On 12/12/18 1:41 AM, Richard Kimberly Heck wrote: > Another issue like the one we had with the 2.3.1 release, related to the > fix for #9158. The commit message explains the issue. The fix seems to > me to make sense anyway. > > The second patch also seems to me to make sense, and it ought to prevent > these kinds of issues arising in future. But maybe it's a bit heavy handed? > > Comments welcome. > > My plan is to go ahead and release 2.3.2 for non-Windows platforms as > planned, as this issue does not seem to affect them. If this fix seems > right, then I'll bulid a new Windows installer for testing and, if that > works, then I'll release a 2.3.2-1 tarball and a Windows installer built > on it as soon as I'm able. > > Comments welcome on that, too.
The other patch....
>From c49dac006b09f214f0255502eae580f98274f87a Mon Sep 17 00:00:00 2001 From: Richard Kimberly Heck <[email protected]> Date: Wed, 12 Dec 2018 01:37:00 -0500 Subject: [PATCH] There is no reason we should need to reload the bibinfo cache for internal buffers. --- src/Buffer.cpp | 3 +++ 1 file changed, 3 insertions(+) diff --git a/src/Buffer.cpp b/src/Buffer.cpp index b71c3fcd83..6343405b5c 100644 --- a/src/Buffer.cpp +++ b/src/Buffer.cpp @@ -2483,6 +2483,9 @@ void Buffer::clearBibFileCache() const void Buffer::reloadBibInfoCache() const { + if (isInternal()) + return; + // use the master's cache Buffer const * const tmp = masterBuffer(); if (tmp != this) { -- 2.17.2
