Re: svn commit: r1002151 - in /subversion/trunk/subversion: libsvn_auth_kwallet/kwallet.cpp svn/main.c svnsync/main.c

2010-10-03 Thread Blair Zajac 

On 10/3/10 2:45 PM, Arfrever Frehtes Taifersar Arahesis wrote:

2010-09-29 14:29:26 Stefan Sperling napisał(a):

On Tue, Sep 28, 2010 at 11:21:26AM -0700, Blair Zajac wrote:

On 9/28/2010 6:07 AM, s...@apache.org wrote:

Author: stsp
Date: Tue Sep 28 13:07:23 2010
New Revision: 1002151

URL: http://svn.apache.org/viewvc?rev=1002151view=rev
Log:
Remove some goo introduced in r878078 and follow-ups, which was related to
the Linux-specific code which has been removed in r1002144.



-  INITIALIZE_APPLICATION
+  QCoreApplication *app;
+  if (! qApp)
+{
+  int argc = 1;
+  app = new QCoreApplication(argc, (char *[1]) {(char *) svn});
+}


Out of curiosity, what does this do?


No idea. I didn't understand this either.
I moved this code back out of the INITIALIZE_APPLICATION macro anyway.
The code was this way before the INITIALIZE_APPLICATION macro was
introduced.


QCoreApplication *app isn't static, so does this do some setup logic
that we don't need to keep track of?


That's what it looks like to me, too.
I suppose Arfrever will definitely know what this is doing.


This code initializes application to avoid segmentation fault.


Should it initialize the application everytime through the function, since app 
is not static, it won't store the result of the new QCoreApplication?


Blair

Re: svn commit: r1002151 - in /subversion/trunk/subversion: libsvn_auth_kwallet/kwallet.cpp svn/main.c svnsync/main.c

2010-09-29 Thread Stefan Sperling 
On Tue, Sep 28, 2010 at 11:21:26AM -0700, Blair Zajac wrote:
 On 9/28/2010 6:07 AM, s...@apache.org wrote:
 Author: stsp
 Date: Tue Sep 28 13:07:23 2010
 New Revision: 1002151
 
 URL: http://svn.apache.org/viewvc?rev=1002151view=rev
 Log:
 Remove some goo introduced in r878078 and follow-ups, which was related to
 the Linux-specific code which has been removed in r1002144.
 
 -  INITIALIZE_APPLICATION
 +  QCoreApplication *app;
 +  if (! qApp)
 +{
 +  int argc = 1;
 +  app = new QCoreApplication(argc, (char *[1]) {(char *) svn});
 +}
 
 Out of curiosity, what does this do?

No idea. I didn't understand this either.
I moved this code back out of the INITIALIZE_APPLICATION macro anyway.
The code was this way before the INITIALIZE_APPLICATION macro was
introduced.

 QCoreApplication *app isn't static, so does this do some setup logic
 that we don't need to keep track of?

That's what it looks like to me, too.
I suppose Arfrever will definitely know what this is doing.

Stefan

Re: svn commit: r1002151 - in /subversion/trunk/subversion: libsvn_auth_kwallet/kwallet.cpp svn/main.c svnsync/main.c

2010-09-28 Thread Blair Zajac 

On 9/28/2010 6:07 AM, s...@apache.org wrote:

Author: stsp
Date: Tue Sep 28 13:07:23 2010
New Revision: 1002151

URL: http://svn.apache.org/viewvc?rev=1002151view=rev
Log:
Remove some goo introduced in r878078 and follow-ups, which was related to
the Linux-specific code which has been removed in r1002144.



-  INITIALIZE_APPLICATION
+  QCoreApplication *app;
+  if (! qApp)
+{
+  int argc = 1;
+  app = new QCoreApplication(argc, (char *[1]) {(char *) svn});
+}


Out of curiosity, what does this do?

QCoreApplication *app isn't static, so does this do some setup logic 
that we don't need to keep track of?


Blair