Reviewing merge request #1243: Porting of Qt to T-Kernel

This makes Qt work on T-Kernel that supports POSIX extensions.
The base is Qt for Embedded linux.
* cross compiles with RVCT v4.0 on Cygwin.
* works with QWS (tested with the VNC driver only)
* tested on ARM11 hardware
* no cups (no printers), no qt3support, no phonon, no largefile, no bearer
no mouse-pc, no mouse-linuxtp, and no kbd-tty

Commits that would be merged:

Version 2
  • Version 1
  • Version 1
  • Version 2
  • 896db16
  • d604047
  • Porting of Qt to T-Kernel

  • 46a620f
  • examples and demos is returned in the original

Showing 896db16-d604047

Comments

Hello Saeki,

thank you very much for your contribution. Would it be possible to split up the one big change into smaller, logical parts? That would make it easier for us to apply.

Also please note that all code in 3rd-party needs to be submitted to the corresponding repository, since Qt only contains a copy of the code. For example, the JavaScriptCore changes should go to webkit.org.

I can’t help but think that a lot of this port has been done a bit backwards.. Qt is designed to keep the details of the systems from the user, yet a lot of this patch is putting implementation details in the hands of the third party, not inside Qt itself – for a single example:

63 #ifdef Q_OS_TKSE
64 // Tkse does not have a picture location.
65 QDesktopServices::DataLocation,
66 #else
67 QDesktopServices::PicturesLocation,
68 #endif

Is it not possible to handle this inside QDesktopServices, so that third party developers won’t have to concern themselves with how TKSE handles things internally?

Another example of this is in QLocalServer stuff:

50 #ifdef Q_OS_TKSE
51 hostLineEdit = new QLineEdit(“/socket/fortune”);
52 #else
50 53 hostLineEdit = new QLineEdit(“fortune”);
54 #endif
51 55

Is there really no way to handle this more transparently?

→ State changed from New to Updated

The examples and demos is returned in the original.

Please let me commit to one.
Because it was one in a big patch to transplant to VxWorks, I also unified it.

reviewing such a huge patch is a real pain.
btw, this needs to be rebased as some things were changed alot…

→ State changed from Updated to Revise and resubmit

hi,

i see no alternative to splitting. only that way you’ll be able to properly justify each change (and improve their quality on the way). see also http://developer.qt.nokia.com/wiki/Commit_Policy
how the patch was originally done is of no concern to us.

thanks

I see.
I will divide commit according to module.

note that this is note about splitting by some file/directory boundaries, but by logical, atomic changes.

→ State changed from Revise and resubmit to Closed

Thank you for your contribution. With Qt 4 having moved [1] under Open
Governance [2] as well, you can now contribute your changes via Gerrit
[3]. At the same time, contributions via Gitorious can no longer be
merged.

We would still like to see your patch on the new platform, and wish
for you to become a member of the Qt Project. Please read [4] and in
particular [5] to familiarize yourself with the new tools. You will
still find a read-only mirror of all Qt repositories on Gitorious.

With a much wider set of possible reviewers, we are also looking
forward to better response times to contributions. Looking forward to
your patch!

Please note that at this point in time Qt 4.8 is a bugfix only branch,
and all fixes should be submitted to Qt 5 first, as far as
applicable. Qt 4 master is closed.

The Nokia Qt team

[1] http://labs.qt.nokia.com/2012/01/10/qt-4-moved-to-open-governance/
[2] http://labs.qt.nokia.com/2011/10/21/the-qt-project-is-live/
[3] http://codereview.qt-project.org
[4] http://www.qt-project.org
[5] http://developer.qt.nokia.com/wiki/Gerrit_Introduction

Add a new comment:

Login or create an account to post a comment

How to apply this merge request to your repository