[BUG] svn_fs_begin_txn

[Yoshiki Hayashi <yo...@xemacs.org>]

svn_fs_begin_txn always return "0" as a transaction ID.  I
traced it to allocate_txn_id but I couldn't find the cause.
It correctly tries to put next_id + 1 into transactions
table but somehow it fails to do so.

Below is the patch for fs-test.c.  You'll see strcmp in this
patch fails.

Index: fs-test.c
===================================================================
RCS file: /cvs/subversion/subversion/tests/libsvn_fs/fs-test.c,v
retrieving revision 1.14
diff -u -r1.14 fs-test.c
--- fs-test.c	2001/02/28 02:19:55	1.14
+++ fs-test.c	2001/02/28 10:08:43
@@ -257,7 +258,7 @@
       if (SVN_NO_ERROR != svn_fs_begin_txn (&txn, fs, 0, pool))
         return fail();
       
-      /* Test that it got id "0", since it's the second txn. */
+      /* Test that it got id "1", since it's the second txn. */
       {
         char *txn_name;
         
@@ -265,7 +266,7 @@
         if (err)
           return fail();
         
-        if (strcmp (txn_name, "0") != 0)
+        if (strcmp (txn_name, "1") != 0)
           return fail();
       }
       


-- 
Yoshiki Hayashi

Re: [BUG] svn_fs_begin_txn

[Yoshiki Hayashi <yo...@xemacs.org>]

Yoshiki Hayashi <yo...@xemacs.org> writes:

> svn_fs_begin_txn always return "0" as a transaction ID.  I
> traced it to allocate_txn_id but I couldn't find the cause.
> It correctly tries to put next_id + 1 into transactions
> table but somehow it fails to do so.

Sorry for the false alert.  This works fine now.

-- 
Yoshiki Hayashi