diff options
author | Philipp Gesang <phg@phi-gamma.net> | 2016-06-15 00:02:42 +0200 |
---|---|---|
committer | GitHub <noreply@github.com> | 2016-06-15 00:02:42 +0200 |
commit | 36cc5c9c567e24916f254203fc362bf124e26d02 (patch) | |
tree | 56cdd0a401ffbb99e8702f47a7865677b0971a8e /src/fontloader/misc/fontloader-font-map.lua | |
parent | 17fbf1d1c26047f1e0e80fc6e5f3331f6183a795 (diff) | |
parent | ba744a4bce3ed03eefbf2b4746fa24e6d388d9ff (diff) | |
download | luaotfload-36cc5c9c567e24916f254203fc362bf124e26d02.tar.gz |
Merge pull request #364 from phi-gamma/master
fixes, 3rd edition
Diffstat (limited to 'src/fontloader/misc/fontloader-font-map.lua')
-rw-r--r-- | src/fontloader/misc/fontloader-font-map.lua | 233 |
1 files changed, 116 insertions, 117 deletions
diff --git a/src/fontloader/misc/fontloader-font-map.lua b/src/fontloader/misc/fontloader-font-map.lua index 509e751..6151b37 100644 --- a/src/fontloader/misc/fontloader-font-map.lua +++ b/src/fontloader/misc/fontloader-font-map.lua @@ -10,7 +10,6 @@ local tonumber, next, type = tonumber, next, type local match, format, find, concat, gsub, lower = string.match, string.format, string.find, table.concat, string.gsub, string.lower local P, R, S, C, Ct, Cc, lpegmatch = lpeg.P, lpeg.R, lpeg.S, lpeg.C, lpeg.Ct, lpeg.Cc, lpeg.match -local utfbyte = utf.byte local floor = math.floor local formatters = string.formatters @@ -45,7 +44,7 @@ of obsolete. Some code may move to runtime or auxiliary modules.</p> -- end -- end -local hex = R("AF","09") +local hex = R("AF","af","09") ----- hexfour = (hex*hex*hex*hex) / function(s) return tonumber(s,16) end ----- hexsix = (hex*hex*hex*hex*hex*hex) / function(s) return tonumber(s,16) end local hexfour = (hex*hex*hex^-2) / function(s) return tonumber(s,16) end @@ -140,7 +139,7 @@ local f_double = formatters["%04X%04X"] -- end -- end -local function tounicode16(unicode,name) +local function tounicode16(unicode) if unicode < 0xD7FF or (unicode > 0xDFFF and unicode <= 0xFFFF) then return f_single(unicode) else @@ -149,7 +148,7 @@ local function tounicode16(unicode,name) end end -local function tounicode16sequence(unicodes,name) +local function tounicode16sequence(unicodes) local t = { } for l=1,#unicodes do local u = unicodes[l] @@ -295,145 +294,145 @@ function mappings.addtounicode(data,filename,checklookups) local ns = 0 local nl = 0 -- - for unic, glyph in next, descriptions do + for du, glyph in next, descriptions do local name = glyph.name if name then - local index = glyph.index - local r = overloads[name] - if r then + local overload = overloads[name] + if overload then -- get rid of weird ligatures - -- glyph.name = r.name - glyph.unicode = r.unicode - elseif not unic or unic == -1 or unic >= private or (unic >= 0xE000 and unic <= 0xF8FF) or unic == 0xFFFE or unic == 0xFFFF then - local unicode = unicodevector[name] or contextvector[name] - if unicode then - glyph.unicode = unicode - ns = ns + 1 - end - -- cidmap heuristics, beware, there is no guarantee for a match unless - -- the chain resolves - if (not unicode) and usedmap then - local foundindex = lpegmatch(oparser,name) - if foundindex then - unicode = cidcodes[foundindex] -- name to number - if unicode then - glyph.unicode = unicode - ns = ns + 1 - else - local reference = cidnames[foundindex] -- number to name - if reference then - local foundindex = lpegmatch(oparser,reference) - if foundindex then - unicode = cidcodes[foundindex] - if unicode then - glyph.unicode = unicode - ns = ns + 1 + -- glyph.name = overload.name + glyph.unicode = overload.unicode + else + local gu = glyph.unicode -- can already be set (number or table) + if not gu or gu == -1 or du >= private or (du >= 0xE000 and du <= 0xF8FF) or du == 0xFFFE or du == 0xFFFF then + local unicode = unicodevector[name] or contextvector[name] + if unicode then + glyph.unicode = unicode + ns = ns + 1 + end + -- cidmap heuristics, beware, there is no guarantee for a match unless + -- the chain resolves + if (not unicode) and usedmap then + local foundindex = lpegmatch(oparser,name) + if foundindex then + unicode = cidcodes[foundindex] -- name to number + if unicode then + glyph.unicode = unicode + ns = ns + 1 + else + local reference = cidnames[foundindex] -- number to name + if reference then + local foundindex = lpegmatch(oparser,reference) + if foundindex then + unicode = cidcodes[foundindex] + if unicode then + glyph.unicode = unicode + ns = ns + 1 + end end - end - if not unicode or unicode == "" then - local foundcodes, multiple = lpegmatch(uparser,reference) - if foundcodes then - glyph.unicode = foundcodes - if multiple then - nl = nl + 1 - unicode = true - else - ns = ns + 1 - unicode = foundcodes + if not unicode or unicode == "" then + local foundcodes, multiple = lpegmatch(uparser,reference) + if foundcodes then + glyph.unicode = foundcodes + if multiple then + nl = nl + 1 + unicode = true + else + ns = ns + 1 + unicode = foundcodes + end end end end end end end - end - -- a.whatever or a_b_c.whatever or a_b_c (no numbers) a.b_ - -- - -- It is not trivial to find a solution that suits all fonts. We tried several alternatives - -- and this one seems to work reasonable also with fonts that use less standardized naming - -- schemes. The extra private test is tested by KE and seems to work okay with non-typical - -- fonts as well. - -- - if not unicode or unicode == "" then - local split = lpegmatch(namesplitter,name) - local nsplit = split and #split or 0 -- add if - if nsplit == 0 then - -- skip - elseif nsplit == 1 then - local base = split[1] - local u = unicodes[base] or unicodevector[base] or contextvector[name] - if not u then + -- a.whatever or a_b_c.whatever or a_b_c (no numbers) a.b_ + -- + -- It is not trivial to find a solution that suits all fonts. We tried several alternatives + -- and this one seems to work reasonable also with fonts that use less standardized naming + -- schemes. The extra private test is tested by KE and seems to work okay with non-typical + -- fonts as well. + -- + if not unicode or unicode == "" then + local split = lpegmatch(namesplitter,name) + local nsplit = split and #split or 0 -- add if + if nsplit == 0 then -- skip - elseif type(u) == "table" then - -- unlikely - if u[1] < private then - unicode = u - glyph.unicode = unicode - end - elseif u < private then - unicode = u - glyph.unicode = unicode - end - else - local t, n = { }, 0 - for l=1,nsplit do - local base = split[l] + elseif nsplit == 1 then + local base = split[1] local u = unicodes[base] or unicodevector[base] or contextvector[name] if not u then - break + -- skip elseif type(u) == "table" then - if u[1] >= private then - break + -- unlikely + if u[1] < private then + unicode = u + glyph.unicode = unicode end - n = n + 1 - t[n] = u[1] - else - if u >= private then + elseif u < private then + unicode = u + glyph.unicode = unicode + end + else + local t, n = { }, 0 + for l=1,nsplit do + local base = split[l] + local u = unicodes[base] or unicodevector[base] or contextvector[name] + if not u then break + elseif type(u) == "table" then + if u[1] >= private then + break + end + n = n + 1 + t[n] = u[1] + else + if u >= private then + break + end + n = n + 1 + t[n] = u + end + end + if n > 0 then + if n == 1 then + unicode = t[1] + else + unicode = t end - n = n + 1 - t[n] = u + glyph.unicode = unicode end end - if n > 0 then - if n == 1 then - unicode = t[1] + nl = nl + 1 + end + -- last resort (we might need to catch private here as well) + if not unicode or unicode == "" then + local foundcodes, multiple = lpegmatch(uparser,name) + if foundcodes then + glyph.unicode = foundcodes + if multiple then + nl = nl + 1 + unicode = true else - unicode = t + ns = ns + 1 + unicode = foundcodes end - glyph.unicode = unicode end end - nl = nl + 1 - end - -- last resort (we might need to catch private here as well) - if not unicode or unicode == "" then - local foundcodes, multiple = lpegmatch(uparser,name) - if foundcodes then - glyph.unicode = foundcodes - if multiple then - nl = nl + 1 - unicode = true - else - ns = ns + 1 - unicode = foundcodes - end + -- check using substitutes and alternates + local r = overloads[unicode] + if r then + unicode = r.unicode + glyph.unicode = unicode + end + -- + if not unicode then + missing[du] = true + nofmissing = nofmissing + 1 end - end - -- check using substitutes and alternates - local r = overloads[unicode] - if r then - unicode = r.unicode - glyph.unicode = unicode - end - -- - if not unicode then - missing[unic] = true - nofmissing = nofmissing + 1 end end - else - -- no name end end if type(checklookups) == "function" then |