jeirmeister/Cockatrice
1
0
Fork 0
mirror of https://github.com/Cockatrice/Cockatrice.git synced 2026-09-21 09:05:10 -07:00
Code Issues Projects Releases Packages Wiki Activity Actions 3

[DeckStorage] Keep the visibility refresh armed until replies land

Browse source 
The single-shot drain was armed with the 500 ms delay at send time, so a
round trip slower than that drained before the server applied the change,
re-read the old state and never re-armed, leaving the column stale until
a manual refresh. The timer is now armed with the full network timeout at
send time (a lost reply still costs one stale refresh) and re-armed with
the short delay every time a reply lands.
This commit is contained in:
Lukas Brübach 2026-09-20 20:17:18 +02:00 • committed by GitHub
parent dd2400d245
commit e0f8fd0091
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 15 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

19
cockatrice/src/interface/widgets/tabs/tab_deck_storage.cpp
View file

@ -127,10 +127,11 @@ TabDeckStorage::TabDeckStorage(TabSupervisor *_tabSupervisor,
SettingsCache::instance().network().getKeepAlive() * 1000));
connect(shareTimeoutTimer, &QTimer::timeout, this, &TabDeckStorage::onShareFromTreeTimeout);
// Restartable single-shot refresh for the Public/Private column. A dropped
// visibility reply must not leave the widget permanently stale, so the tree
// is re-read whenever publishes quiet down instead of waiting on a count
// that can get stuck above zero.
// Restartable single-shot refresh for the Public/Private column. It is
// armed with the full network timeout when a publish is sent (so a dropped
// reply still drains once) and re-armed with the short delay every time a
// reply lands, so the drain cannot fire while a slow round trip is still in
// flight. Either way the tree is re-read once things quiet down.
visibilityRefreshTimer = new QTimer(this);
visibilityRefreshTimer->setSingleShot(true);
visibilityRefreshTimer->setInterval(VISIBILITY_REFRESH_DELAY);
@ -862,6 +863,13 @@ void TabDeckStorage::onShareFromTreeTimeout()
void TabDeckStorage::actPublishDeck()
{
visibilityFailures.clear();
// Arm the drain with the full network timeout so a lost reply still costs
// one refresh instead of a dead column; each reply shrinks it to the short
// delay below, so a slow round trip is never drained before it lands.
const int visibilityFailSafeDelay =
static_cast<int>((static_cast<qint64>(SettingsCache::instance().network().getTimeOut()) + 1) *
SettingsCache::instance().network().getKeepAlive() * 1000);
const auto selection = serverDirView->getCurrentSelection();
for (const auto *node : selection) {
Command_DeckSetVisibility cmd;
@ -884,6 +892,7 @@ void TabDeckStorage::actPublishDeck()
PendingCommand *pend = client->prepareSessionCommand(cmd);
connect(pend, &PendingCommand::finished, this, &TabDeckStorage::setVisibilityFinished);
visibilityRefreshStarted = true;
visibilityRefreshTimer->setInterval(visibilityFailSafeDelay);
visibilityRefreshTimer->start();
client->sendCommand(pend);
}
@ -893,6 +902,7 @@ void TabDeckStorage::setVisibilityFinished(const Response &r, const CommandConta
{
if (r.response_code() == Response::RespOk) {
if (visibilityRefreshStarted) {
visibilityRefreshTimer->setInterval(VISIBILITY_REFRESH_DELAY);
visibilityRefreshTimer->start();
}
return;
@ -904,6 +914,7 @@ void TabDeckStorage::setVisibilityFinished(const Response &r, const CommandConta
.arg(QString::number(static_cast<int>(r.response_code())));
if (visibilityRefreshStarted) {
visibilityFailures.append(message);
visibilityRefreshTimer->setInterval(VISIBILITY_REFRESH_DELAY);
visibilityRefreshTimer->start();
} else {
QMessageBox::critical(this, tr("Error"), message);