Trolltech Home | Qt4-preview-feedback Home | Recent Threads | All Threads | Author | Date
All threads index page 1

Qt4-preview-feedback Archive, December 2007
Another build failure for qt-win-commercial-src-4.4.0-snapshot-20071205.zip


Message 1 in thread

Dear Trolls,

Couple of issues prevented last night's 4.4. snapshot for windows to build 
using MSVC.NET

1. kernel\qwidget.cpp (9753), the closing (i.e. the "*/") of the comment 
block at the end of the file is missing - this is easily fixed.

2. tmp\moc\debug_shared\moc_qprintdialog.cpp (68), error C2039: 
'_q_collapseOrExpandDialog' : is not a member of 'QPrintDialogPrivate':
The problem seems to stem from the declaration of following line in 
qprintdialog.h,
    Q_PRIVATE_SLOT(d_func(), void _q_collapseOrExpandDialog())
the function is declared and implemented in X11 QPrintDialogPrivate (in 
qprintdialog_unix.cpp) but NOT in the private class' Windows counterpart 
(in qprintdialog_win.cpp).

This and the fact that, for one reason or the other, I haven't been able 
to build snapshots for Windows (commerical edition) since Andreas 
announced in his blog that Widgets On The Canvas would become available in 
the nightly snapshot on Nov 22, I couldn't help but wonder if Qt for 
Windows is expected to lag behind other platforms as far as nightly 
snapshot goes and if it would be smarter of me to stop bothering building 
it but instead play with the X11 version?

Pin-Yen Chen
National Instruments
1350 Terra Bella Avenue
Mountain View, CA 94043
(650) 934-0544

Message 2 in thread

Pinyen Chen wrote:
>1. kernel\qwidget.cpp (9753), the closing (i.e. the "*/") of the comment
>block at the end of the file is missing - this is easily fixed.

Interesting. Our internal copy has the closing */.

I added an extra newline to see if it makes a difference.

>2. tmp\moc\debug_shared\moc_qprintdialog.cpp (68), error C2039:
>'_q_collapseOrExpandDialog' : is not a member of 'QPrintDialogPrivate':
>The problem seems to stem from the declaration of following line in
>qprintdialog.h,
>    Q_PRIVATE_SLOT(d_func(), void _q_collapseOrExpandDialog())
>the function is declared and implemented in X11 QPrintDialogPrivate (in
>qprintdialog_unix.cpp) but NOT in the private class' Windows counterpart
>(in qprintdialog_win.cpp).

Indeed. The developer fixed it this morning. This code was added yesterday 
afternoon.

Thanks for pointing out anyways.

>This and the fact that, for one reason or the other, I haven't been able
>to build snapshots for Windows (commerical edition) since Andreas
>announced in his blog that Widgets On The Canvas would become available
> in the nightly snapshot on Nov 22, I couldn't help but wonder if Qt for
> Windows is expected to lag behind other platforms as far as nightly
> snapshot goes and if it would be smarter of me to stop bothering
> building it but instead play with the X11 version?

No, our Windows developers are building it every day. It's supposed to 
work, at least with VS 2005.

-- 
 [ signature omitted ] 

Attachment: signature.asc
Description: This is a digitally signed message part.


Message 3 in thread

Hi, Thiago,

I downloaded last night's 4.4 snapshot for Windows. The problem with 
QPrintDialogPrivate slot is indeed fixed. But the missing of comment block 
closing at the end of kernel\qwidget.cpp persists. I am not sure what 
distinguishes the internal from the external copy (theoratically they 
should be identical, I assume?) It's very easy to verify what's actaully 
in the snapshot, just download a copy from your ftp site and you will see 
the closing is missing. And if it is of any help, I went back and checked 
the X11 version snapshot I downloaded yesterday, the comment block closing 
is missing there too...

Once you get pass the missing comment closing, as far as I can tell, all 
primary 4.4 libraries would build ok. The build however would fail when 
trying to compile the following .ui file for the help\contextsensitivehelp 
example:

uic.exe wateringconfigdialog.ui -o ui_wateringconfigdialog.h
File 'wateringconfigdialog.ui' is not valid

And if you look into what's going on, it's invalid because the end ui tag 
is missing is its closing angel bracket:
...
</connections>
</ui                    <= seems to be truncated somehow

Pin-Yen Chen
National Instruments
1350 Terra Bella Avenue
Mountain View, CA 94043
(650) 934-0544



Thiago Macieira <thiago.macieira@xxxxxxxxxxxxx> 
12/05/2007 01:11 PM

To
qt4-preview-feedback@xxxxxxxxxxxxx
cc

Subject
Re: Another build failure for 
qt-win-commercial-src-4.4.0-snapshot-20071205.zip






Pinyen Chen wrote:
>1. kernel\qwidget.cpp (9753), the closing (i.e. the "*/") of the comment
>block at the end of the file is missing - this is easily fixed.

Interesting. Our internal copy has the closing */.

I added an extra newline to see if it makes a difference.

>2. tmp\moc\debug_shared\moc_qprintdialog.cpp (68), error C2039:
>'_q_collapseOrExpandDialog' : is not a member of 'QPrintDialogPrivate':
>The problem seems to stem from the declaration of following line in
>qprintdialog.h,
>    Q_PRIVATE_SLOT(d_func(), void _q_collapseOrExpandDialog())
>the function is declared and implemented in X11 QPrintDialogPrivate (in
>qprintdialog_unix.cpp) but NOT in the private class' Windows counterpart
>(in qprintdialog_win.cpp).

Indeed. The developer fixed it this morning. This code was added yesterday 

afternoon.

Thanks for pointing out anyways.

>This and the fact that, for one reason or the other, I haven't been able
>to build snapshots for Windows (commerical edition) since Andreas
>announced in his blog that Widgets On The Canvas would become available
> in the nightly snapshot on Nov 22, I couldn't help but wonder if Qt for
> Windows is expected to lag behind other platforms as far as nightly
> snapshot goes and if it would be smarter of me to stop bothering
> building it but instead play with the X11 version?

No, our Windows developers are building it every day. It's supposed to 
work, at least with VS 2005.

-- 
 [ signature omitted ] 

Attachment: signature.asc
Description: Binary data


Message 4 in thread

Pinyen Chen wrote:
>I am not sure what
>distinguishes the internal from the external copy (theoratically they
>should be identical, I assume?)

Unfortunately, that's not true.

There are two major differences between the internal and the external 
copies:

1) license headers
2) extra files

If you're a commercial customer, you'll know that the snapshot and release 
versions differ from the opensource counterparts only by the license the 
files are under (and some extra files).

And if you compare the "all", "mac", "win" and "x11" packages, you'll also 
notice that the difference lies in which files are present.

So, the external copies are created from our internal master copy by 
removing some files and replacing the internal license headers with the 
expected external ones. 

Somewhere in that process, the last bytes of a couple of files are 
removed. We don't know yet why. But thanks for letting us know, we're 
investigating.

-- 
 [ signature omitted ] 

Attachment: signature.asc
Description: This is a digitally signed message part.


Message 5 in thread

Make sense and thanks for the prompt response. I look forward to seeing 
the problem resolved.

Pin-Yen Chen
National Instruments
1350 Terra Bella Avenue
Mountain View, CA 94043
(650) 934-0544



Thiago Macieira <thiago.macieira@xxxxxxxxxxxxx> 
12/06/2007 11:45 PM

To
qt4-preview-feedback@xxxxxxxxxxxxx
cc

Subject
Re: Another build failure for 
qt-win-commercial-src-4.4.0-snapshot-20071205.zip






Pinyen Chen wrote:
>I am not sure what
>distinguishes the internal from the external copy (theoratically they
>should be identical, I assume?)

Unfortunately, that's not true.

There are two major differences between the internal and the external 
copies:

1) license headers
2) extra files

If you're a commercial customer, you'll know that the snapshot and release 

versions differ from the opensource counterparts only by the license the 
files are under (and some extra files).

And if you compare the "all", "mac", "win" and "x11" packages, you'll also 

notice that the difference lies in which files are present.

So, the external copies are created from our internal master copy by 
removing some files and replacing the internal license headers with the 
expected external ones. 

Somewhere in that process, the last bytes of a couple of files are 
removed. We don't know yet why. But thanks for letting us know, we're 
investigating.

-- 
 [ signature omitted ] 

Attachment: signature.asc
Description: Binary data


Message 6 in thread

On Friday 07 December 2007 08:45:24 Thiago Macieira wrote:
> Somewhere in that process, the last bytes of a couple of files are
> removed. We don't know yet why. But thanks for letting us know, we're
> investigating.

This is not a new issue, I remember when building OSS Qt 4.4 somewhere in 
September (downloaded daily snapshot from the ftp site) I also had to 
manually fix a couple of files, all missing some of their (pretty obvious) 
last bytes. Perhaps you are replacing UTF or other multibyte related 
characters (CRLF ?) and not accounting for the difference between characters 
and bytes. And of course it might be something completely unrelated to 
thar :)


To unsubscribe - send "unsubscribe" in the subject to qt4-preview-feedback-request@xxxxxxxxxxxxx