Re: [pgadmin-hackers] pgadmin3 debian package proposal

2003-10-26 Thread Andreas Pflug

Side by side. Mmmm, if we're talking side by side right back to CVS, we really ought to be sure we want to do this. Basically we'd be committing to maintaining our own wx fork indefinately. Moving back to the standard distro might not be so easy in the future...
 

No, I don't want it put into cvs, except for some custom makefiles 
maybe. We should stay with the current way of syncing our snapshot with 
wx from time to time, hoping for an age where we can get back to vanilla wx.

That said, a single build system really would be nice. 

If we have wx in a subdirectory of our source tree top side by side to 
src, the make process can do it in a single build. In this case, we 
wouldn't need to have wx installed, thus minimizing the chance of 
interference with an official wx package.

Regards,
Andreas


---(end of broadcast)---
TIP 7: don't forget to increase your free space map settings


Re: [pgadmin-hackers] pgadmin3 debian package proposal

2003-10-26 Thread Raphaël Enrici
Andreas Pflug wrote:


Side by side. Mmmm, if we're talking side by side right back to CVS, 
we really ought to be sure we want to do this. Basically we'd be 
committing to maintaining our own wx fork indefinately. Moving back 
to the standard distro might not be so easy in the future...
 

No, I don't want it put into cvs, except for some custom makefiles 
maybe. We should stay with the current way of syncing our snapshot 
with wx from time to time, hoping for an age where we can get back to 
vanilla wx. 
I agree with this.



That said, a single build system really would be nice.
If we have wx in a subdirectory of our source tree top side by side to 
src, the make process can do it in a single build. In this case, we 
wouldn't need to have wx installed, thus minimizing the chance of 
interference with an official wx package. 


What I was imaginating was to find a proper way to put, for example, the 
.tar.bz2 of "our" wx source in a specific place and to have a "magic" 
configure (for example) option which could handle the fact that the wx 
to link against is situated in the source tree. I don't want to get the 
wx source always in pgadmin3 source tree, think it has nothing to do 
here and I don't want that we loose cpu time to build wx snap every day 
when building snapshots... By the way, it may not be a good idea to 
"waste" a lot of time to get the configure/makefiles modified to get a 
single build process. May be we could just try to have the proper way of 
launching each configure / make commands and to document it well so that 
all packagers can publish spec, rules or whatever is appropriate to each 
disto. It may than be easier to backout to a dynamic build against 
official wx when they will be ready.

Another thing, may be we should not borrow Michael with these discuss 
anymore ? He had the good idea and may not want to receive all our 
thought about it (?) :)

I'll send back to you my build process as soon as it will be ok (hope it 
will be ok one day) (depends on my free time).

regards,
Raphaël


---(end of broadcast)---
TIP 6: Have you searched our list archives?
  http://archives.postgresql.org


Re: [pgadmin-hackers] pgadmin3 debian package proposal

2003-10-26 Thread Raphaël Enrici
Raphaël Enrici wrote:

Andreas Pflug wrote:



That said, a single build system really would be nice.


If we have wx in a subdirectory of our source tree top side by side 
to src, the make process can do it in a single build. In this case, 
we wouldn't need to have wx installed, thus minimizing the chance of 
interference with an official wx package. 

just for information, I have a first debian package that build without 
external wx. I'm gonna submit it to my "sponsor" so that he can gives 
its advice and gave corrections. I'll send you some news asap.
The method I adopted follows:
pgadmin-1.1.0/wxWindows-pgAdmin3-20031010-5 contains the untared snapshot
The debian/rules (equivalent of rpm spec file) is charged of building 
and installing it in 
pgadmin-1.1.0/wxWindows-pgAdmin3-20031010-5/localinst just before the 
pgA3 configure line is invoked.It refers to it using --with-wx and 
--with-wx-config configure options.

regards,
Raphaël
---(end of broadcast)---
TIP 9: the planner will ignore your desire to choose an index scan if your
 joining column's datatypes do not match


[pgadmin-hackers] New wx patch

2003-10-26 Thread Andreas Pflug
Hi Dave,

I nailed down and fixed another clipboard bug. Actually, it's the same 
problem we had with gtk, so I extended an existing patch.
I uploaded a cumulative patch to snake, but for compiling a new wx 
snapshot you'd probable be better off if you apply the attached patch 
separately. Since three files in the current snapshot are already 
patched, the re-patch probably fails. Just get a fresh copy from wx cvs, 
or I could mail you the files if you like that better.

For gtk, nightly builds do *not* need this additional patch, because the 
new patch is just a reimplementation; this change affects only msw.

Regards,
Andreas
Index: src/gtk/clipbrd.cpp
===
RCS file: /pack/cvsroots/wxwindows/wxWindows/src/gtk/clipbrd.cpp,v
retrieving revision 1.53
diff -u -r1.53 clipbrd.cpp
--- src/gtk/clipbrd.cpp 2003/08/09 12:46:08 1.53
+++ src/gtk/clipbrd.cpp 2003/10/26 15:40:15
@@ -171,11 +171,14 @@
 
 /* make sure we got the data in the correct form (selection type).
if so, copy data to target object */
+#if 0
+// prevents things to work, we already checked above!
 if (selection_data->type != GDK_SELECTION_TYPE_STRING)
 {
 clipboard->m_waiting = FALSE;
 return;
 }
+#endif
 
 data_object->SetData( format, (size_t) selection_data->length, (const char*) 
selection_data->data );
 
@@ -497,6 +500,15 @@
(guint32) GDK_CURRENT_TIME );
 
 while (m_waiting) gtk_main_iteration();
+
+#if defined(__WXGTK20__) && wxUSE_UNICODE
+if (!m_formatSupported && format == wxDataFormat(wxDF_UNICODETEXT))
+{
+// Another try with plain STRING format
+   extern GdkAtom g_altTextAtom;
+   return IsSupported(g_altTextAtom);
+}
+#endif
 
 if (!m_formatSupported) return FALSE;
 
Index: src/gtk/dataobj.cpp
===
RCS file: /pack/cvsroots/wxwindows/wxWindows/src/gtk/dataobj.cpp,v
retrieving revision 1.40
diff -u -r1.40 dataobj.cpp
--- src/gtk/dataobj.cpp 2003/09/09 18:04:57 1.40
+++ src/gtk/dataobj.cpp 2003/10/26 15:40:16
@@ -28,6 +28,7 @@
 //-
 
 GdkAtom  g_textAtom= 0;
+GdkAtom  g_altTextAtom = 0;
 GdkAtom  g_pngAtom = 0;
 GdkAtom  g_fileAtom= 0;
 
@@ -77,12 +78,9 @@
 {
 PrepareFormats();
 
-if (type == wxDF_UNICODETEXT)
-   type = wxDF_TEXT;
-
 m_type = type;
 
-if (m_type == wxDF_TEXT)
+if (m_type == wxDF_TEXT || m_type == wxDF_UNICODETEXT)
 m_format = g_textAtom;
 else
 if (m_type == wxDF_BITMAP)
@@ -143,7 +141,10 @@
 // here (with whom?)
 if (!g_textAtom)
 #if wxUSE_UNICODE
+{
 g_textAtom = gdk_atom_intern( "UTF8_STRING", FALSE );
+g_altTextAtom = gdk_atom_intern( "STRING", FALSE );
+}
 #else
 g_textAtom = gdk_atom_intern( "STRING" /* "text/plain" */, FALSE );
 #endif
@@ -191,6 +192,21 @@
 return n < nFormatCount;
 }
 }
+
+
+// 
+// wxTextDataObject
+// 
+
+#if defined(__WXGTK20__) && wxUSE_UNICODE
+
+void wxTextDataObject::GetAllFormats(wxDataFormat *formats, 
wxDataObjectBase::Direction dir) const
+{
+*formats++ = GetPreferredFormat();
+*formats = g_altTextAtom;
+}
+
+#endif
 
 // 
 // wxFileDataObject
Index: src/msw/ole/dataobj.cpp
===
RCS file: /pack/cvsroots/wxwindows/wxWindows/src/msw/ole/dataobj.cpp,v
retrieving revision 1.75
diff -u -r1.75 dataobj.cpp
--- src/msw/ole/dataobj.cpp 2003/09/12 16:10:00 1.75
+++ src/msw/ole/dataobj.cpp 2003/10/26 15:40:16
@@ -755,6 +755,23 @@
 
 #endif // Debug
 
+
+
+// 
+// wxTextDataObject supports CF_UNICODETEXT, CF_TEXT, CF_OEMTEXT
+// 
+
+
+void wxTextDataObject::GetAllFormats(wxDataFormat *formats,
+   wxDataObjectBase::Direction WXUNUSED(dir)) const
+{
+*formats++ = wxDataFormat(wxDF_UNICODETEXT);
+*formats++ = wxDataFormat(wxDF_TEXT);
+*formats   = wxDataFormat(wxDF_OEMTEXT);
+}
+
+
+
 // 
 // wxBitmapDataObject supports CF_DIB format
 // 
Index: include/wx/dataobj.h
===
RCS file: /pack/cvsroots/wxwindows/wxWindows/include/wx/dataobj.h,v
retrieving revision 1.51
diff -u -r1.51 dataobj.h
--- include/wx/dataobj.h2003/08/09 12:37:10 1.51
+++ i

Re: [pgadmin-hackers] New wx patch

2003-10-26 Thread Hiroshi Saito
Hi Andreas.

setup0.h isn't contained though "wx-20031026.patch" of the accumulation.
you think that it was particularly excluded.
Write it in readme-patches.txt.
Because I was stupid, I did Make-Test again.:-(
Does the thing which contains it become wxWindows-pgAdmin3-20031010-6.tar.bz2?

And ID of the doubt?(828538->828537)
---readme-patches.txt-----
...
wx-20031026.patch
=
...
all 828538 Wrong size calculation in wxCalendarCtrl

Is it 828538?
So, I am not found.

regards,
Hiroshi Saito

- Original Message - 
From: "Andreas Pflug" <[EMAIL PROTECTED]>


> Hi Dave,
> 
> I nailed down and fixed another clipboard bug. Actually, it's the same 
> problem we had with gtk, so I extended an existing patch.
> I uploaded a cumulative patch to snake, but for compiling a new wx 
> snapshot you'd probable be better off if you apply the attached patch 
> separately. Since three files in the current snapshot are already 
> patched, the re-patch probably fails. Just get a fresh copy from wx cvs, 
> or I could mail you the files if you like that better.
> 
> For gtk, nightly builds do *not* need this additional patch, because the 
> new patch is just a reimplementation; this change affects only msw.


---(end of broadcast)---
TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]