checksrc

commit: 8a2982b7c9109eb3a0336f958015a95cd314aeda
parent 2979f91180c081b6e2e8e1f591ba21120b4a434d
Author: Haelwenn (lanodan) Monnier <contact@hacktivis.me>
Date:   Wed, 24 Sep 2025 07:20:38 +0200

check amount of punctuation+numbers vs. letters

Diffstat:
M
README.md
2
+-
M
checksrc.c
40
++++++++++++++++++++++++++++++++++------

2 files changed, 35 insertions(+), 7 deletions(-)
diff --git a/README.md b/README.md
@@ -50,7 +50,7 @@ So `checksrc` can take much longer depending on the payloads.
 Throws a warning but doesn't stops reading the file:
 
 - dump: block of 10 consecutive lines with lengths varying by less than 3 bytes (to detect hex dumps, base64, …)
-- (planned) more punctuation, symbols, and numbers than letters [a-zA-Z\x7C-\xFF] (within 4KB blocks)
+- more punctuation, symbols, and numbers than letters (within 4KB blocks)
 
 ### Major
 Throws an error, stops reading the file, exits unsuccessfully:
diff --git a/checksrc.c b/checksrc.c
@@ -1,15 +1,16 @@
 // SPDX-FileCopyrightText: 2017 Haelwenn (lanodan) Monnier <contact+checksrc@hacktivis.me>
 // SPDX-License-Identifier: MPL-2.0
 #define _DEFAULT_SOURCE
+#include <ctype.h>
 #include <dirent.h>
 #include <errno.h>
 #include <fcntl.h>
 #include <limits.h> // PATH_MAX
+#include <stdbool.h>
+#include <stdint.h> // uint8_t
 #include <stdio.h>
 #include <string.h>
 #include <unistd.h>
-#include <stdbool.h>
-#include <stdint.h> // uint8_t
 
 #if 0
 static const char *generated[] = {
@@ -48,10 +49,12 @@ checkfile(const char *fname)
 	size_t lines_len[10];
 	size_t chars = 0, line = 0;
 	size_t last_dump = 0;
-	int dumps = 0;
+	unsigned int dumps = 0;
+	off_t pos = 0;
 	for(;;)
 	{
 		static uint8_t buf[4096];
+		unsigned int puncts = 0, numbers = 0, letters = 0;
 
 		int nread = read(fd, &buf, 4096);
 		if(nread == 0) break;
@@ -113,8 +116,11 @@ checkfile(const char *fname)
 				continue;
 			}
 
-			/* control characters. Except 0x1B ('^[', ESC). With \n, \r taken care off earlier */
-			if(buf[i] < ' ' && buf[i] != 0x1B)
+			/* ('\e', ESC) */
+			if(buf[i] == 0x1B) continue;
+
+			/* control characters. With \r, \t, \n, \e taken care off earlier */
+			if(buf[i] < ' ')
 			{
 				printf("%s: binary (byte 0x%X found at position 0x%zX)\n", fname, buf[i], i);
 				close(fd);
@@ -126,8 +132,30 @@ checkfile(const char *fname)
 			{
 				chars++;
 				lines_len[line % 10]++;
+
+				if(buf[i] == ' ') continue;
+
+				if(isdigit(buf[i]))
+					numbers++;
+				else if(isalpha(buf[i]))
+					letters++;
+				else
+					puncts++;
 			}
 		}
+
+		if(puncts + numbers > letters)
+		{
+			printf("%s (block 0x%lX to 0x%lX): Got more punctuation(%d) + numbers(%d) than letters(%d)\n",
+			       fname,
+			       pos,
+			       pos + nread,
+			       puncts,
+			       numbers,
+			       letters);
+		}
+
+		pos += nread;
 	}
 
 	// TODO: Check for a "generated" string
@@ -214,7 +242,7 @@ checkdir(const char *dirname)
 					continue;
 				}
 				if(err < 0) return 1;
-				if(err > 0) ret = 1+err;
+				if(err > 0) ret = 1 + err;
 			}
 		}
 	}