Drop langtags on offset text but support on paras

[LesML] / README.markdown

diff --git a/README.markdown b/README.markdown
index 327f253c5d3c63fdb24c0c14ba3e1607a4d43b1e..1979cd30517bf5bec7b8b027ed9a493c543038c8 100644 (file)
--- a/README.markdown
+++ b/README.markdown
@@ -179,10 +179,12 @@ After this classification, each quoted or unquoted paragraph is further
 
 - Otherwise, it is an ordinary paragraph.
 
 
 - Otherwise, it is an ordinary paragraph.
 

-Following this sigil (if any, including trailing white·space) there may
-  be a `¶` followed by zero or more non·white·space characters.
+Following this sigil (if any) there may be a `¶` followed by zero or
+  more non·white·space characters.

 The characters following the `¶` give the identifier for the paragraph,
   which is expected to be unique within a document.
 The characters following the `¶` give the identifier for the paragraph,
   which is expected to be unique within a document.

+This may be suffixed with a language tag beginning with `@` and
+  terminated with `$`.

 
 The remaining characters in a paragraph form its contents.
 Markup within paragraphs is delimited with·out exception by pairs of
 
 The remaining characters in a paragraph form its contents.
 Markup within paragraphs is delimited with·out exception by pairs of


@@ -193,6 +195,16 @@ Markup within paragraphs is delimited with·out exception by pairs of
     (consisting of `U+034F COMBINING GRAPHEME JOINER` for X·M·L
     compatibility).
 
     (consisting of `U+034F COMBINING GRAPHEME JOINER` for X·M·L
     compatibility).
 

+- The characters `{@` and `"}` indicate attribute specifications.
+  The attribute specification must contain at least one `="` which
+    separates the key of the attribute from the value.
+  Attributes attach to the previous element or text node, with
+    white·space‐only text nodes after elements ignored; if there is no
+    such previous element or text node, an empty text node is used
+    instead.
+  Multiple attributes can be given in sequence.
+  Text nodes with attributes are wrapped in `<html:span>`.
+

 - The characters `{🔗` and `>}` indicate a hyperlink to a U·R·L
     (`<html:a>`).
   The hyperlink must contain at least one `<`; the content before the
 - The characters `{🔗` and `>}` indicate a hyperlink to a U·R·L
     (`<html:a>`).
   The hyperlink must contain at least one `<`; the content before the


@@ -217,8 +229,6 @@ Markup within paragraphs is delimited with·out exception by pairs of
 - The characters `⸶` and `⸷` indicate names (`<html:u class="name">`).
 
 - The characters `⟨` and `⟩` indicate offset text (`<html:i>`).
 - The characters `⸶` and `⸷` indicate names (`<html:u class="name">`).
 
 - The characters `⟨` and `⟩` indicate offset text (`<html:i>`).

-  This may be followed by a `@`, a language tag, and a `$` to provide
-    the language of the text.

 
 - The characters `⦃` and `⦄` indicate keyword highlighting
     (`<html:b>`).
 
 - The characters `⦃` and `⦄` indicate keyword highlighting
     (`<html:b>`).


@@ -251,7 +261,9 @@ Finally, any character can be escaped by instead providing its Unicode
   codepoint in the form `{U+NNNN}`, where `NNNN` is one or more
   hexadecimal digits.
 Multiple codepoints may be provided separated by periods, as in
   codepoint in the form `{U+NNNN}`, where `NNNN` is one or more
   hexadecimal digits.
 Multiple codepoints may be provided separated by periods, as in

-  `{U+WWWW.ZZZZ}`
+  `{U+WWWW.ZZZZ}`.
+Due to limitations in X·S·L·T, characters cannot be escaped in
+  attributes (including link targets).

 
 ## Usage
 
 
 ## Usage