bug in config for cygwin

Started by Scottover 23 years ago7 messagesbugs
Jump to latest
#1Scott
scott.ellis@aetas.com.au

Hi,

I'm not sure if what I've noticed fully qualifies as a bug...but, I run cygwin and also do a lot of development with vc++ (6)...Using version 7.2.1 I built the libpq++ libs/dll for win32 for use in my vc++ progs, and also ran the 'configure' script under cygwin...the problem occurs in the pg_config_os.h file, where an "#include <cygwin/version.h>" occurs before an "#ifdef __CYGWIN__" - so naturally my vc++ builds fail. By moving the include to after the ifdef the problem goes away...I thought I'd send this here so somwone with knowledge of patching the configure scripts etc. might have a look.

Regards,
Scott

#2Bruce Momjian
bruce@momjian.us
In reply to: Scott (#1)
Re: bug in config for cygwin

I can't seem to find the file you are having a problem with. Would you
supply the full path of the file?

---------------------------------------------------------------------------

Scott wrote:

Hi,

I'm not sure if what I've noticed fully qualifies as a bug...but,
I run cygwin and also do a lot of development with vc++ (6)...Using
version 7.2.1 I built the libpq++ libs/dll for win32 for use in
my vc++ progs, and also ran the 'configure' script under
cygwin...the problem occurs in the pg_config_os.h file, where
an "#include <cygwin/version.h>" occurs before an "#ifdef
__CYGWIN__" - so naturally my vc++ builds fail. By moving the
include to after the ifdef the problem goes away...I thought
I'd send this here so somwone with knowledge of patching the
configure scripts etc. might have a look.

Regards, Scott

--
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073
#3Scott
scott.ellis@aetas.com.au
In reply to: Bruce Momjian (#2)
Re: bug in config for cygwin

Hi Bruce,

The path is:

postgresql-7.2.1/src/include/pg_config_os.h

Just while I'm here - congrats for a great peice of software, and on such a
quick response time! :)

Another note - I applied that patch for QNX6 (forget authors name - sorry -
but the one that you guys argued about pretty intensly re going into version
7.2) on our QNX (6.2) development system at work...it seems to have worked
no problems. Though I didn't try installing the patched version under cygwin
or building the libpq stuff under windows...

Scott

----- Original Message -----
From: "Bruce Momjian" <pgman@candle.pha.pa.us>
To: "Scott" <scott.ellis@aetas.com.au>
Cc: <pgsql-bugs@postgresql.org>
Sent: Saturday, August 17, 2002 1:17 AM
Subject: Re: [BUGS] bug in config for cygwin

I can't seem to find the file you are having a problem with. Would you
supply the full path of the file?

--------------------------------------------------------------------------

-

Scott wrote:

Hi,

I'm not sure if what I've noticed fully qualifies as a bug...but,
I run cygwin and also do a lot of development with vc++ (6)...Using
version 7.2.1 I built the libpq++ libs/dll for win32 for use in
my vc++ progs, and also ran the 'configure' script under
cygwin...the problem occurs in the pg_config_os.h file, where
an "#include <cygwin/version.h>" occurs before an "#ifdef
__CYGWIN__" - so naturally my vc++ builds fail. By moving the
include to after the ifdef the problem goes away...I thought
I'd send this here so somwone with knowledge of patching the
configure scripts etc. might have a look.

Regards, Scott

--
Bruce Momjian                        |  http://candle.pha.pa.us
pgman@candle.pha.pa.us               |  (610) 359-1001
+  If your life is a hard drive,     |  13 Roberts Road
+  Christ can be your backup.        |  Newtown Square, Pennsylvania

19073

Show quoted text

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org

#4Bruce Momjian
bruce@momjian.us
In reply to: Scott (#3)
Re: bug in config for cygwin

Scott wrote:

Hi Bruce,

The path is:

postgresql-7.2.1/src/include/pg_config_os.h

Oh, OK. I don't see an #include <version.h> in that file so I am
confused how it got there.

Just while I'm here - congrats for a great peice of software, and on such a
quick response time! :)

Another note - I applied that patch for QNX6 (forget authors name - sorry -
but the one that you guys argued about pretty intensly re going into version
7.2) on our QNX (6.2) development system at work...it seems to have worked
no problems. Though I didn't try installing the patched version under cygwin
or building the libpq stuff under windows...

QNX is one of our platforms we want to get working for 7.3 but no one
has started working on it yet. We have refactored some of the SysV
stuff so it should be easier now. If you know anyone who is interested
in submitting a QNX port patch, please let them know.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073
#5Peter Eisentraut
peter_e@gmx.net
In reply to: Scott (#1)
Re: bug in config for cygwin

Scott writes:

I'm not sure if what I've noticed fully qualifies as a bug...but, I run cygwin and also do a lot of development with vc++ (6)...Using version 7.2.1 I built the libpq++ libs/dll for win32 for use in my vc++ progs, and also ran the 'configure' script under cygwin...the problem occurs in the pg_config_os.h file, where an "#include <cygwin/version.h>" occurs before an "#ifdef __CYGWIN__" - so naturally my vc++ builds fail. By moving the include to after the ifdef the problem goes away...I thought I'd send this here so somwone with knowledge of patching the configure scripts etc. might have a look.

If you configure under cygwin then you can't build under not-cygwin. If
you want to build the libraries for native Windows you need to follow the
instructions for the native build.

--
Peter Eisentraut peter_e@gmx.net

#6Scott
scott.ellis@aetas.com.au
In reply to: Peter Eisentraut (#5)
Re: bug in config for cygwin

Yep I thought so - that's why I didn't think it's really a bug - but the
include really ought to be after the define, and if it was I would be able
to use the same headers for both cygwin and non-cygwin builds...would just
be 'nice to have'.

----- Original Message -----
From: "Peter Eisentraut" <peter_e@gmx.net>
To: "Scott" <scott.ellis@aetas.com.au>
Cc: <pgsql-bugs@postgresql.org>
Sent: Sunday, August 18, 2002 1:46 AM
Subject: Re: [BUGS] bug in config for cygwin

Scott writes:

I'm not sure if what I've noticed fully qualifies as a bug...but, I run

cygwin and also do a lot of development with vc++ (6)...Using version 7.2.1
I built the libpq++ libs/dll for win32 for use in my vc++ progs, and also
ran the 'configure' script under cygwin...the problem occurs in the
pg_config_os.h file, where an "#include <cygwin/version.h>" occurs before an
"#ifdef __CYGWIN__" - so naturally my vc++ builds fail. By moving the
include to after the ifdef the problem goes away...I thought I'd send this
here so somwone with knowledge of patching the configure scripts etc. might
have a look.

Show quoted text

If you configure under cygwin then you can't build under not-cygwin. If
you want to build the libraries for native Windows you need to follow the
instructions for the native build.

--
Peter Eisentraut peter_e@gmx.net

---------------------------(end of broadcast)---------------------------
TIP 4: Don't 'kill -9' the postmaster

#7Peter Eisentraut
peter_e@gmx.net
In reply to: Scott (#6)
Re: bug in config for cygwin

Scott writes:

Yep I thought so - that's why I didn't think it's really a bug - but the
include really ought to be after the define, and if it was I would be able
to use the same headers for both cygwin and non-cygwin builds...would just
be 'nice to have'.

Write to pgsql-cygwin@postgresql.org to discuss your proposal. Maybe it's
possible, but someone knowledgeable needs to check that it doesn't break
anything else.

--
Peter Eisentraut peter_e@gmx.net