[ issue #3705 is the "corruption that fsfsverify fixes" issue. ]
One of my old todo notes from issue #3705 is to flush proto rev files.
Looking into the code, I came up with the following.
Does that look right? It seems too short to be the real fix.
[[[
Index: ../libsvn_fs_fs/fs_fs.c
===================================================================
--- ../libsvn_fs_fs/fs_fs.c (revision 1056454)
+++ ../libsvn_fs_fs/fs_fs.c (working copy)
@@ -795,8 +795,11 @@ get_writable_proto_rev_body(svn_fs_t *fs, const vo
/* Now open the prototype revision file and seek to the end. */
err = svn_io_file_open(file, path_txn_proto_rev(fs, txn_id, pool),
- APR_WRITE | APR_BUFFERED, APR_OS_DEFAULT, pool);
+ APR_WRITE, APR_OS_DEFAULT, pool);
+ /* Disable buffering. See notorious issue #3705. */
+ apr_err = apr_file_buffer_set(*file, NULL, 0);
+
/* You might expect that we could dispense with the following seek
and achieve the same thing by opening the file using APR_APPEND.
Unfortunately, APR's buffered file implementation unconditionally
@@ -2029,7 +2032,7 @@ open_and_seek_transaction(apr_file_t **file,
apr_off_t offset;
SVN_ERR(svn_io_file_open(&rev_file, path_txn_proto_rev(fs, txn_id, pool),
- APR_READ | APR_BUFFERED, APR_OS_DEFAULT, pool));
+ APR_READ, APR_OS_DEFAULT, pool));
offset = rep->offset;
SVN_ERR(svn_io_file_seek(rev_file, APR_SET, &offset, pool));
]]]
Daniel
(why disable buffering? I don't remember, but I could refresh my
memory using the issue and IRC logs.)
