
        properties.format   = fonts.formats.tfm      -- better than nothing
        --
        tfmdata.properties  = properties
        tfmdata.resources   = resources
        tfmdata.parameters  = parameters
        tfmdata.shared      = shared
        --
        shared.rawdata      = { resources = resources }
        shared.features     = features
        --
        -- The next branch is only entered when we have a proper encoded file i.e.
        -- unicodes and such. It really nakes no sense to do feature juggling when
        -- we have no names and unicodes.
        --
        if newtfmdata then
            --
            -- Some opentype processing assumes these to be present:
            --
            if not resources.marks then
                resources.marks = { }
            end
            if not resources.sequences then
                resources.sequences = { }
            end
            if not resources.features then
                resources.features = {
                    gsub = { },
                    gpos = { },
                }
            end
            if not tfmdata.changed then
                tfmdata.changed = { }
            end
            if not tfmdata.descriptions then
                tfmdata.descriptions = tfmdata.characters
            end
            --
            -- It might be handy to have this:
            --
            otf.readers.addunicodetable(tfmdata)
            --
            -- We make a pseudo opentype font, e.g. kerns and ligatures etc:
            --
            tfmenhancers.apply(tfmdata,filename)
            --
            -- Now user stuff can kick in.
            --
            constructors.applymanipulators("tfm",tfmdata,features,trace_features,report_tfm)
            --
            -- As that can also mess with names and such, we are now ready for finalizing
            -- the unicode information. This is a different order that for instance type one
            -- (afm) files. First we try to deduce unicodes from already present information.
            --
            otf.readers.unifymissing(tfmdata)
            --
            -- Next we fill in the gaps, based on names from teh agl. Probably not much will
            -- happen here.
            --
            fonts.mappings.addtounicode(tfmdata,filename)
            --
            -- The tounicode data is passed to the backend that constructs the vectors for us.
            --
            tfmdata.tounicode = 1
            local tounicode   = fonts.mappings.tounicode
            for unicode, v in next, tfmdata.characters do
                local u = v.unicode
                if u then
                    v.tounicode = tounicode(u)
                end
            end
            --
            -- However, when we use a bitmap font those vectors can't be constructed because
            -- that information is not carried with those fonts (there is no name info, nor
            -- proper index info, nor unicodes at that end). So, we provide it ourselves.
            --
            if tfmdata.usedbitmap then
                tfm.addtounicode(tfmdata)
            end
        end
        --
        shared.processes    = next(features) and tfm.setfeatures(tfmdata,features) or nil
        --
        parameters.factor        = 1 -- already scaled
        parameters.size          = size
        parameters.slant         = parameters.slant          or parameters[1] or 0
        parameters.space         = parameters.space          or parameters[2] or 0
        parameters.space_stretch = parameters.space_stretch  or parameters[3] or 0
        parameters.space_shrink  = parameters.space_shrink   or parameters[4] or 0
        parameters.x_height      = parameters.x_height       or parameters[5] or 0
        parameters.quad          = parameters.quad           or parameters[6] or 0
        parameters.extra_space   = parameters.extra_space    or parameters[7] or 0
        --
        constructors.enhanceparameters(parameters) -- official copies for us
        --
        if newtfmdata then
            --
            -- We do nothing as we assume flat tfm files. It would become real messy
            -- otherwise and I don't have something for testing on my system anyway.
            --
        elseif constructors.resolvevirtualtoo then
            fonts.loggers.register(tfmdata,file.suffix(filename),specification) -- strange, why here
            local vfname = findbinfile(specification.name, 'ovf')
            if vfname and vfname ~= "" then
                local vfdata = font.read_vf(vfname,size) -- not cached, fast enough
                if vfdata then
                    local chars = tfmdata.characters
                    for k,v in next, vfdata.characters do
                        chars[k].commands = v.commands
                    end
                    properties.virtualized = true
                    tfmdata.fonts = vfdata.fonts
                    tfmdata.type = "virtual" -- else nested calls with cummulative scaling
                    local fontlist = vfdata.fonts
                    local name = file.nameonly(filename)
                    for i=1,#fontlist do
                        local n = fontlist[i].name
                        local s = fontlist[i].size
                        local d = depth[filename]
                        s = constructors.scaled(s,vfdata.designsize)
                        if d > tfm.maxnestingdepth then
                            report_defining("too deeply nested virtual font %a with size %a, max nesting depth %s",n,s,tfm.maxnestingdepth)
                            fontlist[i] = { id = 0 }
                        elseif (d > 1) and (s > tfm.maxnestingsize) then
                            report_defining("virtual font %a exceeds size %s",n,s)
                            fontlist[i] = { id = 0 }
                        else
                            local t, id = fonts.constructors.readanddefine(n,s)
                            fontlist[i] = { id = id }
                        end
                    end
                end
            end
        end
        --
        -- This is for old times sake (and context specific) so we comment it. It has
        -- to do with encoding prefixes (a context naming that was later adopted by
        -- the lm/gyre project)
        --
     -- if not features.encoding then
     --     local encoding, filename = match(properties.filename,"^(.-)%-(.*)$")
     --     if filename and encoding and encodings.known and encodings.known[encoding] then
     --         features.encoding = encoding
     --     end
     -- end
        --
        -- Some afterthoughts:
        --
        properties.haskerns     = true
        properties.hasligatures = true
        resources.unicodes      = { }
        resources.lookuptags    = { }
        --
        depth[filename] = depth[filename] - 1
        --
        return tfmdata
    else
        depth[filename] = depth[filename] - 1
    end
end

local function check_tfm(specification,fullname) -- we could split up like afm/otf
    local foundname = findbinfile(fullname, 'tfm') or ""
    if foundname == "" then
        foundname = findbinfile(fullname, 'ofm') or "" -- not needed in context
    end
    if foundname == "" then
        foundname = fonts.names.getfilename(fullname,"tfm") or ""
    end
    if foundname ~= "" then
        specification.filename = foundname
        specification.format   = "ofm"
        return read_from_tfm(specification)
    elseif trace_defining then
        report_defining("loading tfm with name %a fails",specification.name)
    end
end

readers.check_tfm = check_tfm

function readers.tfm(specification)
    local fullname = specification.filename or ""
    if fullname == "" then
        local forced = specification.forced or ""
        if forced ~= "" then
            fullname = specification.name .. "." .. forced
        else
            fullname = specification.name
        end
    end
    return check_tfm(specification,fullname)
end

readers.ofm = readers.tfm

-- The reencoding acts upon the 'reencode' feature which can have values 'auto' or
-- an enc file. You can also specify a 'pfbfile' feature (but it defaults to the
-- tfm filename) and a 'bitmap' feature. When no enc file is givven (auto) we will
-- get the vectors from the pfb file.

do

    local outfiles = { }

    local tfmcache = table.setmetatableindex(function(t,tfmdata)
        local id = font.define(tfmdata)
        t[tfmdata] = id
        return id
    end)

    local encdone  = table.setmetatableindex("table")

    function tfm.reencode(tfmdata,specification)

        local features = specification.features

        if not features then
            return
        end

        local features = features.normal

        if not features then
            return
        end

        local tfmfile = file.basename(tfmdata.name)
        local encfile = features.reencode -- or features.enc
        local pfbfile = features.pfbfile  -- or features.pfb
        local bitmap  = features.bitmap   -- or features.pk

        if not encfile then
            return
        end

        local pfbfile = outfiles[tfmfile]

        if pfbfile == nil then
            if bitmap then
                pfbfile = false
            elseif type(pfbfile) ~= "string" then
                pfbfile = tfmfile
            end
            if type(pfbfile) == "string" then
                pfbfile = file.addsuffix(pfbfile,"pfb")
             -- pdf.mapline(tfmfile .. "<" .. pfbfile)
                report_tfm("using type1 shapes from %a for %a",pfbfile,tfmfile)
            else
                report_tfm("using bitmap shapes for %a",tfmfile)
                pfbfile = false -- use bitmap
            end
            outfiles[tfmfile] = pfbfile
        end

        local encoding = false
        local vector   = false

        if type(pfbfile) == "string" then
            local pfb = fonts.constructors.handlers.pfb
            if pfb and pfb.loadvector then
                local v, e = pfb.loadvector(pfbfile)
                if v then
                    vector = v
                end
                if e then
                    encoding = e
                end
            end
        end
        if type(encfile) == "string" and encfile ~= "auto" then
            encoding = fonts.encodings.load(file.addsuffix(encfile,"enc"))
            if encoding then
                encoding = encoding.vector
            end
        end
        if not encoding then
            report_tfm("bad encoding for %a, quitting",tfmfile)
            return
        end

        local unicoding  = fonts.encodings.agl and fonts.encodings.agl.unicodes
        local virtualid  = tfmcache[tfmdata]
        local tfmdata    = table.copy(tfmdata) -- good enough for small fonts
        local characters = { }
        local originals  = tfmdata.characters
        local indices    = { }
        local parentfont = { "font", 1 }
        local private    = fonts.constructors.privateoffset
        local reported   = encdone[tfmfile][encfile]

        -- create characters table

        local backmap = vector and table.swapped(vector)
        local done    = { } -- prevent duplicate

        for index, name in sortedhash(encoding) do -- predictable order
            local unicode  = unicoding[name]
            local original = originals[index]
            if original then
                if unicode then
                    original.unicode = unicode
                else
                    unicode = private
                    private = private + 1
                    if not reported then
                        report_tfm("glyph %a in font %a with encoding %a gets unicode %U",name,tfmfile,encfile,unicode)
                    end
                end
                characters[unicode] = original
                indices[index]      = unicode
                original.name       = name -- so one can lookup weird names
                if backmap then
                    original.index     = backmap[name]
                else -- probably bitmap
                    original.commands = { parentfont, { "char", index } }
                    original.oindex   = index
                end
                done[name] = true
            elseif not done[name] then
                report_tfm("bad index %a in font %a with name %a",index,tfmfile,name)
            end
        end

        encdone[tfmfile][encfile] = true

        -- redo kerns and ligatures

        for k, v in next, characters do
            local kerns = v.kerns
            if kerns then
                local t = { }
                for k, v in next, kerns do
                    local i = indices[k]
                    if i then
                        t[i] = v
                    end
                end
                v.kerns = next(t) and t or nil
            end
            local ligatures = v.ligatures
            if ligatures then
                local t = { }
                for k, v in next, ligatures do
                    local i = indices[k]
                    if i then
                        t[i] = v
                        v.char = indices[v.char]
                    end
                end
                v.ligatures = next(t) and t or nil
            end
        end

        -- wrap up

        tfmdata.fonts         = { { id = virtualid } }
        tfmdata.characters    = characters
        tfmdata.fullname      = tfmdata.fullname or tfmdata.name
        tfmdata.psname        = file.nameonly(pfbfile or tfmdata.name)
        tfmdata.filename      = pfbfile
        tfmdata.encodingbytes = 2
        tfmdata.format        = "type1"
        tfmdata.tounicode     = 1
        tfmdata.embedding     = "subset"
        tfmdata.usedbitmap    = bitmap and virtualid

        return tfmdata
    end

end

-- This code adds a ToUnicode vector for bitmap fonts. We don't bother about
-- ranges because we have small fonts. it works ok with acrobat but fails with
-- the other viewers (they get confused by the bitmaps I guess).

do

    local template = [[
/CIDInit /ProcSet findresource begin
  12 dict begin
  begincmap
    /CIDSystemInfo << /Registry (TeX) /Ordering (bitmap-%s) /Supplement 0 >> def
    /CMapName /TeX-bitmap-%s def
    /CMapType 2 def
    1 begincodespacerange
      <00> <FF>
    endcodespacerange
    %s beginbfchar
%s
    endbfchar
  endcmap
CMapName currentdict /CMap defineresource pop end
end
end
]]

    local flushstreamobject = lpdf and lpdf.flushstreamobject
    local setfontattributes = pdf.setfontattributes

    if not flushstreamobject then
        flushstreamobject = function(data)
            return pdf.obj {
                immediate = true,
                type      = "stream",
                string    = data,
            }
        end
    end

    if not setfontattributes then
        setfontattributes = function(id,data)
            print(format("your luatex is too old so no tounicode bitmap font%i",id))
        end
    end

    function tfm.addtounicode(tfmdata)
        local id   = tfmdata.usedbitmap
        local map  = { }
        local char = { } -- no need for range, hardly used
        for k, v in next, tfmdata.characters do
            local index     = v.oindex
            local tounicode = v.tounicode
            if index and tounicode then
                map[index] = tounicode
            end
        end
        for k, v in sortedhash(map) do
            char[#char+1] = format("<%02X> <%s>",k,v)
        end
        char  = concat(char,"\n")
        local stream    = format(template,id,id,#char,char)
        local reference = flushstreamobject(stream,nil,true)
        setfontattributes(id,format("/ToUnicode %i 0 R",reference))
    end

end

-- Now we implement the regular features handlers. We need to convert the
-- tfm specific structures to opentype structures. In basemode they are
-- converted back so that is a bti of a waste but it's fast enough.

do

    local everywhere = { ["*"] = { ["*"] = true } } -- or: { ["*"] = { "*" } }
    local noflags    = { false, false, false, false }

    local function enhance_normalize_features(data)
        local ligatures  = setmetatableindex("table")
        local kerns      = setmetatableindex("table")
        local characters = data.characters
        for u, c in next, characters do
            local l = c.ligatures
            local k = c.kerns
            if l then
                ligatures[u] = l
                for u, v in next, l do
                    l[u] = { ligature = v.char }
                end
                c.ligatures = nil
            end
            if k then
                kerns[u] = k
                for u, v in next, k do
                    k[u] = v -- { v, 0 }
                end
                c.kerns = nil
            end
        end

        for u, l in next, ligatures do
            for k, v in next, l do
                local vl = v.ligature
                local dl = ligatures[vl]
                if dl then
                    for kk, vv in next, dl do
                        v[kk] = vv -- table.copy(vv)
                    end
                end
            end
        end

        local features = {
            gpos = { },
            gsub = { },
        }
        local sequences = {
            -- only filled ones
        }
        if next(ligatures) then
            features.gsub.liga = everywhere
            data.properties.hasligatures = true
            sequences[#sequences+1] = {
                features = {
                    liga = everywhere,
                },
                flags    = noflags,
                name     = "s_s_0",
                nofsteps = 1,
                order    = { "liga" },
                type     = "gsub_ligature",
                steps    = {
                    {
                        coverage = ligatures,
                    },
                },
            }
        end
        if next(kerns) then
            features.gpos.kern = everywhere
            data.properties.haskerns = true
            sequences[#sequences+1] = {
                features = {
                    kern = everywhere,
                },
                flags    = noflags,
                name     = "p_s_0",
                nofsteps = 1,
                order    = { "kern" },
                type     = "gpos_pair",
                steps    = {
                    {
                        format   = "kern",
                        coverage = kerns,
                    },
                },
            }
        end
        data.resources.features  = features
        data.resources.sequences = sequences
        data.shared.resources = data.shared.resources or resources
    end

    registertfmenhancer("normalize features",   enhance_normalize_features)
    registertfmenhancer("check extra features", otfenhancers.enhance)

end

-- As with type one (afm) loading, we just use the opentype ones:

registertfmfeature {
    name         = "mode",
    description  = "mode",
    initializers = {
        base = otf.modeinitializer,
        node = otf.modeinitializer,
    }
}

registertfmfeature {
    name         = "features",
    description  = "features",
    default      = true,
    initializers = {
        base     = otf.basemodeinitializer,
        node     = otf.nodemodeinitializer,
    },
    processors   = {
        node     = otf.featuresprocessor,
    }
}