[Raw Msg Headers][Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: 2.99.35 -- test results release :)



> > From zmailer-owner@nic.funet.fi  Fri Aug  9 02:04:08 1996
> >
> > plus recent Linuxes (ELF), and all (?) BSD systems.  Oh yes, anybody
> > with HP-UX ?  Any reports to share ?
> 
> I managed to compile the zmailer-2.99.35 at HPUX 9.01 with gcc-2.7.0 
> after making the following changes:
> (1) comment out line 76 or add #if !defined(_INCLUDE_HPUX_SOURCE) to
> 	#include <sys/select.h>
>     in libresolv/res_send.c

	Hmm..  That was added to handle AIX 3.2.5 ...
	I think I need to convince Paul Vixie to use autoconf at the bind..
	Alternatively you can forget using configuration option:
		--with-bundled-libresolv
	and you won't get into this trouble..
	(I added "#ifdef _AIX" around that include)

> (2) line 59 of smtpserver/fdstatfs.c, append one more condition:
>  	|| defined(_INCLUDE_HPUX_SOURCE)

	Naeh, there is a better way -- now done.
	Basically the  autoconf  provides tests for all known
	statfs() variants (a few more than I use, in fact),
	and now the module is rewritten with them.

> (3) line 36 of scheduler/qprint.c, we don't need to re-define struct
>     utimbuf and so, I use another #if to deal with it.
>     The same is applied to line 86 of transport/mailbox/mailbox.c

	Yes, also a bit hairy thing.
	Am I right that your system has  <utime.h>,  and that it defines
	the structure ?  (It is POSIX thing..)   Are there systems where
	utime() call is defined, but no  <utime.h> -header exists ?
	[ that question is for EVERYBODY, any knowledge out there ?
	  BSD 4.3 does not have utime(), nor the header.. ]

> (4) make sure that the following line is commented out:
> 	#define HAVE_UTIMES

	It is defined, if your system has  utimes() routine.
	Perhaps unfortunately the utimes() has had precedence so far.
	I will change things to have preference of  utime()  over the
	utimes() -- the utimes() is BSDish thing.

> ---------------------------------------------------------------------
> In addition, we got some warning message for select() in identuser.c:
> identuser.c:231: warning: passing arg 2 of `select' from incompatible 
> pointer type
> identuser.c:231: warning: passing arg 3 of `select' from incompatible 
> pointer type

	Where and how is your  1) fd_set defined, 2) select defined ?

> ---------------------------------------------------------------------
> another warning message from dbm.c:
> dbm.c: In function `open_dbm':
> dbm.c:40: warning: return makes integer from pointer without a cast

	Oops, right. (should be:  return -1;)

> ---------------------------------------------------------------------
> transport.c: In function `mux' `queryipccheck'and :
> transport.c:620: warning: passing arg 2 of `select' from incompatible 
> pointer type ...

	Same question as at identuser.c

> ---------------------------------------------------------------------
> smtp.c: In function `select_sleep':
> smtp.c:2212: warning: passing arg 2 of `select' from incompatible 
> pointer type

	again the same question

> ---------------------------------------------------------------------
> 
> That's all I have got. Thanks for your attention.

	You are welcome,

> -- W.W. Fung

	/Matti Aarnio <mea@nic.funet.fi>