custom_uniq: uniq -u buggy? workaround

2 files changed, 90 insertions, 0 deletions

diff --git a/Makefile b/Makefile
index 3deee66..b805715 100644
--- a/Makefile
+++ b/Makefile
@@ -12,12 +12,14 @@ CFLAGS += -I include
 
 
 all: \
+	out/custom_uniq \
 	out/mem_internal_check \
 	out/split_for_sort \
 	out/tree_based_check
 
 
 debug: \
+	out/debug/custom_uniq \
 	out/debug/mem_internal_check \
 	out/debug/split_for_sort \
 	out/debug/tree_based_check
@@ -38,6 +40,12 @@ out/debug: out
 
 # Lazy, not splitting compiler & linker calls
 
+out/custom_uniq: out src/custom_uniq.c include/trace_macros.h
+	${CC} -o $@ ${CFLAGS} ${PROD_CFLAGS} src/custom_uniq.c
+
+out/debug/custom_uniq: out/debug src/custom_uniq.c include/trace_macros.h
+	${CC} -o $@ ${CFLAGS} ${DEBUG_CFLAGS} src/custom_uniq.c
+
 out/mem_internal_check: out src/mem_internal_check.c \
 	include/trace_macros.h include/hex_conversion.h
 	${CC} -o $@ ${CFLAGS} ${PROD_CFLAGS} src/mem_internal_check.c
diff --git a/src/custom_uniq.c b/src/custom_uniq.c
new file mode 100644
index 0000000..82ce3da
--- /dev/null
+++ b/src/custom_uniq.c
@@ -0,0 +1,82 @@
+/* SPDX-License-Identifier: Apache-2.0 */
+
+/* Copyright 2025 Thorsten Töpper
+ *
+ * I noticed GNU uniq -u not always removed all overlaps, instead
+ * of spending hours in creating a reproducable test case and code
+ * review, a simple custom implementation of that mode.
+ *
+ * vim:ts=4:sw=4:expandtab
+ */
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <errno.h>
+
+#include "trace_macros.h"
+
+#define LINE_LENGTH 4096
+
+/* prev=(cur)?0:1; less cycles? */
+#define INDEXSWITCH cur=(cur+1)%2; prev=(prev+1)%2;
+
+int main(int argc, char **argv) {
+    FILE *fdin = stdin;
+    char *lines[2];
+    int cur = 1, prev = 0;
+    bool duplicate = false;
+
+    if (argc > 1) {
+        if (strncmp(argv[1], "-", 2) != 0) {
+            if ((fdin=fopen(argv[1], "r")) == NULL) {
+                LOGERR("ERROR: Failed to open file %s: %s\n",
+                        argv[1], strerror(errno));
+                return EXIT_FAILURE;
+            }
+        }
+    }
+
+    lines[0] = calloc(LINE_LENGTH, sizeof(char));
+    lines[1] = calloc(LINE_LENGTH, sizeof(char));
+    if (lines[0] == NULL || lines[1] == NULL) {
+        LOGERR("ERROR: Failed to allocate 8kiB...\n");
+        return EXIT_FAILURE;
+    }
+
+    /* preparation */
+    if (fgets(lines[prev], LINE_LENGTH, fdin) == NULL) {
+        LOGERR("ERROR: Failed to read input line.\n");
+        return EXIT_FAILURE;
+    }
+
+    /* parsing input */
+    while (fgets(lines[cur], LINE_LENGTH, fdin) != NULL) {
+        if (strcmp(lines[cur], lines[prev]) == 0) {
+            duplicate = true;
+            continue;
+        }
+
+        if ( ! duplicate ) {
+            fputs(lines[prev], stdout);
+        } else {
+            duplicate = false;
+        }
+
+        INDEXSWITCH;
+    }
+
+    /* last line */
+    if ( ! duplicate ) {
+        fputs(lines[prev], stdout);
+    }
+
+
+    if (fdin != stdin)
+        fclose(fdin);
+
+    return EXIT_SUCCESS;
+}
+