From 6ca012d4fb97c02a2206aebd42aef1f9cd5e1993 Mon Sep 17 00:00:00 2001
From: David Robillard <d@drobilla.net>
Date: Sun, 26 Feb 2023 16:57:46 -0500
Subject: [WIP] Partially resurrect CURIE nodes

---
 include/serd/node.h   | 12 ++++++++++--
 include/serd/reader.h | 30 +++++++++++++++++++++++++-----
 2 files changed, 35 insertions(+), 7 deletions(-)

(limited to 'include')

diff --git a/include/serd/node.h b/include/serd/node.h
index 941be5d1..78f5baae 100644
--- a/include/serd/node.h
+++ b/include/serd/node.h
@@ -70,6 +70,14 @@ typedef enum {
   */
   SERD_URI = 2,
 
+  /**
+     CURIE, a shortened URI.
+
+     Value is an unquoted CURIE string relative to the current environment,
+     e.g. "rdf:type".  @see [CURIE Syntax 1.0](http://www.w3.org/TR/curie)
+  */
+  SERD_CURIE = 3,
+
   /**
      A blank node.
 
@@ -80,7 +88,7 @@ typedef enum {
      @see [RDF 1.1
      Turtle](http://www.w3.org/TR/turtle/#grammar-production-BLANK_NODE_LABEL)
   */
-  SERD_BLANK = 3,
+  SERD_BLANK = 4,
 
   /**
      A variable node.
@@ -93,7 +101,7 @@ typedef enum {
      @see [SPARQL 1.1 Query
      Language](https://www.w3.org/TR/sparql11-query/#rVar)
   */
-  SERD_VARIABLE = 4,
+  SERD_VARIABLE = 5,
 } SerdNodeType;
 
 /// Node flags, which ORed together make a #SerdNodeFlags
diff --git a/include/serd/reader.h b/include/serd/reader.h
index 57c8b2c3..a500bb88 100644
--- a/include/serd/reader.h
+++ b/include/serd/reader.h
@@ -54,6 +54,26 @@ typedef enum {
   */
   SERD_READ_VARIABLES = 1U << 1U,
 
+  /**
+     Read prefixed name (CURIE) references exactly without expanding them.
+
+     Normally, the reader expands all prefixed names to full URIs based on the
+     prefixes in the current environment, and considers failure to expand a
+     syntax error.  This flag disables that expansion so prefixed names will be
+     emitted directly as CURIE nodes.
+
+     Note that these nodes rely on some context which can change over time, and
+     may even be undefined initially, so this flag should be used with caution.
+     Most applications should leave it off and avoid using CURIE nodes
+     entirely, because they are error-prone compared to working with complete
+     URIs.  However, it can be useful for error-tolerance, or in constrained or
+     high-performance streaming contexts.  For example, to re-indent a Turtle
+     file and ignore any possibly undefined prefixed names, this flag can be
+     used to disable expansion, which also boosts performance since it avoids
+     the lookup and expansion overhead.
+  */
+  SERD_READ_CURIES = 1U << 2U,
+
   /**
      Read relative URI references exactly without resolving them.
 
@@ -61,7 +81,7 @@ typedef enum {
      flag disables that, so that URI references are passed to the sink exactly
      as they are in the input.
   */
-  SERD_READ_RELATIVE = 1U << 2U,
+  SERD_READ_RELATIVE = 1U << 3U,
 
   /**
      Read blank node labels without adding a prefix unique to the document.
@@ -75,7 +95,7 @@ typedef enum {
      corruption.  Specifically, if data from separate documents parsed with
      this flag is combined, the IDs from each document may clash.
   */
-  SERD_READ_GLOBAL = 1U << 3U,
+  SERD_READ_GLOBAL = 1U << 4U,
 
   /**
      Read generated blank node labels exactly without adjusting them.
@@ -90,7 +110,7 @@ typedef enum {
      anonymous nodes, the generated IDs for those nodes may clash with IDs from
      the input document.
   */
-  SERD_READ_GENERATED = 1U << 4U,
+  SERD_READ_GENERATED = 1U << 5U,
 
   /**
      Generate blank node labels with suffixes left-padded with zeros.
@@ -100,7 +120,7 @@ typedef enum {
      nodes).  In particular, this can be used to preserve blank node ordering
      from documents when the statements are sorted, such as in a model.
   */
-  SERD_READ_ORDERED = 1U << 5U,
+  SERD_READ_ORDERED = 1U << 6U,
 
   /**
      Read URIs with percent-encoded UTF-8 characters decoded.
@@ -109,7 +129,7 @@ typedef enum {
      preserved.  This flags enables UTF-8 decoding of URIs, so octet escapes
      like "%7E" in URIs will be decoded to UTF-8 characters like "~".
   */
-  SERD_READ_DECODED = 1U << 6U,
+  SERD_READ_DECODED = 1U << 7U,
 } SerdReaderFlag;
 
 /// Bitwise OR of SerdReaderFlag values
-- 
cgit v1.2.1