From 52f28c38d0d1d7d6036495b3ecadf3cd82dd1e50 Mon Sep 17 00:00:00 2001
From: Lieuwe Rooijakkers <lieuwerooijakkers@gmail.com>
Date: Mon, 19 Aug 2024 01:02:26 +0200
Subject: toilet: don't read whole file in memory

---
 src/toilet.c | 96 +++++++++++++++++++++++++++++++++++++-----------------------
 1 file changed, 60 insertions(+), 36 deletions(-)

(limited to 'src')

diff --git a/src/toilet.c b/src/toilet.c
index de32003..74cafb2 100644
--- a/src/toilet.c
+++ b/src/toilet.c
@@ -1,9 +1,11 @@
 #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>
@@ -11,7 +13,7 @@
 
 #include "util/debug.h"
 #include "util/error.h"
-#include "util/loop_files.h"
+#include "util/loop_args.h"
 #include "util/map.h"
 #include "util/versie.h"
 
@@ -73,66 +75,88 @@ static char** parse_options(int argc, char **argv, int *modeMap) {
   return argv + optind;
 }
 
-static size_t get_count(enum MODE mode, struct filebuf *fb) {
-  switch (mode) {
-    case M_BYTES:
-      return fb->sz;
+static size_t count_lines(char *fname, FILE *f) {
+  size_t nlines = 0;
 
-    case M_WORDS: {
-#define IN(i) (i < fb->sz)
+  {
+    char *line = NULL;
+    size_t linen = 0;
+    while ((errno = 0, getline(&line, &linen, f)) != -1) {
+      nlines++;
+    }
+    free(line);
+  }
 
-      size_t words = 0;
+  if (errno != 0) {
+    printf("toilet: fout bij lezen uit bestand '%s'\n", fname);
+    exit(1);
+  }
 
-      // (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++;
-      }
+  rewind(f);
+  return nlines;
+}
 
-      return words;
+static size_t count_words(FILE *f) {
+#define BUF_SIZE 4096
 
-#undef IN
+  size_t nwords = 0;
+  static char buf[BUF_SIZE];
+
+  while (!feof(f)) {
+    const size_t n = fread(buf, 1, BUF_SIZE, f);
+
+#define IN(i) (i < n)
+    for (size_t i = 0; IN(i);) {
+      size_t previ = i;
+      while (IN(i) && !isspace(buf[i])) i++;
+      nwords += i != previ;
+      while (IN(i) && isspace(buf[i])) i++;
     }
+#undef IN
 
-    case M_LINES: {
-      size_t lines = 0;
-      size_t i = 0;
+  }
 
-      while (i != fb->sz) {
-        if (fb->buf[i] == '\n') lines++;
-        i++;
-      }
+  rewind(f);
+  return nwords;
 
-      // handle case if file does not have trailing newline
-      if (fb->buf[i - 1] != '\n') {
-        lines++;
-      }
+#undef BUF_SIZE
+}
 
-      return lines;
+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;
     }
 
+    case M_WORDS:
+      return count_words(f);
+
+    case M_LINES:
+      return count_lines(fname, f);
+
     default:
       assert(false);
   }
 }
 
-static int process(struct filebuf *fb, char *fname, bool) {
+static int process(char *fname, bool isstdin) {
+  FILE *f = isstdin ? stdin : fopen(fname, "r");
+
   for (enum MODE mode = 1; mode <= M_BYTES; mode <<= 1) {
     if (mode & modeMap) {
-      const size_t count = get_count(mode, fb);
+      const size_t count = get_count(mode, fname, f);
       printf("%li ", count);
     }
   }
   printf("%s\n", fname);
-  free_filebuf(fb);
+
+  if (!isstdin) fclose(f);
   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);
@@ -140,5 +164,5 @@ int entry_toilet(int argc, char **argv) {
     modeMap = INT_MAX;
   }
 
-  return loop_files(args, process);
+  return loop_args(args, process);
 }
-- 
cgit v1.2.3-70-g09d2