[Libreoffice-commits] online.git: Branch 'distro/collabora/collabora-online-3' - wsd/ClientSession.cpp

[Libreoffice Gerrit user]

 wsd/ClientSession.cpp |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

New commits:
commit 7253dd1499db54550a4e4e887820027c9c8a4c31
Author:     Tamás Zolnai <tamas.zolnai at collabora.com>
AuthorDate: Fri Sep 21 16:04:53 2018 +0200
Commit:     Jan Holesovsky <kendy at collabora.com>
CommitDate: Mon Oct 1 12:16:38 2018 +0200

    Use a bigger timeout for waiting tileprocessed message
    
    The main purpose of this time-out is to avoid waiting forever for lost
    tile messages, but since it rare to loose them we can use bigger value,
    so we can avoid to send new tiles to a slow network.
    
    Note that the used time stamp does not mean the time when the tile
    actually send to the client, but the time when it gets to the sender
    queue.
    
    Change-Id: I230d85c38b3a5dafd195851d0cf4caac23149e3e
    (cherry picked from commit b623aca57ea47c785ab70ff6ccec58acbff63d75)
    Reviewed-on: https://gerrit.libreoffice.org/60879
    Reviewed-by: Jan Holesovsky <kendy at collabora.com>
    Tested-by: Jan Holesovsky <kendy at collabora.com>

diff --git a/wsd/ClientSession.cpp b/wsd/ClientSession.cpp
index 9ef6cf07f..ca43bc426 100644
--- a/wsd/ClientSession.cpp
+++ b/wsd/ClientSession.cpp
@@ -1066,7 +1066,7 @@ void ClientSession::removeOutdatedTilesOnFly()
     {
         auto tileIter = _tilesOnFly.begin();
         double elapsedTimeMs = std::chrono::duration_cast<std::chrono::milliseconds>(std::chrono::steady_clock::now() - tileIter->second).count();
-        if(elapsedTimeMs > 3000)
+        if(elapsedTimeMs > 5000.0)
         {
             LOG_WRN("Tracker tileID was dropped because of time out. Tileprocessed message did not arrive");
             _tilesOnFly.erase(tileIter);