Re: release 3.0.2 again

[John W. Eaton]

On  8-Sep-2008, John W. Eaton wrote:

| On 30-Aug-2008, Thomas Treichl wrote:
| 
| | Hi,
| | 
| | it took me a while (like always ;) but finally I found the origin of that 
| | problem. There has been a change to load-path.cc from Octave version 3.0.0 
to 
| | 3.0.1 from
| | 
| |    static void
| |    maybe_add_path_elts (std::string& path, const std::string& dir)
| |    {
| |      std::string tpath = genpath (dir);
| | 
| |      if (! tpath.empty ())
| |        path += dir_path::path_sep_str + tpath;
| |    }
| | 
| | into
| | 
| |    static void
| |    maybe_add_path_elts (std::string& path, const std::string& dir)
| |    {
| |      std::string tpath = genpath (dir);
| | 
| |      if (! tpath.empty ())
| |        {
| |          if (path.empty ())
| |       path = tpath;
| |          else
| |       path += dir_path::path_sep_str + tpath;
| |        }
| |    }
| | 
| | With this new function the '--path' option from command line does not work 
| | anymore in 3.0.1 and 3.0.2 on my system. I replaced that new function in 
the 
| | 3.0.1 and 3.0.2 sources with the function from the 3.0.0 sources and 
recreated 
| | the binaries of Octave.app once again. Both (3.0.1 and 3.0.2) then accepted 
the 
| | '--path' option correctly.
| | 
| | Maybe this also addresses Vic's problem who found out another strange path 
| | problem at startup (but I'm not sure about this as I don't have any BBedit 
| | available)
| | 
| |    
https://www-old.cae.wisc.edu/pipermail/help-octave/2008-August/010660.html
| | 
| | Could we revert the change of maybe_add_path_elts for the 3.0.3 version and 
use 
| | the 3.0.0 function once again?
| 
| It seems to me that the new version is correct because it only does
| 
|   PATH = PATH + SEP + TPATH
| 
| if PATH is not empty.  Otherwise, it simply sets PATH to TPATH.  Isn't
| that the correct thing to do?  The previous version of the function
| could result in
| 
|   PATH = SEP + TPATH
| 
| which I think caused some problems.  So I don't think that this is the
| proper fix to the problem you are seeing.

Does the following patch (without the change you suggest above) also
solve the problem for you?

jwe

# HG changeset patch
# User Kim Hansen
# Date 1211316542 14400
# Node ID 37ff0c21c17dec2de4e8681cbec18ee90d7bd63f
# Parent  9a656cd34f3463f2aacfd8551cba47549353a6b6
load-path.cc (load_path::initialize): include separator when appending sys_path

diff --git a/src/ChangeLog b/src/ChangeLog
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,8 @@
+2008-05-20  Kim Hansen  <address@hidden>
+
+       * load-path.cc (load_path::do_initialize):
+       Include separator when appending sys_path.
+
 2008-05-20  Bill Denney  <address@hidden>
 
        * file-io.cc (Ffgets, Fgets, Ffputs, Ffscanf): Doc fix (seealso).
diff --git a/src/load-path.cc b/src/load-path.cc
--- a/src/load-path.cc
+++ b/src/load-path.cc
@@ -444,7 +444,7 @@
     xpath += dir_path::path_sep_str + tpath;
 
   if (! sys_path.empty ())
-    xpath += sys_path;
+    xpath += dir_path::path_sep_str + sys_path;
 
   do_set (xpath, false);
 }