Add jau::codec::base::ascii38_alphabet: Drop lower case letters suitable for unique vfat filename and use it for temp-dir

Base 38 is still big enough to provide > INT_MAX @ 6 digits.

5 files changed, 98 insertions, 6 deletions

diff --git a/include/jau/base_codec.hpp b/include/jau/base_codec.hpp
index 62e74c9..d5d9e3c 100644
--- a/include/jau/base_codec.hpp
+++ b/include/jau/base_codec.hpp
@@ -256,6 +256,45 @@ namespace jau::codec::base {
     };
 
     /**
+     * Safe base 38 alphabet with ASCII code-point sorting order.
+     *
+     * - Value: `-0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_`
+     * - Padding: `=`
+     *
+     * ### Properties
+     * - 7-bit ASCII
+     * - Code page 437 compatible
+     * - Safe URL and filename use
+     * - Excludes forbidden [v]fat chars: `<>:"/\|?*`
+     * - Only using upper-case letters for unique filename under vfat
+     * - Excludes quoting chars: "'$ and space
+     * - Supporting ASCII code-point sorting.
+     * - Order: `-` < `0` < `A` < `a` < `z`
+     */
+    class ascii38_alphabet : public alphabet {
+        private:
+            static inline constexpr const std::string_view data = "-0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_";
+
+            static int s_code_point(const char c) noexcept {
+                if ('0' <= c && c <= '9') {
+                    return c - '0' + 1;
+                } else if ('A' <= c && c <= 'Z') {
+                    return c - 'A' + 11;
+                } else if ('-' == c) {
+                    return 0;
+                } else if ('_' == c) {
+                    return 37;
+                } else {
+                    return -1;
+                }
+            }
+
+        public:
+            ascii38_alphabet() noexcept
+            : alphabet("ascii38", 38, data, '=', s_code_point) {}
+    };
+
+    /**
      * Safe base 64 alphabet with ASCII code-point sorting order.
      *
      * - Value: `-0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz`
diff --git a/java_base/org/jau/util/BaseCodec.java b/java_base/org/jau/util/BaseCodec.java
index 500eec7..b7c8f47 100644
--- a/java_base/org/jau/util/BaseCodec.java
+++ b/java_base/org/jau/util/BaseCodec.java
@@ -246,6 +246,48 @@ public class BaseCodec {
     }
 
     /**
+     * Safe base 38 alphabet with ASCII code-point sorting order.
+     *
+     * - Value: `-0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_`
+     * - Padding: `=`
+     *
+     * ### Properties
+     * - 7-bit ASCII
+     * - Code page 437 compatible
+     * - Safe URL and filename use
+     * - Excludes forbidden [v]fat chars: `<>:"/\|?*`
+     * - Only using upper-case letters for unique filename under vfat
+     * - Excludes quoting chars: "'$ and space
+     * - Supporting ASCII code-point sorting.
+     * - Order: `-` < `0` < `A` < `a` < `z`
+     *
+     * @see encodeBase()
+     * @see decodeBase()
+     */
+    public static class Ascii38Alphabet extends Alphabet {
+        private static final String data = "-0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_";
+
+        @Override
+        public int code_point(final char c) {
+            if ('0' <= c && c <= '9') {
+                return c - '0' + 1;
+            } else if ('A' <= c && c <= 'Z') {
+                return c - 'A' + 11;
+            } else if ('-' == c) {
+                return 0;
+            } else if ('_' == c) {
+                return 37;
+            } else {
+                return -1;
+            }
+        }
+
+        public Ascii38Alphabet() {
+            super("ascii38", 38, data, '=');
+        }
+    }
+
+    /**
      * Safe base 64 alphabet with ASCII code-point sorting order.
      *
      * - Value: `-0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz`
diff --git a/src/file_util.cpp b/src/file_util.cpp
index 3e8cb23..9fdaf68 100644
--- a/src/file_util.cpp
+++ b/src/file_util.cpp
@@ -1531,7 +1531,7 @@ static bool copy_push_mkdir(const file_stats& dst_stats, copy_context_t& ctx) no
     } else if( !dst_stats.exists() ) {
         new_dir = true;
         constexpr const int32_t val_min = 888;
-        constexpr const int32_t val_max = std::numeric_limits<int32_t>::max(); // 6 digits base 64 > INT_MAX
+        constexpr const int32_t val_max = std::numeric_limits<int32_t>::max(); // 6 digits base 38 > INT_MAX
         uint64_t mkdir_cntr = 0;
         std::mt19937_64 prng;
         std::uniform_int_distribution<int32_t> prng_dist(val_min, val_max);
@@ -1539,7 +1539,7 @@ static bool copy_push_mkdir(const file_stats& dst_stats, copy_context_t& ctx) no
         do {
             ++mkdir_cntr;
             const int32_t val_d = prng_dist(prng);
-            basename_ = "."+jau::codec::base::encode(val_d, 64, jau::codec::base::ascii64_alphabet(), 6); // base 64, 6 digits
+            basename_ = "."+jau::codec::base::encode(val_d, 38, jau::codec::base::ascii38_alphabet(), 6); // base 38, 6 digits
             if( 0 == ::mkdirat(dest_dirfd, basename_.c_str(), jau::fs::posix_protection_bits(fmode_t::rwx_usr)) ) {
                 mkdir_ok = true;
             } else if (errno != EINTR && errno != EEXIST) {
diff --git a/test/java/jau/test/util/TestBaseCodec.java b/test/java/jau/test/util/TestBaseCodec.java
index ca164f3..d5ea37e 100644
--- a/test/java/jau/test/util/TestBaseCodec.java
+++ b/test/java/jau/test/util/TestBaseCodec.java
@@ -187,7 +187,13 @@ public class TestBaseCodec {
     }
 
     @Test
-    public void test01IntegerBase64() {
+    public void test01IntegerBase38() {
+        testRadix_3digits_int32(38, new BaseCodec.Ascii38Alphabet());
+        testRadix_3digits_int32(38, new BaseCodec.Ascii64Alphabet());
+    }
+
+    @Test
+    public void test02IntegerBase64() {
         testIntegerBase64(new BaseCodec.Base64Alphabet());
         testIntegerBase64(new BaseCodec.Base64urlAlphabet());
         testIntegerBase64(new BaseCodec.Natural86Alphabet());
@@ -196,7 +202,7 @@ public class TestBaseCodec {
     }
 
     @Test
-    public void test02IntegerBase86() {
+    public void test03IntegerBase86() {
         testIntegerBase86(new BaseCodec.Natural86Alphabet());
         testIntegerBase86(new BaseCodec.Ascii86Alphabet());
     }
diff --git a/test/test_codec_base01.cpp b/test/test_codec_base01.cpp
index 86063e9..e5f9f2b 100644
--- a/test/test_codec_base01.cpp
+++ b/test/test_codec_base01.cpp
@@ -183,7 +183,12 @@ static void testIntegerBase86(const jau::codec::base::alphabet& aspec) {
     // testRadix_int64(86, aspec, 0x0_i64, 0x7FFFFFFFFFFFFFFF_i64);
 }
 
-TEST_CASE( "Integer Base 64 Encoding Test 01", "[integer][type]" ) {
+TEST_CASE( "Integer Base 38 Encoding Test 01", "[integer][type]" ) {
+    testRadix_3digits_int32(38, jau::codec::base::ascii38_alphabet());
+    testRadix_3digits_int32(38, jau::codec::base::ascii64_alphabet());
+}
+
+TEST_CASE( "Integer Base 64 Encoding Test 02", "[integer][type]" ) {
     testIntegerBase64(jau::codec::base::base64_alphabet());
     testIntegerBase64(jau::codec::base::base64url_alphabet());
     testIntegerBase64(jau::codec::base::natural86_alphabet());
@@ -191,7 +196,7 @@ TEST_CASE( "Integer Base 64 Encoding Test 01", "[integer][type]" ) {
     testIntegerBase64(jau::codec::base::ascii86_alphabet());
 }
 
-TEST_CASE( "Integer Base 86 Encoding Test 02", "[integer][type]" ) {
+TEST_CASE( "Integer Base 86 Encoding Test 03", "[integer][type]" ) {
     testIntegerBase86(jau::codec::base::natural86_alphabet());
     testIntegerBase86(jau::codec::base::ascii86_alphabet());
 }