From: Diane Trout <diane@caltech.edu>
Date: Thu, 7 Sep 2006 00:32:32 +0000 (+0000)
Subject: go back to using const char * in seqcomp
X-Git-Url: http://woldlab.caltech.edu/gitweb/?p=mussa.git;a=commitdiff_plain;h=a1b5ecc1ff79d588ffd1b68bc443e695825291c5

go back to using const char * in seqcomp
it appears that the string iterator version is about half as fast
as just using raw const char * pointers.

Hopefully the string pointers won't get moved around behind the scenes
while seqcomp is running. (Currently shouldn't as once a sequence is
set it's supposed to be constant.
---

diff --git a/alg/flp_seqcomp.cpp b/alg/flp_seqcomp.cpp
index 2c7ad7a..04eca42 100644
--- a/alg/flp_seqcomp.cpp
+++ b/alg/flp_seqcomp.cpp
@@ -52,8 +52,8 @@ FLPs::seqcomp(const Sequence& sseq1, const Sequence& sseq2, bool is_RC)
   int start_i, seq1_i, seq2_i, win_i;      // loop variables
   int matches;                    // number of matches in to a window
   int i2_offset;
-  Sequence::const_iterator seq1 = sseq1.begin();
-  Sequence::const_iterator seq2 = sseq2.begin();
+  const char *seq1 = sseq1.c_str();
+  const char *seq2 = sseq2.c_str();
 
   int seq1_win_num = sseq1.size() - window_size + 1;
   int seq2_win_num = sseq2.size() - window_size + 1;
diff --git a/alg/sequence.cpp b/alg/sequence.cpp
index 550d747..b2fa011 100644
--- a/alg/sequence.cpp
+++ b/alg/sequence.cpp
@@ -610,6 +610,10 @@ Sequence::clear()
   motif_list.clear();
 }
 
+const char *Sequence::c_str() const
+{
+  return seq->c_str();
+}
 
 Sequence::const_iterator Sequence::begin() const
 {
diff --git a/alg/sequence.hpp b/alg/sequence.hpp
index f77160c..34c28b6 100644
--- a/alg/sequence.hpp
+++ b/alg/sequence.hpp
@@ -133,6 +133,8 @@ public:
   const_reference at(size_type n) const;
   //! clear the sequence and its annotations
   void clear();
+  //! return c pointer to the sequence data
+  const char *c_str() const;
   //! forward iterator
   const_iterator begin() const;
   //! last iterator