VSEARCH 1.0.12: Temporary fix for long FASTA headers

torognes · Feb 6, 2015 · 3d7ad4d · 3d7ad4d
1 parent e0cc843
commit 3d7ad4d
Show file tree

Hide file tree

Showing 6 changed files with 18 additions and 13 deletions.
diff --git a/doc/vsearch.1 b/doc/vsearch.1
@@ -1,5 +1,5 @@
 .\" ============================================================================
-.TH vsearch 1 "February 5, 2015" "version 1.0.11" "USER COMMANDS"
+.TH vsearch 1 "February 6, 2015" "version 1.0.12" "USER COMMANDS"
 .\" ============================================================================
 .SH NAME
 vsearch \(em chimera detection, clustering, dereplication, masking, pairwise alignment, searching, shuffling and sorting of amplicons from metagenomic projects.
@@ -1519,6 +1519,9 @@ affecting clustering and dereplication commands.
 .TP
 .BR v1.0.11\~ "released February 5th, 2015"
 Introduces the possibility to output results in SAM format.
+.TP
+.BR v1.0.12\~ "released February 6th, 2015"
+Temporarily fixes a problem with long headers in FASTA files.
 .LP
 .\" ============================================================================
 .\" TODO:

diff --git a/doc/vsearch_manual.pdf b/doc/vsearch_manual.pdf
diff --git a/src/align.cc b/src/align.cc
@@ -409,7 +409,7 @@ void nw_align(char * dseq,
 #if 1
   if (score != dist)
   {
-    fprintf(stderr, "Warning: Error with query no %lu and db sequence no %lu:\n", queryno, dbseqno);
+    fprintf(stderr, "WARNING: Error with query no %lu and db sequence no %lu:\n", queryno, dbseqno);
     fprintf(stderr, "Initial and recomputed alignment score disagreement: %lu %lu\n", dist, score);
     fprintf(stderr, "Alignment: %s\n", cigar);
   }

diff --git a/src/db.cc b/src/db.cc
@@ -163,7 +163,10 @@ void db_read(const char * filename, int upcase)
       unsigned long hdrbegin = datalen;
 
       if (line[0] != '>')
-        fatal("Illegal header line in fasta file.");
+        fatal("Illegal header line in fasta file (not starting with '>' character).");
+
+      if (strlen(line) + 1 == LINEALLOC)
+        fatal("FASTA header line in database too long");
 
       /* terminate header at first space or end of line */
 
@@ -407,7 +410,7 @@ void db_read(const char * filename, int upcase)
 
   if (stripped_count)
     {
-      fprintf(stderr, "Warning: invalid characters stripped from sequence:");
+      fprintf(stderr, "WARNING: invalid characters stripped from sequence:");
       for (int i=0; i<256;i++)
         if (stripped[i])
           fprintf(stderr, " %c(%d)", i, stripped[i]);

diff --git a/src/query.cc b/src/query.cc
@@ -1,5 +1,5 @@
 /*
-    Copyright (C) 2014 Torbjorn Rognes & Tomas Flouri
+    Copyright (C) 2014-2015 Torbjorn Rognes & Tomas Flouri
 
     This program is free software: you can redistribute it and/or modify
     it under the terms of the GNU Affero General Public License as
@@ -25,12 +25,8 @@
    allocated here for the query header and sequence. This buffers will
    be overwritten on the next call of query_getnext. */
 
-#ifndef LINE_MAX
-#define LINE_MAX 2048
-#endif
-
 #define MEMCHUNK 4096
-#define LINEALLOC LINE_MAX
+#define LINEALLOC 1048576
 
 extern unsigned int chrstatus[256];
 
@@ -184,7 +180,7 @@ void query_close()
 
   if (query_stripped_count)
     {
-      fprintf(stderr, "Warning: invalid characters stripped from query:");
+      fprintf(stderr, "WARNING: invalid characters stripped from query:");
       for (int i=0; i<256;i++)
         if (query_stripped[i])
           fprintf(stderr, " %c(%d)", i, query_stripped[i]);
@@ -222,8 +218,11 @@ int query_getnext(char ** head, int * head_len,
       /* read header */
 
       if (query_line[0] != '>')
-        fatal("Illegal header line in query fasta file");
+        fatal("Illegal header line in query fasta file (not starting with '>' character).");
 
+      if (strlen(query_line) + 1 == LINEALLOC)
+        fatal("FASTA header line in query too long");
+
       /* terminate header at first space or end of line */
 
       char * z0 = query_line + 1;

diff --git a/src/vsearch.h b/src/vsearch.h
@@ -84,7 +84,7 @@
 #include "allpairs.h"
 
 #define PROG_NAME "vsearch"
-#define PROG_VERSION "v1.0.11"
+#define PROG_VERSION "v1.0.12"
 
 #ifdef __APPLE__
 #define PROG_ARCH "osx_x86_64"