Re: [RFC/PATCH] Modifying internal FS transaction properties
From: Julian Foad <julianfoad_at_btopenworld.com>
Date: Tue, 3 Mar 2015 10:19:39 +0000
Evgeny Kotkov wrote:
+1 (concept)
That sounds like the logical approach. I haven't reviewed the patch itself.
> [...] As a caller, I would expect a
s/successive/successful/
> 2) These internal properties currently are a part of what svn_fs_txn_proplist()
+1.
The only test of svn_fs_txn_proplist() is fs-test.c 14 "transaction_props", and it fails if I change its begin-txn code from
SVN_ERR(svn_fs_begin_txn(&txn, fs, 0, pool));
to (for example)
SVN_ERR(svn_fs_begin_txn2(&txn, fs, 0, SVN_FS_TXN_CHECK_OOD, pool));
because it does not expect to see any txn-props that it did not create itself, except for svn:date.
The only other place where we currently acknowledge this leakage is in another test:
$ grep -E -w "SVN_FS__(CHECK_LOCKS|CHECK_OOD|CLIENT_DATE)|check-locks|check-ood|client-date" \
I think it would be good to change this now. I don't think there is any legitimate reason to keep exposing these properties just because we have been doing so for a long time.
- Julian
|
This is an archived mail posted to the Subversion Dev mailing list.
This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.