D19389: Rewrite kworkspace logout, shutdown and suspend API

[Kai Uwe Broulik]

broulik added inline comments.

INLINE COMMENTS

> loginddbustypes.h:103
> +
> +class NamedUserPath
> +{

You're using `struct` and `class` with all public inconsistently

> loginddbustypes.h:135
> +    QString mode;
> +    int userId;
> +    uint processId;

`userId` is unsigned

> sessionmanagement.h:74
> +
> +
> +    SessionManagement(QObject *parent = nullptr);

Whitespace

> sessionmanagement.h:80
> +
> +    bool canShutdown();
> +    bool canReboot();

Why are these not `const`?

> sessionmanagement.h:93
> +     * These requestX methods will either launch a prompt to shutdown or
> +     * The user may cancel it at any point in the process
> +     */

Didn't we have a "don't ask" thing that also doesn't ask for apps to close?

> sessionmanagementbackend.cpp:63
> +{
> +    return m_kserverConfig->group("General").readEntry("ConfirmLogout", true); // TODO
> +}

what's the TODO?

> sessionmanagementbackend.cpp:84
> +        } else {
> +            // both "yes" and "challenge" will show up in the UI
> +            *argToUpdate = reply.value() != QLatin1String("no") ? true : false;

Better explicitly check for `yes` and `challenge`, as there is also:

> If "na" is returned the operation is not available because hardware, kernel or drivers do not support it.

> sessionmanagementbackend.cpp:91
> +            emit stateChanged();
> +            emit canShutdownChanged();
> +            emit canRebootChanged();

Can we emit those only if actually different now?

> sessionmanagementbackend.cpp:200
> +    auto canStop = m_ck->CanStop();
> +    canStop.waitForFinished();
> +    m_canShutdown = canStop.value();

Why is logind all async but consolekit is not? :)

REPOSITORY
  R120 Plasma Workspace

REVISION DETAIL
  https://phabricator.kde.org/D19389

To: davidedmundson, #plasma
Cc: pino, broulik, plasma-devel, LeGast00n, jraleigh, fbampaloukas, GB_2, ragreen, Pitel, ZrenBot, himcesjf, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, apol, mart
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/plasma-devel/attachments/20190621/52690a37/attachment-0001.html>