From 73dd2967c8e1e4f6d7334ee9e539a323d6e66cbd Mon Sep 17 00:00:00 2001
From: Bartek Fabiszewski <bfabiszewski@users.noreply.github.com>
Date: Mon, 27 Nov 2017 12:19:26 +0100
Subject: [PATCH] Discard too long lines in dictionary file (#14)

* Discard too long lines in dictionary file

* Discard too long lines in dictionary file: add warning and test
---
 hyphen.c             | 22 ++++++++++++++++++----
 tests/Makefile.am    |  5 +++++
 tests/longlines.hyph |  1 +
 tests/longlines.pat  | 10 ++++++++++
 tests/longlines.test |  4 ++++
 tests/longlines.word |  1 +
 6 files changed, 39 insertions(+), 4 deletions(-)
 create mode 100644 tests/longlines.hyph
 create mode 100644 tests/longlines.pat
 create mode 100755 tests/longlines.test
 create mode 100644 tests/longlines.word

diff --git a/hyphen.c b/hyphen.c
index 9a132d0..9f2b711 100644
--- a/hyphen.c
+++ b/hyphen.c
@@ -438,11 +438,25 @@ for (k = 0; k < 2; k++) {
   }
 
   if (k == 0 || nextlevel) {
-    while (fgets (buf, sizeof(buf), f) != NULL) {
+    while (fgets(buf, sizeof(buf), f) != NULL) {
+      
+      /* discard lines that don't fit in buffer */
+      if (!feof(f) && strchr(buf, '\n') == NULL) {
+        int c;
+        while ((c = fgetc(f)) != '\n' && c != EOF);
+        /* issue warning if not a comment */
+        if (buf[0] != '%') {
+          fprintf(stderr, "Warning: skipping too long pattern (more than %lu chars)\n", sizeof(buf));
+        }
+        continue;
+      }
+      
       if (strncmp(buf, "NEXTLEVEL", 9) == 0) {
-	nextlevel = 1;
-	break;
-      } else if (buf[0] != '%') hnj_hyphen_load_line(buf, dict[k], hashtab);
+        nextlevel = 1;
+        break;
+      } else if (buf[0] != '%') {
+        hnj_hyphen_load_line(buf, dict[k], hashtab);
+      }
     }
   } else if (k == 1) {
     /* default first level: hyphen and ASCII apostrophe */
diff --git a/tests/Makefile.am b/tests/Makefile.am
index 2b08c67..e72863c 100644
--- a/tests/Makefile.am
+++ b/tests/Makefile.am
@@ -11,6 +11,7 @@ alt4.test \
 alt5.test \
 alt6.test \
 alt7.test \
+longlines.test \
 compound.test \
 compound2.test \
 compound3.test \
@@ -50,6 +51,9 @@ alt6.word \
 alt7.hyph \
 alt7.pat \
 alt7.word \
+longlines.hyph \
+longlines.pat \
+longlines.word \
 alt.hyph \
 alt.pat \
 alt.word \
@@ -105,6 +109,7 @@ alt4.test \
 alt5.test \
 alt6.test \
 alt7.test \
+longlines.test \
 alt.test \
 basealt2.test \
 basealt.test \
diff --git a/tests/longlines.hyph b/tests/longlines.hyph
new file mode 100644
index 0000000..8b0ee7f
--- /dev/null
+++ b/tests/longlines.hyph
@@ -0,0 +1 @@
+a=bc=d
diff --git a/tests/longlines.pat b/tests/longlines.pat
new file mode 100644
index 0000000..290aef4
--- /dev/null
+++ b/tests/longlines.pat
@@ -0,0 +1,10 @@
+ISO8859-1
+LEFTHYPHENMIN 1
+RIGHTHYPHENMIN 1
+% Check whether characters over MAX_CHARS are not treated as new line
+% This test is valid as long as MAX_CHARS is 100
+%
+% Following pattern should result in a=bc=d hyphenation
+a1b2c1d
+% and should not be overriden by pattern from too long comment (over MAX_CHARS characters)
+%|------------------------------ this part is 100 characters long --------------------------------|a8b9c8d
diff --git a/tests/longlines.test b/tests/longlines.test
new file mode 100755
index 0000000..f144145
--- /dev/null
+++ b/tests/longlines.test
@@ -0,0 +1,4 @@
+#!/bin/sh
+DIR="`dirname $0`"
+NAME="`basename $0 .test`"
+$DIR/test.sh $NAME.pat $NAME.word $NAME.hyph
diff --git a/tests/longlines.word b/tests/longlines.word
new file mode 100644
index 0000000..acbe86c
--- /dev/null
+++ b/tests/longlines.word
@@ -0,0 +1 @@
+abcd