[Webkit-unassigned] [Bug 30710] [Qt] QWebHistory::saveState() is inconsistent with the Qt API

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=30710





--- Comment #3 from jedrzej.nowacki at nokia.com  2009-10-28 01:47:55 PDT ---
Few month ago there was a discussion in the Oslo office about these functions.
The main argument for having saveState() and restoreState() was to keep
coherence with other Qt classes (like QApplication, QMainWindow...). Second
issue was a problem of multiple version of QWebHistory. The QDataSteream
version was changed on most of Qt minor, but not in bug fix releases and the
version format is X.Y (enum QDataStream::Version). As most of the QtWebkit bug
fix are done by upstreaming the Webkit part, there is a chance that we will
have to bump the version in a bug fix release. Qt's and Webkit's version number
are out of sync.

(In reply to comment #0)
> The function saveState is used for visible object like widgets (QFileDialog,
> QSplitter, etc). The bytearray returned is typically small, and is saved with
> QSettings.
> 
> For all other objects (QColor, QImage, etc), QDataStream & operator<<(Q...) is
> preferred since it can be used to write directly on the filesystem.
The QApplication is not a widget and it is hard to say that it is visible part
of an application.
QWebHistory is not so big, so comparison to QImage is an abuse. I don't think
that writing history is a bottle neck, and it have to be done as a real stream.
But of course optimizations are welcomed :-).

Summary:
The case is not so obvious, but I prefer streaming operators too. And I think
that removing of save&restore functions is perfectly save for the first Qt 4.6
release.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.