Re: [pgadmin-hackers] CVS compile error

2003-10-03 Thread Dave Page
Title: Message



 

  
  -Original Message-From: Adam H. 
  Pendleton [mailto:[EMAIL PROTECTED] Sent: 03 October 2003 
  20:41To: Dave PageCc: 
  [EMAIL PROTECTED]Subject: Re: [pgadmin-hackers] CVS 
  compile errorDave Page wrote:

-Original Message-
From: Adam H.Pendleton [mailto:[EMAIL PROTECTED]] 
Sent: 02 October 2003 23:32
To: [EMAIL PROTECTED]
Subject: [pgadmin-hackers] CVS compile error


I get this when compiling a fresh CVS tree:

utils/utffile.cpp: In member function `off_t 
wxUtfFile::Read(wxString&, 
long
long int)':
utils/utffile.cpp:85: no matching function for call to 
`wxMBConv::MB2WC(
wxStringBuffer, char*&, long unsigned int)'
/usr/local/wx2/include/wx/strconv.h:46: candidates are: virtual size_t
wxMBConv::MB2WC(wchar_t*, const char*, long unsigned int) const
make[2]: *** [utffile.o] Error 1
make[2]: *** Waiting for unfinished jobs
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

You need the 20031001 wx snapshot from the wxWindows site. Currently we
haven't patched it.

Regards, Dave.

  
  Does this mean that the CVS head from wxWindows doesn't work? 
Dunno 
about that, but that's the same error I got before upgrading to that snapshot 
from our previous one.
 
Regards, Dave.


Re: [pgadmin-hackers] CVS compile error

2003-10-03 Thread Adam H. Pendleton




Dave Page wrote:

  
  
  
-Original Message-
From: Adam H.Pendleton [mailto:[EMAIL PROTECTED]] 
Sent: 02 October 2003 23:32
To: [EMAIL PROTECTED]
Subject: [pgadmin-hackers] CVS compile error


I get this when compiling a fresh CVS tree:

utils/utffile.cpp: In member function `off_t 
wxUtfFile::Read(wxString&, 
long
long int)':
utils/utffile.cpp:85: no matching function for call to 
`wxMBConv::MB2WC(
wxStringBuffer, char*&, long unsigned int)'
/usr/local/wx2/include/wx/strconv.h:46: candidates are: virtual size_t
wxMBConv::MB2WC(wchar_t*, const char*, long unsigned int) const
make[2]: *** [utffile.o] Error 1
make[2]: *** Waiting for unfinished jobs
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

  
  
You need the 20031001 wx snapshot from the wxWindows site. Currently we
haven't patched it.

Regards, Dave.

  

Does this mean that the CVS head from wxWindows doesn't work?

ahp




Re: [pgadmin-hackers] [pgadmin-support] Bug in pgadmin III displaying/exporting

2003-10-03 Thread Dave Page


> -Original Message-
> From: Andreas Pflug [mailto:[EMAIL PROTECTED] 
> Sent: 03 October 2003 17:42
> To: [EMAIL PROTECTED]
> Cc: pgadmin-hackers; [EMAIL PROTECTED]; Adam H. Pendleton
> Subject: Re: [pgadmin-hackers] [pgadmin-support] Bug in 
> pgadmin III displaying/exporting
> 
> 
> >
> So that distri isn't compiled for SSL.

SSL is broken on Unix due to a build problem - see my email from this
morning. Hopefully Adam will have a fix soon.

Regards, Dave

---(end of broadcast)---
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])


Re: [pgadmin-hackers] [pgadmin-support] Bug in pgadmin III displa

2003-10-03 Thread Godshall Michael
Title: RE: [pgadmin-hackers] [pgadmin-support] Bug in pgadmin III displaying/exporting dat





Hi,


I did download the binary(pgadmin3-0.9.0-1.i386.rpm).  The folder on the mirror now only has a RedHat9.0 folder.  I think it used to be named just RedHat but maybe I originally mis-read it.

I still have the rpm if that would help.


Mike


-Original Message-
From: Andreas Pflug [mailto:[EMAIL PROTECTED]]
Sent: Friday, October 03, 2003 11:42 AM
To: [EMAIL PROTECTED]
Cc: pgadmin-hackers; [EMAIL PROTECTED]; Adam H. Pendleton
Subject: Re: [pgadmin-hackers] [pgadmin-support] Bug in pgadmin III
displaying/exporting dat



Jean-Michel POURE wrote:


>Le Vendredi 3 Octobre 2003 17:52, Andreas Pflug a écrit :
>  
>
>>Or did Jean-Michel just forget to use --enable-static when building?
>>    
>>
>
>No, the packages use %configure --enable-debug --enable-static 
>--with-wx-config=%{withwxconfig}. And I can confirm that the SSL combo does 
>not display under RedHat 9.
>
So that distri isn't compiled for SSL.


Michael, did you use said binary package? If not, which did you use?


If Michael really did use that RH9 package, this would mean that 
libcrypto and libssl are linked, but pgAdmin3 is SSL-disabled. This 
certainly wouldn't make sense.


So if SSL is not defined, the ssl libs may not be linked to it either.


Regards,
Andreas






Re: [pgadmin-hackers] [pgadmin-support] Bug in pgadmin III displa ying/exporting dat

2003-10-03 Thread Jean-Michel POURE
Le Vendredi 3 Octobre 2003 18:55, Godshall Michael a écrit :
> So if SSL is not defined, the ssl libs may not be linked to it either.

I can remember using an old snapshot with the SSL combo. The libraries should 
be installed on the build host. What is the name of the libraries needed?

Cheers, Jean-Michel


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


Re: [pgadmin-hackers] [pgadmin-support] Bug in pgadmin III displaying/exporting

2003-10-03 Thread Andreas Pflug
Jean-Michel POURE wrote:

Le Vendredi 3 Octobre 2003 17:52, Andreas Pflug a écrit :
 

Or did Jean-Michel just forget to use --enable-static when building?
   

No, the packages use %configure --enable-debug --enable-static 
--with-wx-config=%{withwxconfig}. And I can confirm that the SSL combo does 
not display under RedHat 9.

So that distri isn't compiled for SSL.

Michael, did you use said binary package? If not, which did you use?

If Michael really did use that RH9 package, this would mean that 
libcrypto and libssl are linked, but pgAdmin3 is SSL-disabled. This 
certainly wouldn't make sense.

So if SSL is not defined, the ssl libs may not be linked to it either.

Regards,
Andreas
---(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


Re: [pgadmin-hackers] [pgadmin-support] Bug in pgadmin III displaying/exporting

2003-10-03 Thread Adam H. Pendleton
Andreas Pflug wrote:

Hi Adam,

obviously the Redhat binary distribution needs dynamically linked 
libcrypto and libssl. IMHO we should spend the additional 10kb and 
have both libs statically linked to pgAdmin3, to avoid these 
dependencies. Or did Jean-Michel just forget to use --enable-static 
when building?

AFAIR we confirmed that by default dynamic linking should be used, 
while our binary distributions should be as static as possible to 
minimize dependencies.

Okay, I will take a look at this again and see if the --disable-shared 
flag can statically link as much as possible.

ahp

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


Re: [pgadmin-hackers] [pgadmin-support] Bug in pgadmin III displaying/exporting dat

2003-10-03 Thread Jean-Michel POURE
Le Vendredi 3 Octobre 2003 17:52, Andreas Pflug a écrit :
> Or did Jean-Michel just forget to use --enable-static when building?

No, the packages use %configure --enable-debug --enable-static 
--with-wx-config=%{withwxconfig}. And I can confirm that the SSL combo does 
not display under RedHat 9.

Cheers, Jean-Michel


---(end of broadcast)---
TIP 3: if posting/reading through Usenet, please send an appropriate
  subscribe-nomail command to [EMAIL PROTECTED] so that your
  message can get through to the mailing list cleanly


Re: [pgadmin-hackers] [pgadmin-support] Bug in pgadmin III displaying/exporting dat

2003-10-03 Thread Andreas Pflug
Adam H. Pendleton wrote:

Andreas Pflug wrote:

Godshall Michael wrote:

I use linux at home and would test it there if I can get pgadminIII 
installed on it.  Last time I tried it failed because I did not have 
ssl and crypto so.4 libraries.  My Redhat 8.0 box has versions 
so.2.  Maybe I downloaded the wrong version of pgadminIII.

Adam,

can't we link ssl and crypto static too? Alternatively, we'd have to 
distribute non-SSL versions too..

I'm not sure I follow. 


Hi Adam,

obviously the Redhat binary distribution needs dynamically linked 
libcrypto and libssl. IMHO we should spend the additional 10kb and have 
both libs statically linked to pgAdmin3, to avoid these dependencies. Or 
did Jean-Michel just forget to use --enable-static when building?

AFAIR we confirmed that by default dynamic linking should be used, while 
our binary distributions should be as static as possible to minimize 
dependencies.

Regards,
Andreas


---(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


Re: [pgadmin-hackers] [pgadmin-support] Bug in pgadmin III displaying/exporting dat

2003-10-03 Thread Adam H. Pendleton
Andreas Pflug wrote:

Godshall Michael wrote:

I use linux at home and would test it there if I can get pgadminIII 
installed on it.  Last time I tried it failed because I did not have 
ssl and crypto so.4 libraries.  My Redhat 8.0 box has versions so.2.  
Maybe I downloaded the wrong version of pgadminIII.

Adam,

can't we link ssl and crypto static too? Alternatively, we'd have to 
distribute non-SSL versions too..

I'm not sure I follow.

ahp

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


Re: [pgadmin-hackers] wxWindows Upgrade

2003-10-03 Thread Adam H. Pendleton




Dave Page wrote:

  
  
  
-Original Message-
From: Jean-Michel POURE [mailto:[EMAIL PROTECTED]] 
Sent: 03 October 2003 08:31
To: Dave Page
Cc: pgadmin-hackers
Subject: Re: [pgadmin-hackers] wxWindows Upgrade


Le Jeudi 2 Octobre 2003 16:28, Dave Page a écrit :


  Add --enable-monolithic when you configure it to get the old single 
library.
  

Thanks for the information. wxGTK 2.5.1 will be released this 
week-end. 
Therefore, it is high time I compile the official RPMs (which are not 
monolithic) today and hand them to the wxWindows team.

  
  
OK. If you want something to do, perhaps you could look into what's required to get pgAdmin to build with a non-monolithic wx? I'm guessing it's just an appropriate patch to /acinclude.m4

  

Yes, just a patch to acinclude.  I will take care of this one while I
am fixing the -DSSL bug.

ahp




Re: [pgadmin-hackers] wxWindows Upgrade

2003-10-03 Thread Dave Page


> -Original Message-
> From: Jean-Michel POURE [mailto:[EMAIL PROTECTED] 
> Sent: 03 October 2003 08:31
> To: Dave Page
> Cc: pgadmin-hackers
> Subject: Re: [pgadmin-hackers] wxWindows Upgrade
> 
> 
> Le Jeudi 2 Octobre 2003 16:28, Dave Page a écrit :
> > Add --enable-monolithic when you configure it to get the old single 
> > library.
> 
> Thanks for the information. wxGTK 2.5.1 will be released this 
> week-end. 
> Therefore, it is high time I compile the official RPMs (which are not 
> monolithic) today and hand them to the wxWindows team.

OK. If you want something to do, perhaps you could look into what's required to get 
pgAdmin to build with a non-monolithic wx? I'm guessing it's just an appropriate patch 
to /acinclude.m4

> This way, pgAdmin3 should be published by Linux vendors more 
> quickly. Will get 
> back to you during the day.

Cool :-)

BTW, we made Newsforge and apparently Segfault though I can't confirm that one, as 
well as a number of regional Linux sites...

Regards, Dave.

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


Re: [pgadmin-hackers] CVS compile error

2003-10-03 Thread Dave Page


> -Original Message-
> From: Adam H.Pendleton [mailto:[EMAIL PROTECTED] 
> Sent: 02 October 2003 23:32
> To: [EMAIL PROTECTED]
> Subject: [pgadmin-hackers] CVS compile error
> 
> 
> I get this when compiling a fresh CVS tree:
> 
> utils/utffile.cpp: In member function `off_t 
> wxUtfFile::Read(wxString&, 
> long
> long int)':
> utils/utffile.cpp:85: no matching function for call to 
> `wxMBConv::MB2WC(
> wxStringBuffer, char*&, long unsigned int)'
> /usr/local/wx2/include/wx/strconv.h:46: candidates are: virtual size_t
> wxMBConv::MB2WC(wchar_t*, const char*, long unsigned int) const
> make[2]: *** [utffile.o] Error 1
> make[2]: *** Waiting for unfinished jobs
> make[1]: *** [all-recursive] Error 1
> make: *** [all] Error 2

You need the 20031001 wx snapshot from the wxWindows site. Currently we
haven't patched it.

Regards, Dave.

---(end of broadcast)---
TIP 3: if posting/reading through Usenet, please send an appropriate
  subscribe-nomail command to [EMAIL PROTECTED] so that your
  message can get through to the mailing list cleanly


Re: [pgadmin-hackers] gcc ignoring defined macros?

2003-10-03 Thread Dave Page


> -Original Message-
> From: Adam H. Pendleton [mailto:[EMAIL PROTECTED] 
> Sent: 02 October 2003 23:29
> To: Dave Page
> Cc: Hiroshi Saito; [EMAIL PROTECTED]
> Subject: Re: [pgadmin-hackers] gcc ignoring defined macros?
> 
> 
> 
> How are you adding the -DSSL flag to the command line?  It might be 
> mistaken for a linker flag because of its location in the 
> args.  Just a 
> shot in the dark, really...

Ah-ha!!

Right, I editted the line in src/Makefile that compiles frmConnect.cpp.
I added -DSSL so the final command read:

if g++ -DSSL -DHAVE_CONFIG_H -I. 

And that works. :-)

Further investigation of the command line:

if g++ -DHAVE_CONFIG_H -I. -I. -I..
-DDATA_DIR=\"/usr/local/pgadmin3/share/pgadmin3/\" -Wall -g
-I../src/include -I../src/agent/include -I -DSSL
-I/usr/local/pgsql/include -I/usr/local/lib/wx/include/gtk2ud-2.5
-D__WXDEBUG__ -D__WXGTK__ -D_FILE_OFFSET_BITS=64 -D_LARGE_FILES
-I/usr/local/include -DDATA_DIR=\"/usr/local/pgadmin3/share/pgadmin3/\"
-Wall -g -I../src/include -I../src/agent/include -I -g -O2 -MT
frmConnect.o -MD -MP -MF ".deps/frmConnect.Tpo" \
  -c -o frmConnect.o `test -f 'ui/frmConnect.cpp' || echo
'./'`ui/frmConnect.cpp; \
then mv ".deps/frmConnect.Tpo" ".deps/frmConnect.Po"; \
else rm -f ".deps/frmConnect.Tpo"; exit 1; \
fi

shows up the problem. Firstly, a bunch of it seems to be included twice.
Note the double inclusion of:

-DDATA_DIR=\"/usr/local/pgadmin3/share/pgadmin3/\" -Wall -g
-I../src/include -I../src/agent/include -I

Then secondly, notice the -I on the end which does not have a path -
that's what seems to be losing the SSL definition because it gets picked
up as the argument to -I. Further investigation shows that that path is
set to ${WXHOME} which is ovbviously empty when the command is created.
Adding a / to the end of it makes things work nicely.

Not sure about the best way to fix this properly, so I guess it's over
to you Adam :-)

Regards, Dave.

---(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


Re: [pgadmin-hackers] wxWindows Upgrade

2003-10-03 Thread Jean-Michel POURE
Le Jeudi 2 Octobre 2003 16:28, Dave Page a écrit :
> Add --enable-monolithic when you configure it to get the old single
> library.

Thanks for the information. wxGTK 2.5.1 will be released this week-end. 
Therefore, it is high time I compile the official RPMs (which are not 
monolithic) today and hand them to the wxWindows team.

This way, pgAdmin3 should be published by Linux vendors more quickly. Will get 
back to you during the day.

Cheers, Jean-Michel


---(end of broadcast)---
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])


Re: [pgadmin-hackers] wxWindows Upgrade

2003-10-03 Thread Dave Page


> -Original Message-
> From: Andreas Pflug [mailto:[EMAIL PROTECTED] 
> Sent: 02 October 2003 19:15
> Cc: [EMAIL PROTECTED]
> Subject: [pgadmin-hackers] wxWindows Upgrade
> 
> 
> Just to have this clear:
> 
> pgAdmin3 1.0.0 branch will compile with both our own snapshot 
> or with a 
> reasonable wx cvs version (2003-10-01 is tested). If you 
> don't use our 
> snapshot, the clipboard won't work correctly under gtk. At 
> the moment, 
> we recommend using our wx snapshot, and we expect that this 
> remain true 
> in the future.
> 
> pgAdmin3 cvs head requires a new wx cvs snapshot (2003-10-01, later 
> might work also). Clipboard still won't work, but if you're using the 
> cvs head you're probably developing/testing new things/ports 
> and won't 
> bother.

I'm working on both - can I get another PC and second slackware box
please?

:-)

Regards, Dave.

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