On Thu, Sep 08, 2011 at 01:34:14PM +0100, Colin Watson wrote:
> Thanks to a note from John W. Eaton that the development version doesn't
> suffer from this problem, I tracked down the fix and have prepared a
> suitable patch against the Debian package. Please consider this to fix
> the octave-symbolic build failure.
Urgh, I missed the existence of debian/in/series. Updated patch
follows.
* New patch: clear_symbol_table: Clean up top-level variables when exiting
Octave (closes: #628353, LP: #831157).
diff -u octave3.2-3.2.4/debian/patches/series
octave3.2-3.2.4/debian/patches/series
--- octave3.2-3.2.4/debian/patches/series
+++ octave3.2-3.2.4/debian/patches/series
@@ -14,2 +14,3 @@
avoid-implicit-conversion
+clear_symbol_table
diff -u octave3.2-3.2.4/debian/in/series octave3.2-3.2.4/debian/in/series
--- octave3.2-3.2.4/debian/in/series
+++ octave3.2-3.2.4/debian/in/series
@@ -24,2 +24,3 @@
avoid-implicit-conversion
+clear_symbol_table
:]
only in patch2:
unchanged:
--- octave3.2-3.2.4.orig/debian/patches/clear_symbol_table
+++ octave3.2-3.2.4/debian/patches/clear_symbol_table
@@ -0,0 +1,90 @@
+Description: Clean up top-level variables when exiting Octave
+ This fixes this assertion failure:
+ .
+ Inconsistency detected by ld.so: dl-close.c: 736: _dl_close: Assertion
+ `map->l_init_called' failed!
+ .
+ This patch is slightly modified from the upstream commits to add
+ __attribute__ ((visibility ("hidden"))) to the symbol_table::cleanup
+ method, in order that this patch does not change the externally-visible
+ ABI. That modification can and should be dropped when moving to a new
+ upstream version containing this fix.
+Origin: upstream, http://hg.savannah.gnu.org/hgweb/octave/rev/692ab4eaf965
+Origin: upstream, http://hg.savannah.gnu.org/hgweb/octave/rev/a5a05b2ebb9d
+Author: Jaroslav Hajek <high...@gmail.com>
+Bug: http://savannah.gnu.org/bugs/?34195
+Bug-Debian: http://bugs.debian.org/633719
+Bug-Debian: http://bugs.debian.org/628353
+Bug-Ubuntu: https://bugs.launchpad.net/bugs/831157
+
+Index: b/src/symtab.cc
+===================================================================
+--- a/src/symtab.cc
++++ b/src/symtab.cc
+@@ -1193,6 +1193,39 @@
+ }
+ }
+
++void symbol_table::cleanup (void)
++{
++ // Clear variables in top scope.
++ all_instances[xtop_scope]->clear_variables ();
++
++ // Clear function table. This is a hard clear, ignoring mlocked functions.
++ fcn_table.clear ();
++
++ // Clear variables in global scope.
++ // FIXME: are there any?
++ all_instances[xglobal_scope]->clear_variables ();
++
++ // Clear global variables.
++ global_table.clear ();
++
++ // Delete all possibly remaining scopes.
++ for (all_instances_iterator iter = all_instances.begin ();
++ iter != all_instances.end (); iter++)
++ {
++ scope_id scope = iter->first;
++ if (scope != xglobal_scope && scope != xtop_scope)
++ scope_id_cache::free (scope);
++
++ // First zero the table entry to avoid possible duplicate delete.
++ symbol_table *inst = iter->second;
++ iter->second = 0;
++
++ // Now delete the scope. Note that there may be side effects, such as
++ // deleting other scopes.
++ delete inst;
++ }
++}
++
+ DEFUN (ignore_function_time_stamp, args, nargout,
+ "-*- texinfo -*-\n\
+ @deftypefn {Built-in Function} {@var{val} =} ignore_function_time_stamp ()\n\
+Index: b/src/symtab.h
+===================================================================
+--- a/src/symtab.h
++++ b/src/symtab.h
+@@ -1838,6 +1838,8 @@
+ static void stash_dir_name_for_subfunctions (scope_id scope,
+ const std::string& dir_name);
+
++ static void cleanup (void) __attribute__ ((visibility ("hidden")));
++
+ private:
+
+ typedef std::map<std::string, symbol_record>::const_iterator
table_const_iterator;
+Index: b/src/toplev.cc
+===================================================================
+--- a/src/toplev.cc
++++ b/src/toplev.cc
+@@ -673,6 +673,9 @@
+ {
+ do_octave_atexit ();
+
++ // Clean up symbol table.
++ SAFE_CALL (symbol_table::cleanup, ());
++
+ SAFE_CALL (sysdep_cleanup, ())
+
+ if (octave_exit)
Thanks,
--
Colin Watson [cjwat...@ubuntu.com]
--
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
