[jena] 01/02: Fix IRIs.asBase

[andy]

This is an automated email from the ASF dual-hosted git repository.

andy pushed a commit to branch jena5
in repository https://gitbox.apache.org/repos/asf/jena.git

commit 53baf40c39d565df2f16db5589045819df471f78
Author: Andy Seaborne <a...@apache.org>
AuthorDate: Sun Oct 22 20:22:15 2023 +0100

    Fix IRIs.asBase
---
 .../src/main/java/org/apache/jena/irix/IRIs.java   | 30 ++++++++++++++--------
 .../main/java/org/apache/jena/irix/SystemIRIx.java |  2 --
 2 files changed, 19 insertions(+), 13 deletions(-)

diff --git a/jena-core/src/main/java/org/apache/jena/irix/IRIs.java 
b/jena-core/src/main/java/org/apache/jena/irix/IRIs.java
index 2cdbbc8bd4..12209317de 100644
--- a/jena-core/src/main/java/org/apache/jena/irix/IRIs.java
+++ b/jena-core/src/main/java/org/apache/jena/irix/IRIs.java
@@ -76,24 +76,32 @@ public class IRIs {
     /**
      * Given a candidate baseURI string, which may be a filename,
      * turn it into a IRI suitable as a base IRI.
+     * This includes encoding characters in a filename (e.g. spaces).
      */
-    public static String toBase(String baseURI) {
-        if ( baseURI == null )
+    public static String toBase(String uriForBase) {
+        if ( uriForBase == null )
             return getBaseStr();
-        String scheme = scheme(baseURI);
+        String scheme = scheme(uriForBase);
         if ( Sys.isWindows ) {
             // Assume a scheme of one letter is a Windows drive letter.
             if ( scheme != null && scheme.length() == 1 )
                 scheme = "file";
         }
-        // File scheme if the URI is "file:" or there is no scheme
-        // and the system base is a "file:" URI
-        boolean isFile = ( scheme != null )
-                ? scheme.equals("file")
-                : IRIs.getSystemBase().hasScheme("file");
-        if ( isFile )
-            baseURI = IRILib.encodeFileURL(baseURI);
-        return IRIs.getSystemBase().resolve(baseURI).toString();
+        if ( scheme == null  ) {
+            // Relative name: it the base is a file: URI, encode the relative
+            // name if it does not look like it is already encoded.
+            boolean isFileBase = IRIs.getSystemBase().hasScheme("file");
+            if ( isFileBase && ! uriForBase.contains("%") )
+                uriForBase = IRILib.encodeFileURL(uriForBase);
+        } else {
+            // If the scheme of the proposed base URI is file: then assume it 
is a legal IRI as intended.
+            // Pragmatically, fix-up a few characters that are illegal.
+            if ( scheme.equals("file") ) {
+                uriForBase = uriForBase.replace(" ", "%20");
+                uriForBase = uriForBase.replace("\\", "/");
+            }
+        }
+        return IRIs.getSystemBase().resolve(uriForBase).toString();
     }
 
     /** Return a general purpose resolver, with the current system base as its 
base IRI. */
diff --git a/jena-core/src/main/java/org/apache/jena/irix/SystemIRIx.java 
b/jena-core/src/main/java/org/apache/jena/irix/SystemIRIx.java
index a22fbb740d..ccffc388bf 100644
--- a/jena-core/src/main/java/org/apache/jena/irix/SystemIRIx.java
+++ b/jena-core/src/main/java/org/apache/jena/irix/SystemIRIx.java
@@ -123,8 +123,6 @@ public class SystemIRIx {
         if ( baseStr == null )
             return IRIx.create(fallbackBaseURI);
         try {
-            if ( !baseStr.endsWith("/") )
-                baseStr = baseStr+"/";
             IRIx base = IRIx.create(baseStr);
             if ( ! base.isReference() )
                 Log.error(IRIs.class, "System base URI is not a reference URI: 
must have scheme, host and path");