From 038261ea529bc4516d7cee501db70020938dbf2b Mon Sep 17 00:00:00 2001 From: Albert Krewinkel Date: Mon, 5 Apr 2021 21:45:52 +0200 Subject: JATS writer: escape disallows chars in identifiers XML identifiers must start with an underscore or letter, and can contain only a limited set of punctuation characters. Any IDs not adhering to these rules are rewritten by writing the offending characters as Uxxxx, where `xxxx` is the character's hex code. --- test/Tests/Writers/JATS.hs | 205 +++++++++++++++++++++++++-------------------- 1 file changed, 115 insertions(+), 90 deletions(-) (limited to 'test') diff --git a/test/Tests/Writers/JATS.hs b/test/Tests/Writers/JATS.hs index 2f501c890..23c1686dc 100644 --- a/test/Tests/Writers/JATS.hs +++ b/test/Tests/Writers/JATS.hs @@ -1,21 +1,21 @@ {-# LANGUAGE OverloadedStrings #-} module Tests.Writers.JATS (tests) where -import Data.Text (unpack) +import Data.Text (Text) import Test.Tasty import Tests.Helpers import Text.Pandoc import Text.Pandoc.Arbitrary () import Text.Pandoc.Builder +import qualified Data.Text as T -jats :: (ToPandoc a) => a -> String -jats = unpack - . purely (writeJATS def{ writerWrapText = WrapNone }) - . toPandoc +jats :: (ToPandoc a) => a -> Text +jats = purely (writeJATS def{ writerWrapText = WrapNone }) + . toPandoc -jatsArticleAuthoring :: (ToPandoc a) => a -> String -jatsArticleAuthoring = unpack - . purely (writeJatsArticleAuthoring def{ writerWrapText = WrapNone }) +jatsArticleAuthoring :: (ToPandoc a) => a -> Text +jatsArticleAuthoring = + purely (writeJatsArticleAuthoring def{ writerWrapText = WrapNone }) . toPandoc {- @@ -32,89 +32,114 @@ which is in turn shorthand for infix 4 =: (=:) :: (ToString a, ToPandoc a) - => String -> (a, String) -> TestTree + => String -> (a, Text) -> TestTree (=:) = test jats tests :: [TestTree] -tests = [ testGroup "inline code" - [ "basic" =: code "@&" =?> "

@&

" - , "lang" =: codeWith ("", ["c"], []) "@&" =?> "

@&

" - ] - , testGroup "block code" - [ "basic" =: codeBlock "@&" =?> "@&" - , "lang" =: codeBlockWith ("", ["c"], []) "@&" =?> "@&" - ] - , testGroup "images" - [ "basic" =: - image "/url" "title" mempty - =?> "" - ] - , testGroup "inlines" - [ "Emphasis" =: emph "emphasized" - =?> "

emphasized

" +tests = + [ testGroup "inline code" + [ "basic" =: code "@&" =?> "

@&

" + , "lang" =: codeWith ("", ["c"], []) "@&" =?> "

@&

" + ] + , testGroup "block code" + [ "basic" =: codeBlock "@&" =?> "@&" + , "lang" =: codeBlockWith ("", ["c"], []) "@&" =?> "@&" + ] + , testGroup "images" + [ "basic" =: + image "/url" "title" mempty + =?> "" + ] + , testGroup "inlines" + [ "Emphasis" =: emph "emphasized" + =?> "

emphasized

" + + , test jatsArticleAuthoring "footnote in articleauthoring tag set" + ("test" <> note (para "footnote") =?> + unlines [ "

test" + , "

footnote

" + , "

" + ]) + ] + , "bullet list" =: bulletList [ plain $ text "first" + , plain $ text "second" + , plain $ text "third" + ] + =?> "\n\ + \ \n\ + \

first

\n\ + \ \n\ + \ \n\ + \

second

\n\ + \ \n\ + \ \n\ + \

third

\n\ + \ \n\ + \" + , testGroup "definition lists" + [ "with internal link" =: definitionList [(link "#go" "" (str "testing"), + [plain (text "hi there")])] =?> + "\n\ + \ \n\ + \ testing\n\ + \ \n\ + \

hi there

\n\ + \ \n\ + \ \n\ + \" + ] + , testGroup "math" + [ "escape |" =: para (math "\\sigma|_{\\{x\\}}") =?> + "

\n\ + \\n\ + \σ|{x}

" + ] + , testGroup "headers" + [ "unnumbered header" =: + headerWith ("foo",["unnumbered"],[]) 1 + (text "Header 1" <> note (plain $ text "note")) =?> + "\n\ + \ Header 1<xref ref-type=\"fn\" rid=\"fn1\">1</xref>\n\ + \" + , "unnumbered sub header" =: + headerWith ("foo",["unnumbered"],[]) 1 + (text "Header") + <> headerWith ("foo",["unnumbered"],[]) 2 + (text "Sub-Header") =?> + "\n\ + \ Header\n\ + \ \n\ + \ Sub-Header\n\ + \ \n\ + \" + , "containing image" =: + header 1 (image "imgs/foo.jpg" "" (text "Alt text")) =?> + "\n\ + \ <inline-graphic mimetype=\"image\" mime-subtype=\"jpeg\" xlink:href=\"imgs/foo.jpg\" />\n\ + \" + ] + + , testGroup "ids" + [ "non-ASCII in header ID" =: + headerWith ("smørbrød",[],[]) 1 (text "smørbrød") =?> + T.unlines [ "" + , " smørbrød" + , "" + ] + + , "disallowed symbol in header id" =: + headerWith ("i/o",[],[]) 1 (text "I/O") =?> + T.unlines [ "" + , " I/O" + , "" + ] + + , "disallowed symbols in internal link target" =: + link "#foo:bar" "" "baz" =?> + "

baz

" - , test jatsArticleAuthoring "footnote in articleauthoring tag set" - ("test" <> note (para "footnote") =?> - unlines [ "

test" - , "

footnote

" - , "

" - ]) - ] - , "bullet list" =: bulletList [ plain $ text "first" - , plain $ text "second" - , plain $ text "third" - ] - =?> "\n\ - \ \n\ - \

first

\n\ - \ \n\ - \ \n\ - \

second

\n\ - \ \n\ - \ \n\ - \

third

\n\ - \ \n\ - \" - , testGroup "definition lists" - [ "with internal link" =: definitionList [(link "#go" "" (str "testing"), - [plain (text "hi there")])] =?> - "\n\ - \ \n\ - \ testing\n\ - \ \n\ - \

hi there

\n\ - \ \n\ - \ \n\ - \" - ] - , testGroup "math" - [ "escape |" =: para (math "\\sigma|_{\\{x\\}}") =?> - "

\n\ - \\n\ - \σ|{x}

" - ] - , testGroup "headers" - [ "unnumbered header" =: - headerWith ("foo",["unnumbered"],[]) 1 - (text "Header 1" <> note (plain $ text "note")) =?> - "\n\ - \ Header 1<xref ref-type=\"fn\" rid=\"fn1\">1</xref>\n\ - \" - , "unnumbered sub header" =: - headerWith ("foo",["unnumbered"],[]) 1 - (text "Header") - <> headerWith ("foo",["unnumbered"],[]) 2 - (text "Sub-Header") =?> - "\n\ - \ Header\n\ - \ \n\ - \ Sub-Header\n\ - \ \n\ - \" - , "containing image" =: - header 1 (image "imgs/foo.jpg" "" (text "Alt text")) =?> - "\n\ - \ <inline-graphic mimetype=\"image\" mime-subtype=\"jpeg\" xlink:href=\"imgs/foo.jpg\" />\n\ - \" - ] - ] + , "code id starting with a number" =: + codeWith ("7y",[],[]) "print 5" =?> + "

print 5

" + ] + ] -- cgit v1.2.3