Groupies working hard! :D

A0/file.c

@@ -8,12 +8,15 @@ enum FileType{ASCII, ISO, UTF, data};
 
 bool is_utf8(FILE* f) {
     rewind(f);
-    char byte;
+    unsigned char byte;
     int char_length = -1;
     for (;;) {
         if (fread(&byte, 1, 1, f) == 0) {
             break;
         }
+        if (byte == 0){
+            return false;
+        }
         for (int j = 0 ; j <= 8 ; j++) {
             if (byte >> 7 == 0) {
                 char_length = j;
@@ -48,20 +51,20 @@ bool is_iso8859(FILE* f) {
     for (int i = 0 ; i < 95 ; i++) {
         iso8859_CHARACTERS[i+102] = i+160;
     }
-    char byte;
+    unsigned char byte;
-    bool is_iso8859;
+    bool iso_byte;
     for (;;) {
         if (fread(&byte, 1, 1, f) == 0) {
             break;
         }
-        is_iso8859 = false;
+        iso_byte = false;
-        for (int i = 0 ; i < 95 ; i++) {
+        for (int i = 0 ; i < 197 ; i++) {
             if (byte == iso8859_CHARACTERS[i]) {
-                is_iso8859 = true;
+                iso_byte = true;
                 break;
             }
         }
-        if (! is_iso8859) {
+        if (! iso_byte) {
             return false;
         }
 
@@ -81,7 +84,7 @@ bool is_ascii(FILE* f) {
         ASCII_CHARACTERS[i+8] = i+32;
     }
 
-    char byte;
+    unsigned char byte;
     bool ascii_byte;
     for (;;) {
         if (fread(&byte, 1, 1, f) == 0) {
@@ -126,6 +129,8 @@ int main(int argc, char* argv[]) {
         fprintf(stdout, "%s: ASCII text\n", argv[1]);
     } else if (is_iso8859(f)) {
         fprintf(stdout, "%s: ISO-8859 text\n", argv[1]);
+    } else if (is_utf8(f)) {
+        fprintf(stdout, "%s: UTF-8 Unicode text\n", argv[1]);
     } else {
         fprintf(stdout, "%s: data\n", argv[1]);
     }

A0/test.sh

@@ -12,26 +12,28 @@ rm -f test_files/*
 
 
 echo "Generating test files.."
-printf "Hello, World!\n" > test_files/ascii.input
+printf "Hello, World\x04!\n" > test_files/ascii.input
 printf "Hello, World!" > test_files/ascii2.input
+printf "Test file\nTest\nThis is a test file\n\n\n\n" > test_files/ascii3.input
 printf "Hello,\x00World!\n" > test_files/data.input
-printf "Hællo,\x00World!\n" > test_files/iso1.input
+printf "Hello,\x00World!\n" > test_files/data2.input
-printf "Hello,\x00Wørld!\n" > test_files/iso2.input
+printf "Hello,\xa0World!\n" > test_files/iso1.input
-printf "Hello,\x00World!\n" > test_files/utf81.input
+printf "Hello,\xa1\xbbWorld!\n" > test_files/iso2.input
-printf "Hello,\x00World!\n" > test_files/utf82.input
+printf "Hello, \xaaWorld!\n" > test_files/iso3.input
+printf "\x24\xe0\xa4\xb9\xf0\x90\x8d\x88\n" > test_files/utf81.input
+printf "مرحبا عالمي\n" > test_files/utf82.input
+printf "Hello, World or should I say, 안녕하세요 세계!\n" > test_files/utf83.input
 printf "" > test_files/empty.input
-### TODO: Generate more test files ###
-
 
 echo "Running the tests.."
 exitcode=0
-for f in test_files/*.input
+for f in test_files/*.input # For loop
 do
-  echo ">>> Testing ${f}.."
+  echo ">>> Testing ${f}.." # print
-  file    "${f}" | sed 's/ASCII text.*/ASCII text/' > "${f}.expected"
+  file    "${f}" | sed 's/ASCII text.*/ASCII text/' > "${f}.expected" # fjerner tekst der kun bliver givet ved ASCII filer og gemmer svaret i ".expected" filen.
-  ./file  "${f}" > "${f}.actual"
+  ./file  "${f}" > "${f}.actual" # giver vores *egen* kodes svar i ".actual" filen.
 
-  if ! diff -u "${f}.expected" "${f}.actual"
+  if ! diff -u "${f}.expected" "${f}.actual" # sammenligner vores egen kodes svar og "file" kommandoens svar
   then
     echo ">>> Failed :-("
     exitcode=1