From d93e07620b6b09ea341031fe3a1aae0d2d97714e Mon Sep 17 00:00:00 2001
From: Lieuwe Rooijakkers <lieuwerooijakkers@gmail.com>
Date: Mon, 19 Aug 2024 01:50:35 +0200
Subject: Revert "toilet: fix segfault in word count"

This reverts back to commit 7298453390da42d4fcd21d2911e2d14b2187ceff.
---
 src/toilet.c | 105 +++++++++++++++++++++--------------------------------------
 1 file changed, 38 insertions(+), 67 deletions(-)

(limited to 'src/toilet.c')

diff --git a/src/toilet.c b/src/toilet.c
index ff51ba8..de32003 100644
--- a/src/toilet.c
+++ b/src/toilet.c
@@ -1,11 +1,9 @@
 #include <assert.h>
 #include <ctype.h>
-#include <errno.h>
 #include <getopt.h>
 #include <limits.h>
 #include <pwd.h>
 #include <stdbool.h>
-#include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <sys/types.h>
@@ -13,7 +11,7 @@
 
 #include "util/debug.h"
 #include "util/error.h"
-#include "util/loop_args.h"
+#include "util/loop_files.h"
 #include "util/map.h"
 #include "util/versie.h"
 
@@ -21,8 +19,6 @@
 
 static int modeMap;
 
-#define BUF_SIZE 4096
-
 static void usage(FILE *f) {
   fprintf(f,
       "Gebruik: toilet [-nchV] [BESTAND]...\n"
@@ -77,91 +73,66 @@ static char** parse_options(int argc, char **argv, int *modeMap) {
   return argv + optind;
 }
 
-static size_t count_lines(char *fname, FILE *f) {
-  size_t nlines = 0;
-  bool last_was_nl = false;
-
-  while (!feof(f)) {
-    static char buf[BUF_SIZE];
-    const size_t n = fread(buf, 1, BUF_SIZE, f);
-
-    for (size_t i = 0; i < n; i++) {
-      last_was_nl = false;
-      if (buf[i] == '\n') {
-        nlines++;
-        last_was_nl = true;
-      }
-    }
-  }
+static size_t get_count(enum MODE mode, struct filebuf *fb) {
+  switch (mode) {
+    case M_BYTES:
+      return fb->sz;
 
-  if (!last_was_nl) nlines++;
+    case M_WORDS: {
+#define IN(i) (i < fb->sz)
 
-  if (errno != 0) {
-    printf("toilet: fout bij lezen uit bestand '%s'\n", fname);
-    exit(1);
-  }
+      size_t words = 0;
 
-  rewind(f);
-  return nlines;
-}
-
-static size_t count_words(FILE *f) {
-  size_t nwords = 0;
+      // (c) Tom Forging
+      for (size_t i = 0; IN(i);) {
+        size_t previ = i;
+        while (IN(i) && !isspace(fb->buf[i])) i++;
+        words += i != previ;
+        while (IN(i) && isspace(fb->buf[i])) i++;
+      }
 
-  while (!feof(f)) {
-    static char buf[BUF_SIZE];
-    const size_t n = fread(buf, 1, BUF_SIZE, f);
+      return words;
 
-    // (c) Tom Forging
-    for (size_t i = 0; i<n;) {
-      size_t previ = i;
-      while (i<n && !isspace(buf[i])) i++;
-      nwords += i != previ;
-      while (i<n && isspace(buf[i])) i++;
+#undef IN
     }
-  }
 
-  rewind(f);
-  return nwords;
-}
+    case M_LINES: {
+      size_t lines = 0;
+      size_t i = 0;
 
-static size_t get_count(enum MODE mode, char *fname, FILE *f) {
-  switch (mode) {
-    case M_BYTES: {
-      fseek(f, 0, SEEK_END);
-      long offset = ftell(f);
-      rewind(f);
-      return offset;
-    }
+      while (i != fb->sz) {
+        if (fb->buf[i] == '\n') lines++;
+        i++;
+      }
 
-    case M_WORDS:
-      return count_words(f);
+      // handle case if file does not have trailing newline
+      if (fb->buf[i - 1] != '\n') {
+        lines++;
+      }
 
-    case M_LINES:
-      return count_lines(fname, f);
+      return lines;
+    }
 
     default:
       assert(false);
   }
 }
 
-static int process(char *fname, bool isstdin) {
-  FILE *f = isstdin ? stdin : fopen(fname, "rb");
-
+static int process(struct filebuf *fb, char *fname, bool) {
   for (enum MODE mode = 1; mode <= M_BYTES; mode <<= 1) {
     if (mode & modeMap) {
-      const size_t count = get_count(mode, fname, f);
+      const size_t count = get_count(mode, fb);
       printf("%li ", count);
     }
   }
-
-  if (fname != NULL) printf("%s", fname);
-  putchar('\n');
-
-  if (!isstdin) fclose(f);
+  printf("%s\n", fname);
+  free_filebuf(fb);
   return 0;
 }
 
+// TODO: be smarter, toilet doesn't have to read the whole file in memory (for
+// unmappable files)
+
 int entry_toilet(int argc, char **argv) {
   modeMap = 0;
   char **args = parse_options(argc, argv, &modeMap);
@@ -169,5 +140,5 @@ int entry_toilet(int argc, char **argv) {
     modeMap = INT_MAX;
   }
 
-  return loop_args(args, process);
+  return loop_files(args, process);
 }
-- 
cgit v1.2.3-70-g09d2