summaryrefslogtreecommitdiff
path: root/lualibs-url.lua
diff options
context:
space:
mode:
Diffstat (limited to 'lualibs-url.lua')
-rw-r--r--lualibs-url.lua335
1 files changed, 271 insertions, 64 deletions
diff --git a/lualibs-url.lua b/lualibs-url.lua
index e3e6f81..83b8fcc 100644
--- a/lualibs-url.lua
+++ b/lualibs-url.lua
@@ -6,101 +6,291 @@ if not modules then modules = { } end modules ['l-url'] = {
license = "see context related readme files"
}
-local char, gmatch, gsub = string.char, string.gmatch, string.gsub
+local char, gmatch, gsub, format, byte, find = string.char, string.gmatch, string.gsub, string.format, string.byte, string.find
+local concat = table.concat
local tonumber, type = tonumber, type
-local lpegmatch = lpeg.match
+local P, C, R, S, Cs, Cc, Ct, Cf, Cg, V = lpeg.P, lpeg.C, lpeg.R, lpeg.S, lpeg.Cs, lpeg.Cc, lpeg.Ct, lpeg.Cf, lpeg.Cg, lpeg.V
+local lpegmatch, lpegpatterns, replacer = lpeg.match, lpeg.patterns, lpeg.replacer
--- from the spec (on the web):
+-- from wikipedia:
--
--- foo://example.com:8042/over/there?name=ferret#nose
--- \_/ \______________/\_________/ \_________/ \__/
--- | | | | |
--- scheme authority path query fragment
--- | _____________________|__
--- / \ / \
--- urn:example:animal:ferret:nose
-
-url = url or { }
-
-local function tochar(s)
- return char(tonumber(s,16))
-end
+-- foo://username:password@example.com:8042/over/there/index.dtb?type=animal;name=narwhal#nose
+-- \_/ \_______________/ \_________/ \__/ \___/ \_/ \______________________/ \__/
+-- | | | | | | | |
+-- | userinfo hostname port | | query fragment
+-- | \________________________________/\_____________|____|/
+-- scheme | | | |
+-- | authority path | |
+-- | | |
+-- | path interpretable as filename
+-- | ___________|____________ |
+-- / \ / \ |
+-- urn:example:animal:ferret:nose interpretable as extension
+
+url = url or { }
+local url = url
-local colon, qmark, hash, slash, percent, endofstring = lpeg.P(":"), lpeg.P("?"), lpeg.P("#"), lpeg.P("/"), lpeg.P("%"), lpeg.P(-1)
+local tochar = function(s) return char(tonumber(s,16)) end
-local hexdigit = lpeg.R("09","AF","af")
-local plus = lpeg.P("+")
-local escaped = (plus / " ") + (percent * lpeg.C(hexdigit * hexdigit) / tochar)
+local colon = P(":")
+local qmark = P("?")
+local hash = P("#")
+local slash = P("/")
+local percent = P("%")
+local endofstring = P(-1)
+
+local hexdigit = R("09","AF","af")
+local plus = P("+")
+local nothing = Cc("")
+local escapedchar = (percent * C(hexdigit * hexdigit)) / tochar
+local escaped = (plus / " ") + escapedchar
-- we assume schemes with more than 1 character (in order to avoid problems with windows disks)
+-- we also assume that when we have a scheme, we also have an authority
+--
+-- maybe we should already split the query (better for unescaping as = & can be part of a value
+
+local schemestr = Cs((escaped+(1-colon-slash-qmark-hash))^2)
+local authoritystr = Cs((escaped+(1- slash-qmark-hash))^0)
+local pathstr = Cs((escaped+(1- qmark-hash))^0)
+----- querystr = Cs((escaped+(1- hash))^0)
+local querystr = Cs(( (1- hash))^0)
+local fragmentstr = Cs((escaped+(1- endofstring))^0)
+
+local scheme = schemestr * colon + nothing
+local authority = slash * slash * authoritystr + nothing
+local path = slash * pathstr + nothing
+local query = qmark * querystr + nothing
+local fragment = hash * fragmentstr + nothing
+
+local validurl = scheme * authority * path * query * fragment
+local parser = Ct(validurl)
+
+lpegpatterns.url = validurl
+lpegpatterns.urlsplitter = parser
-local scheme = lpeg.Cs((escaped+(1-colon-slash-qmark-hash))^2) * colon + lpeg.Cc("")
-local authority = slash * slash * lpeg.Cs((escaped+(1- slash-qmark-hash))^0) + lpeg.Cc("")
-local path = slash * lpeg.Cs((escaped+(1- qmark-hash))^0) + lpeg.Cc("")
-local query = qmark * lpeg.Cs((escaped+(1- hash))^0) + lpeg.Cc("")
-local fragment = hash * lpeg.Cs((escaped+(1- endofstring))^0) + lpeg.Cc("")
+local escapes = { }
-local parser = lpeg.Ct(scheme * authority * path * query * fragment)
+setmetatable(escapes, { __index = function(t,k)
+ local v = format("%%%02X",byte(k))
+ t[k] = v
+ return v
+end })
+
+local escaper = Cs((R("09","AZ","az")^1 + P(" ")/"%%20" + S("-./_")^1 + P(1) / escapes)^0) -- space happens most
+local unescaper = Cs((escapedchar + 1)^0)
+
+lpegpatterns.urlunescaped = escapedchar
+lpegpatterns.urlescaper = escaper
+lpegpatterns.urlunescaper = unescaper
-- todo: reconsider Ct as we can as well have five return values (saves a table)
-- so we can have two parsers, one with and one without
-function url.split(str)
+local function split(str)
return (type(str) == "string" and lpegmatch(parser,str)) or str
end
--- todo: cache them
+local isscheme = schemestr * colon * slash * slash -- this test also assumes authority
-function url.hashed(str)
- local s = url.split(str)
- local somescheme = s[1] ~= ""
- return {
- scheme = (somescheme and s[1]) or "file",
- authority = s[2],
- path = s[3],
- query = s[4],
- fragment = s[5],
- original = str,
- noscheme = not somescheme,
- }
+local function hasscheme(str)
+ if str then
+ local scheme = lpegmatch(isscheme,str) -- at least one character
+ return scheme ~= "" and scheme or false
+ else
+ return false
+ end
end
-function url.hasscheme(str)
- return url.split(str)[1] ~= ""
+--~ print(hasscheme("home:"))
+--~ print(hasscheme("home://"))
+
+-- todo: cache them
+
+local rootletter = R("az","AZ")
+ + S("_-+")
+local separator = P("://")
+local qualified = P(".")^0 * P("/")
+ + rootletter * P(":")
+ + rootletter^1 * separator
+ + rootletter^1 * P("/")
+local rootbased = P("/")
+ + rootletter * P(":")
+
+local barswapper = replacer("|",":")
+local backslashswapper = replacer("\\","/")
+
+-- queries:
+
+local equal = P("=")
+local amp = P("&")
+local key = Cs(((escapedchar+1)-equal )^0)
+local value = Cs(((escapedchar+1)-amp -endofstring)^0)
+
+local splitquery = Cf ( Ct("") * P { "sequence",
+ sequence = V("pair") * (amp * V("pair"))^0,
+ pair = Cg(key * equal * value),
+}, rawset)
+
+-- hasher
+
+local function hashed(str) -- not yet ok (/test?test)
+ if str == "" then
+ return {
+ scheme = "invalid",
+ original = str,
+ }
+ end
+ local s = split(str)
+ local rawscheme = s[1]
+ local rawquery = s[4]
+ local somescheme = rawscheme ~= ""
+ local somequery = rawquery ~= ""
+ if not somescheme and not somequery then
+ s = {
+ scheme = "file",
+ authority = "",
+ path = str,
+ query = "",
+ fragment = "",
+ original = str,
+ noscheme = true,
+ filename = str,
+ }
+ else -- not always a filename but handy anyway
+ local authority, path, filename = s[2], s[3]
+ if authority == "" then
+ filename = path
+ elseif path == "" then
+ filename = ""
+ else
+ filename = authority .. "/" .. path
+ end
+ s = {
+ scheme = rawscheme,
+ authority = authority,
+ path = path,
+ query = lpegmatch(unescaper,rawquery), -- unescaped, but possible conflict with & and =
+ queries = lpegmatch(splitquery,rawquery), -- split first and then unescaped
+ fragment = s[5],
+ original = str,
+ noscheme = false,
+ filename = filename,
+ }
+ end
+ return s
end
-function url.addscheme(str,scheme)
- return (url.hasscheme(str) and str) or ((scheme or "file:///") .. str)
+-- inspect(hashed("template://test"))
+
+-- Here we assume:
+--
+-- files: /// = relative
+-- files: //// = absolute (!)
+
+--~ table.print(hashed("file://c:/opt/tex/texmf-local")) -- c:/opt/tex/texmf-local
+--~ table.print(hashed("file://opt/tex/texmf-local" )) -- opt/tex/texmf-local
+--~ table.print(hashed("file:///opt/tex/texmf-local" )) -- opt/tex/texmf-local
+--~ table.print(hashed("file:////opt/tex/texmf-local" )) -- /opt/tex/texmf-local
+--~ table.print(hashed("file:///./opt/tex/texmf-local" )) -- ./opt/tex/texmf-local
+
+--~ table.print(hashed("c:/opt/tex/texmf-local" )) -- c:/opt/tex/texmf-local
+--~ table.print(hashed("opt/tex/texmf-local" )) -- opt/tex/texmf-local
+--~ table.print(hashed("/opt/tex/texmf-local" )) -- /opt/tex/texmf-local
+
+url.split = split
+url.hasscheme = hasscheme
+url.hashed = hashed
+
+function url.addscheme(str,scheme) -- no authority
+ if hasscheme(str) then
+ return str
+ elseif not scheme then
+ return "file:///" .. str
+ else
+ return scheme .. ":///" .. str
+ end
end
-function url.construct(hash)
- local fullurl = hash.sheme .. "://".. hash.authority .. hash.path
- if hash.query then
- fullurl = fullurl .. "?".. hash.query
+function url.construct(hash) -- dodo: we need to escape !
+ local fullurl, f = { }, 0
+ local scheme, authority, path, query, fragment = hash.scheme, hash.authority, hash.path, hash.query, hash.fragment
+ if scheme and scheme ~= "" then
+ f = f + 1 ; fullurl[f] = scheme .. "://"
end
- if hash.fragment then
- fullurl = fullurl .. "?".. hash.fragment
+ if authority and authority ~= "" then
+ f = f + 1 ; fullurl[f] = authority
end
- return fullurl
+ if path and path ~= "" then
+ f = f + 1 ; fullurl[f] = "/" .. path
+ end
+ if query and query ~= "" then
+ f = f + 1 ; fullurl[f] = "?".. query
+ end
+ if fragment and fragment ~= "" then
+ f = f + 1 ; fullurl[f] = "#".. fragment
+ end
+ return lpegmatch(escaper,concat(fullurl))
end
-function url.filename(filename)
- local t = url.hashed(filename)
+function url.filename(filename) -- why no lpeg here ?
+ local t = hashed(filename)
return (t.scheme == "file" and (gsub(t.path,"^/([a-zA-Z])([:|])/)","%1:"))) or filename
end
+local function escapestring(str)
+ return lpegmatch(escaper,str)
+end
+
+url.escape = escapestring
+
+-- function url.query(str) -- separator could be an option
+-- if type(str) == "string" then
+-- local t = { }
+-- for k, v in gmatch(str,"([^&=]*)=([^&=]*)") do
+-- t[k] = v
+-- end
+-- return t
+-- else
+-- return str
+-- end
+-- end
+
function url.query(str)
if type(str) == "string" then
- local t = { }
- for k, v in gmatch(str,"([^&=]*)=([^&=]*)") do
- t[k] = v
- end
- return t
+ return lpegmatch(splitquery,str) or ""
else
return str
end
end
+function url.toquery(data)
+ local td = type(data)
+ if td == "string" then
+ return #str and escape(data) or nil -- beware of double escaping
+ elseif td == "table" then
+ if next(data) then
+ local t = { }
+ for k, v in next, data do
+ t[#t+1] = format("%s=%s",k,escapestring(v))
+ end
+ return concat(t,"&")
+ end
+ else
+ -- nil is a signal that no query
+ end
+end
+
+-- /test/ | /test | test/ | test => test
+
+function url.barepath(path)
+ if not path or path == "" then
+ return ""
+ else
+ return (gsub(path,"^/?(.-)/?$","%1"))
+ end
+end
+
--~ print(url.filename("file:///c:/oeps.txt"))
--~ print(url.filename("c:/oeps.txt"))
--~ print(url.filename("file:///oeps.txt"))
@@ -108,12 +298,30 @@ end
--~ print(url.filename("/oeps.txt"))
--~ from the spec on the web (sort of):
---~
---~ function test(str)
---~ print(table.serialize(url.hashed(str)))
+
+--~ local function test(str)
+--~ local t = url.hashed(str)
+--~ t.constructed = url.construct(t)
+--~ print(table.serialize(t))
--~ end
---~
---~ test("%56pass%20words")
+
+--~ inspect(url.hashed("http://www.pragma-ade.com/test%20test?test=test%20test&x=123%3d45"))
+--~ inspect(url.hashed("http://www.pragma-ade.com/test%20test?test=test%20test&x=123%3d45"))
+
+--~ test("sys:///./colo-rgb")
+
+--~ test("/data/site/output/q2p-develop/resources/ecaboperception4_res/topicresources/58313733/figuur-cow.jpg")
+--~ test("file:///M:/q2p/develop/output/q2p-develop/resources/ecaboperception4_res/topicresources/58313733")
+--~ test("M:/q2p/develop/output/q2p-develop/resources/ecaboperception4_res/topicresources/58313733")
+--~ test("file:///q2p/develop/output/q2p-develop/resources/ecaboperception4_res/topicresources/58313733")
+--~ test("/q2p/develop/output/q2p-develop/resources/ecaboperception4_res/topicresources/58313733")
+
+--~ test("file:///cow%20with%20spaces")
+--~ test("file:///cow%20with%20spaces.pdf")
+--~ test("cow%20with%20spaces.pdf")
+--~ test("some%20file")
+--~ test("/etc/passwords")
+--~ test("http://www.myself.com/some%20words.html")
--~ test("file:///c:/oeps.txt")
--~ test("file:///c|/oeps.txt")
--~ test("file:///etc/oeps.txt")
@@ -127,7 +335,6 @@ end
--~ test("tel:+1-816-555-1212")
--~ test("telnet://192.0.2.16:80/")
--~ test("urn:oasis:names:specification:docbook:dtd:xml:4.1.2")
---~ test("/etc/passwords")
--~ test("http://www.pragma-ade.com/spaced%20name")
--~ test("zip:///oeps/oeps.zip#bla/bla.tex")
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-01 02:37:44 +0000