From ffd46fd134188c423054645fc0a73af8223f9b29 Mon Sep 17 00:00:00 2001
From: tmikolov <tmikolov@c84ef02e-58a5-4c83-e53e-41fc32d635eb>
Date: Mon, 8 Sep 2014 20:14:01 +0000
Subject: [PATCH] fixed minor bugs

---
 compute-accuracy.c | 8 +++++++-
 distance.c         | 8 +++++++-
 word-analogy.c     | 8 +++++++-
 word2vec.c         | 4 ++--
 4 files changed, 23 insertions(+), 5 deletions(-)

diff --git a/compute-accuracy.c b/compute-accuracy.c
index d83fcbb..61c31ac 100644
--- a/compute-accuracy.c
+++ b/compute-accuracy.c
@@ -53,7 +53,13 @@ int main(int argc, char **argv)
     return -1;
   }
   for (b = 0; b < words; b++) {
-    fscanf(f, "%s%c", &vocab[b * max_w], &ch);
+    a = 0;
+    while (1) {
+      vocab[b * max_w + a] = fgetc(f);
+      if (feof(f) || (vocab[b * max_w + a] == ' ')) break;
+      if ((a < max_w) && (vocab[b * max_w + a] != '\n')) a++;
+    }
+    vocab[b * max_w + a] = 0;
     for (a = 0; a < max_w; a++) vocab[b * max_w + a] = toupper(vocab[b * max_w + a]);
     for (a = 0; a < size; a++) fread(&M[a + b * size], sizeof(float), 1, f);
     len = 0;
diff --git a/distance.c b/distance.c
index fbeb24a..731999f 100644
--- a/distance.c
+++ b/distance.c
@@ -51,7 +51,13 @@ int main(int argc, char **argv) {
     return -1;
   }
   for (b = 0; b < words; b++) {
-    fscanf(f, "%s%c", &vocab[b * max_w], &ch);
+    a = 0;
+    while (1) {
+      vocab[b * max_w + a] = fgetc(f);
+      if (feof(f) || (vocab[b * max_w + a] == ' ')) break;
+      if ((a < max_w) && (vocab[b * max_w + a] != '\n')) a++;
+    }
+    vocab[b * max_w + a] = 0;
     for (a = 0; a < size; a++) fread(&M[a + b * size], sizeof(float), 1, f);
     len = 0;
     for (a = 0; a < size; a++) len += M[a + b * size] * M[a + b * size];
diff --git a/word-analogy.c b/word-analogy.c
index ea91ab1..8c486f6 100644
--- a/word-analogy.c
+++ b/word-analogy.c
@@ -50,7 +50,13 @@ int main(int argc, char **argv) {
     return -1;
   }
   for (b = 0; b < words; b++) {
-    fscanf(f, "%s%c", &vocab[b * max_w], &ch);
+    a = 0;
+    while (1) {
+      vocab[b * max_w + a] = fgetc(f);
+      if (feof(f) || (vocab[b * max_w + a] == ' ')) break;
+      if ((a < max_w) && (vocab[b * max_w + a] != '\n')) a++;
+    }
+    vocab[b * max_w + a] = 0;
     for (a = 0; a < size; a++) fread(&M[a + b * size], sizeof(float), 1, f);
     len = 0;
     for (a = 0; a < size; a++) len += M[a + b * size] * M[a + b * size];
diff --git a/word2vec.c b/word2vec.c
index 6763cfa..b9d3dc6 100644
--- a/word2vec.c
+++ b/word2vec.c
@@ -152,9 +152,9 @@ void SortVocab() {
   train_words = 0;
   for (a = 0; a < size; a++) {
     // Words occuring less than min_count times will be discarded from the vocab
-    if (vocab[a].cn < min_count) {
+    if ((vocab[a].cn < min_count) && (a != 0)) {
       vocab_size--;
-      free(vocab[vocab_size].word);
+      free(vocab[a].word);
     } else {
       // Hash will be re-computed, as after the sorting it is not actual
       hash=GetWordHash(vocab[a].word);