Merge pull request #3901 from maxonfjvipon/bug/#3864/unhex-via-xsl

bug(#3864): unhexing via XSL

eo-maven-plugin/src/main/java/org/eolang/maven/LintMojo.java

@@ -36,7 +36,6 @@
 import java.util.concurrent.ConcurrentHashMap;
 import org.apache.maven.plugins.annotations.LifecyclePhase;
 import org.apache.maven.plugins.annotations.Mojo;
-import org.apache.maven.plugins.annotations.Parameter;
 import org.cactoos.list.ListOf;
 import org.eolang.lints.Defect;
 import org.eolang.lints.Program;
@@ -72,26 +71,20 @@ public final class LintMojo extends SafeMojo {
      */
     static final String CACHE = "linted";
 
-    /**
-     * Whether we should fail on warning.
-     *
-     * @checkstyle MemberNameCheck (11 lines)
-     */
-    @SuppressWarnings("PMD.ImmutableField")
-    @Parameter(property = "eo.failOnWarning", required = true, defaultValue = "true")
-    private boolean failOnWarning;
+    @Override
+    void exec() throws IOException {
+        if (this.skipLinting) {
+            Logger.info(this, "Linting is skipped because eo:skipLinting is TRUE");
+        } else {
+            this.lint();
+        }
+    }
 
     /**
-     * Whether we should lint all the sources together as package.
-     *
-     * @checkstyle MemberNameCheck (11 lines)
+     * Lint.
+     * @throws IOException If fails
      */
-    @SuppressWarnings("PMD.ImmutableField")
-    @Parameter(property = "eo.lintAsPackage", required = true, defaultValue = "true")
-    private boolean lintAsPackage;
-
-    @Override
-    void exec() throws IOException {
+    private void lint() throws IOException {
         final long start = System.currentTimeMillis();
         final Collection<ForeignTojo> tojos = this.scopedTojos().withShaken();
         final ConcurrentHashMap<Severity, Integer> counts = new ConcurrentHashMap<>();
@@ -156,7 +149,7 @@ private int lintOne(final ForeignTojo tojo,
         final Path target = new Place(name).make(base, AssembleMojo.XMIR);
         tojo.withLinted(
             new FpDefault(
-                src -> LintMojo.lint(xmir, counts).toString(),
+                src -> LintMojo.linted(xmir, counts).toString(),
                 this.cache.toPath().resolve(LintMojo.CACHE),
                 this.plugin.getVersion(),
                 new TojoHash(tojo),
@@ -283,8 +276,7 @@ private static String summary(final ConcurrentHashMap<Severity, Integer> counts)
      * @param counts Counts of errors, warnings, and critical
      * @return XML after linting
      */
-    private static XML lint(final XML xmir,
-        final ConcurrentHashMap<Severity, Integer> counts) {
+    private static XML linted(final XML xmir, final ConcurrentHashMap<Severity, Integer> counts) {
         final Directives dirs = new Directives();
         final Collection<Defect> defects = new Program(xmir).defects();
         if (!defects.isEmpty()) {

eo-maven-plugin/src/main/java/org/eolang/maven/SafeMojo.java

@@ -379,6 +379,36 @@ abstract class SafeMojo extends AbstractMojo {
     @SuppressWarnings("PMD.ImmutableField")
     protected boolean ignoreTransitive;
 
+    /**
+     * Whether we should fail on warning.
+     *
+     * @checkstyle MemberNameCheck (10 lines)
+     * @checkstyle VisibilityModifierCheck (7 lines)
+     */
+    @SuppressWarnings("PMD.ImmutableField")
+    @Parameter(property = "eo.failOnWarning", required = true, defaultValue = "true")
+    protected boolean failOnWarning;
+
+    /**
+     * Whether we should lint all the sources together as package.
+     *
+     * @checkstyle MemberNameCheck (10 lines)
+     * @checkstyle VisibilityModifierCheck (7 lines)
+     */
+    @SuppressWarnings("PMD.ImmutableField")
+    @Parameter(property = "eo.lintAsPackage", required = true, defaultValue = "true")
+    protected boolean lintAsPackage;
+
+    /**
+     * Whether we should skip linting at all.
+     *
+     * @checkstyle MemberNameCheck (10 lines)
+     * @checkstyle VisibilityModifierCheck (7 lines)
+     */
+    @SuppressWarnings("PMD.ImmutableField")
+    @Parameter(property = "eo.skipLinting", required = true, defaultValue = "false")
+    protected boolean skipLinting;
+
     /**
      * The current version of eo-maven-plugin.
      * Maven 3 only.

eo-maven-plugin/src/main/resources/org/eolang/maven/shake/cti-adds-errors.xsl

@@ -26,6 +26,7 @@ SOFTWARE.
   <!--
   For every cti objects add error messages.
   -->
+  <xsl:import href="/org/eolang/parser/_funcs.xsl"/>
   <xsl:output encoding="UTF-8" method="xml"/>
   <xsl:template match="/program/errors">
     <xsl:copy>
@@ -52,35 +53,14 @@ SOFTWARE.
         <xsl:value-of select="@line"/>
       </xsl:attribute>
       <xsl:attribute name="severity">
-        <xsl:value-of select="eo:hex-to-utf8(o[last() - 1]/o[1]/text())"/>
+        <xsl:value-of select="eo:bytes-to-string(o[last() - 1]/o[1]/text())"/>
       </xsl:attribute>
-      <xsl:value-of select="eo:hex-to-utf8(o[last()]/o[1]/text())"/>
+      <xsl:value-of select="eo:bytes-to-string(o[last()]/o[1]/text())"/>
     </xsl:element>
   </xsl:template>
   <xsl:template match="node()|@*">
     <xsl:copy>
       <xsl:apply-templates select="node()|@*"/>
     </xsl:copy>
   </xsl:template>
-  <!--Converts hex sting into readable UTF-8 string-->
-  <xsl:function name="eo:hex-to-utf8">
-    <xsl:param name="str"/>
-    <xsl:variable name="hex" select="'0123456789ABCDEF'"/>
-    <xsl:variable name="tail" select="translate($str, '-', '')"/>
-    <!-- Base case: Return empty string if input is empty or invalid -->
-    <xsl:if test="string-length($tail) &gt;= 2">
-      <!-- Extract first 2 digits -->
-      <xsl:variable name="first" select="substring($tail, 1, 1)"/>
-      <xsl:variable name="second" select="substring($tail, 2, 1)"/>
-      <!-- Get their hex values -->
-      <xsl:variable name="val1" select="string-length(substring-before($hex, $first))"/>
-      <xsl:variable name="val2" select="string-length(substring-before($hex, $second))"/>
-      <!-- Ensure valid character range -->
-      <xsl:variable name="codepoint" select="$val1 * 16 + $val2"/>
-      <xsl:if test="$codepoint &gt; 0">
-        <xsl:variable name="head" select="codepoints-to-string($codepoint)"/>
-        <xsl:value-of select="concat($head, eo:hex-to-utf8(substring($tail, 3)))"/>
-      </xsl:if>
-    </xsl:if>
-  </xsl:function>
 </xsl:stylesheet>

eo-parser/src/main/java/org/eolang/parser/Xmir.java

@@ -31,6 +31,7 @@
 import com.yegor256.xsline.StEndless;
 import com.yegor256.xsline.TrClasspath;
 import com.yegor256.xsline.TrDefault;
+import com.yegor256.xsline.TrJoined;
 import com.yegor256.xsline.Train;
 import com.yegor256.xsline.Xsline;
 import java.util.Collection;
@@ -57,23 +58,22 @@
  */
 @SuppressWarnings("PMD.TooManyMethods")
 public final class Xmir implements XML {
-    /**
-     * Unhex transformation.
-     */
-    private static final Shift UNHEX = new StUnhex();
-
     /**
      * Train of transformations that prepare XMIR for conversion to EO.
      */
     private static final Train<Shift> FOR_EO = new TrFull(
-        new TrDefault<>(
-            new StEndless(
-                new StClasspath("/org/eolang/parser/print/tuples-to-stars.xsl")
+        new TrJoined<>(
+            new TrDefault<>(
+                new StEndless(
+                    new StClasspath("/org/eolang/parser/print/tuples-to-stars.xsl")
+                )
             ),
-            new StClasspath("/org/eolang/parser/print/dataized-to-const.xsl"),
-            Xmir.UNHEX,
-            new StClasspath("/org/eolang/parser/print/wrap-data.xsl"),
-            new StClasspath("/org/eolang/parser/print/to-eo.xsl")
+            new TrClasspath<>(
+                "/org/eolang/parser/print/dataized-to-const.xsl",
+                "/org/eolang/parser/print/unhex-data.xsl",
+                "/org/eolang/parser/print/wrap-data.xsl",
+                "/org/eolang/parser/print/to-eo.xsl"
+            ).back()
         )
     );
 
@@ -166,7 +166,7 @@ public String toPhi(final boolean conservative) {
         return this.converted(
             new TrFull(
                 new TrDefault<>(
-                    Xmir.UNHEX,
+                    new StClasspath("/org/eolang/parser/print/unhex-data.xsl"),
                     new StClasspath(
                         "/org/eolang/parser/phi/to-phi.xsl",
                         String.format("conservative %b", conservative)

eo-parser/src/main/resources/org/eolang/parser/_funcs.xsl

@@ -22,7 +22,7 @@ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 SOFTWARE.
 -->
-<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:eo="https://www.eolang.org" xmlns:xs="http://www.w3.org/2001/XMLSchema" id="_funcs" version="2.0">
+<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:math="http://www.w3.org/2005/xpath-functions/math" xmlns:eo="https://www.eolang.org" xmlns:xs="http://www.w3.org/2001/XMLSchema" id="_funcs" version="2.0">
   <xsl:function name="eo:has-data" as="xs:boolean">
     <xsl:param name="o" as="element()"/>
     <xsl:sequence select="normalize-space(string-join($o/text(), '')) != ''"/>
@@ -35,4 +35,135 @@ SOFTWARE.
     <xsl:param name="o" as="element()"/>
     <xsl:sequence select="$o/@base='∅'"/>
   </xsl:function>
+  <!-- BYTES TO STRING -->
+  <xsl:function name="eo:bytes-to-string" as="xs:string">
+    <xsl:param name="bytes" as="xs:string"/>
+    <xsl:choose>
+      <xsl:when test="$bytes = '--'">
+        <xsl:sequence select="''"/>
+      </xsl:when>
+      <xsl:otherwise>
+        <xsl:variable name="decoded">
+          <xsl:for-each select="eo:decode-bytes(for $byte in (if (ends-with($bytes, '-')) then substring-before($bytes, '-') else tokenize($bytes, '-')) return eo:hex-to-utf8($byte))">
+            <xsl:choose>
+              <xsl:when test=".=10">
+                <xsl:value-of select="'\n'"/>
+              </xsl:when>
+              <xsl:when test=".=9">
+                <xsl:value-of select="'\t'"/>
+              </xsl:when>
+              <xsl:when test=".=13">
+                <xsl:value-of select="'\r'"/>
+              </xsl:when>
+              <!-- Keep ASCII characters -->
+              <xsl:when test=". ge 32 and . le 126">
+                <xsl:variable name="char" select="codepoints-to-string(.)"/>
+                <xsl:if test="$char='\' or $char='&quot;'">
+                  <xsl:text>\</xsl:text>
+                </xsl:if>
+                <xsl:value-of select="$char"/>
+              </xsl:when>
+              <!-- Convert non-ASCII to \uXXXX -->
+              <xsl:when test=". le 65535">
+                <xsl:value-of select="concat('\u', eo:int-to-hex(xs:int(.)))"/>
+              </xsl:when>
+              <!-- Handle surrogate pairs for code points above U+FFFF -->
+              <xsl:otherwise>
+                <!-- 55296 = 0xD800 -->
+                <xsl:variable name="cp1" select="xs:int(floor((. - 65536) div 1024) + 55296)"/>
+                <!-- 56320 = 0xDC00 -->
+                <xsl:variable name="cp2" select="xs:int(((. - 65536) mod 1024) + 56320)"/>
+                <xsl:value-of select="concat('\u', eo:int-to-hex($cp1), '\u', eo:int-to-hex($cp2))"/>
+              </xsl:otherwise>
+            </xsl:choose>
+          </xsl:for-each>
+        </xsl:variable>
+        <xsl:sequence select="$decoded"/>
+      </xsl:otherwise>
+    </xsl:choose>
+  </xsl:function>
+  <!-- BYTES TO NUMBER, e.g. 40-14-00-00-00-00-00-00 => 5 -->
+  <xsl:function name="eo:bytes-to-number" as="xs:anyAtomicType">
+    <xsl:param name="bytes"/>
+    <!-- Undash -->
+    <xsl:variable name="hex" select="translate($bytes, '-', '')"/>
+    <xsl:variable name="map" as="element()*">
+      <entry h="0" b="0000"/>
+      <entry h="1" b="0001"/>
+      <entry h="2" b="0010"/>
+      <entry h="3" b="0011"/>
+      <entry h="4" b="0100"/>
+      <entry h="5" b="0101"/>
+      <entry h="6" b="0110"/>
+      <entry h="7" b="0111"/>
+      <entry h="8" b="1000"/>
+      <entry h="9" b="1001"/>
+      <entry h="A" b="1010"/>
+      <entry h="B" b="1011"/>
+      <entry h="C" b="1100"/>
+      <entry h="D" b="1101"/>
+      <entry h="E" b="1110"/>
+      <entry h="F" b="1111"/>
+    </xsl:variable>
+    <xsl:variable name="bin" as="xs:string" select="string-join(for $c in string-to-codepoints(upper-case($hex)) return $map[@h = codepoints-to-string($c)]/@b, '')"/>
+    <!-- Sign bit (1 for negative, 0 for positive) -->
+    <xsl:variable name="sign" select="if (substring($bin, 1, 1) = '1') then -1 else 1"/>
+    <!-- Extract exponent (11 bits) and convert to integer -->
+    <xsl:variable name="exponentBits" select="substring($bin, 2, 11)"/>
+    <xsl:variable name="exponent" select="sum(for $i in 1 to string-length($exponentBits) return xs:double(substring($exponentBits, $i, 1)) * math:pow(2, string-length($exponentBits) - $i)) - 1023"/>
+    <!-- Extract mantissa (52 bits) -->
+    <xsl:variable name="mantissaBits" select="substring($bin, 13, 52)"/>
+    <xsl:variable name="mantissaValue">
+      <xsl:sequence select="sum(for $i in 1 to string-length($mantissaBits) return xs:double(substring($mantissaBits, $i, 1)) div math:pow(2, $i))"/>
+    </xsl:variable>
+    <!-- Compute final double value -->
+    <xsl:sequence select="$sign * (1 + $mantissaValue) * math:pow(2, $exponent)"/>
+  </xsl:function>
+  <!-- HELPER FUNCTIONS -->
+  <!-- Function to decode UTF-8 bytes into Unicode code points -->
+  <xsl:function name="eo:decode-bytes" as="xs:integer*">
+    <xsl:param name="bytes" as="xs:integer*"/>
+    <xsl:choose>
+      <!-- 1-byte sequence: 0xxxxxxx -->
+      <xsl:when test="$bytes[1] lt 128">
+        <xsl:sequence select="$bytes[1]"/>
+        <xsl:sequence select="eo:decode-bytes(subsequence($bytes, 2))"/>
+      </xsl:when>
+      <!-- 2-byte sequence: 110xxxxx 10xxxxxx -->
+      <xsl:when test="$bytes[1] ge 192 and $bytes[1] lt 224">
+        <xsl:variable name="code-point" select="(($bytes[1] - 192) * 64) + ($bytes[2] - 128)"/>
+        <xsl:sequence select="$code-point"/>
+        <xsl:sequence select="eo:decode-bytes(subsequence($bytes, 3))"/>
+      </xsl:when>
+      <!-- 3-byte sequence: 1110xxxx 10xxxxxx 10xxxxxx -->
+      <xsl:when test="$bytes[1] ge 224 and $bytes[1] lt 240">
+        <xsl:variable name="code-point" select="(($bytes[1] - 224) * 4096) + (($bytes[2] - 128) * 64) + ($bytes[3] - 128)"/>
+        <xsl:sequence select="$code-point"/>
+        <xsl:sequence select="eo:decode-bytes(subsequence($bytes, 4))"/>
+      </xsl:when>
+      <!-- 4-byte sequence: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx -->
+      <xsl:when test="$bytes[1] ge 240 and $bytes[1] lt 248">
+        <xsl:variable name="code-point" select="(($bytes[1] - 240) * 262144) + (($bytes[2] - 128) * 4096) + (($bytes[3] - 128) * 64) + ($bytes[4] - 128)"/>
+        <xsl:sequence select="$code-point"/>
+        <xsl:sequence select="eo:decode-bytes(subsequence($bytes, 5))"/>
+      </xsl:when>
+      <!-- Otherwise, return empty (should not occur if input is valid UTF-8) -->
+      <xsl:otherwise/>
+    </xsl:choose>
+  </xsl:function>
+  <!-- Function to convert integer to 4-digit hex string -->
+  <xsl:function name="eo:int-to-hex" as="xs:string">
+    <xsl:param name="value" as="xs:integer"/>
+    <xsl:variable name="hex-chars" select="'0123456789ABCDEF'"/>
+    <xsl:variable name="hex" select="concat(substring($hex-chars, floor($value idiv 4096) + 1, 1), substring($hex-chars, floor(($value mod 4096) idiv 256) + 1, 1), substring($hex-chars, floor(($value mod 256) idiv 16) + 1, 1), substring($hex-chars, ($value mod 16) + 1, 1))"/>
+    <xsl:sequence select="$hex"/>
+  </xsl:function>
+  <xsl:function name="eo:hex-to-utf8" as="xs:integer">
+    <xsl:param name="hex" as="xs:string"/>
+    <xsl:variable name="hex-upper" select="upper-case(normalize-space($hex))"/>
+    <xsl:variable name="length" select="string-length($hex-upper)"/>
+    <xsl:variable name="hex-digits" select="string-to-codepoints('0123456789ABCDEF')"/>
+    <xsl:variable name="decimal" select="sum(for $i in 1 to $length return (index-of($hex-digits, string-to-codepoints(substring($hex-upper, $i, 1))) - 1) * math:pow(16, $length - $i))"/>
+    <xsl:value-of select="xs:int($decimal)"/>
+  </xsl:function>
 </xsl:stylesheet>