From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jeffrey A Law <law@cygnus.com>
To: Mark Klein <mklein@dis.com>
Cc: gcc@gcc.gnu.org
Subject: Re: MPE Port 
Date: Tue, 26 Oct 1999 19:05:00 -0000
Message-id: <18403.940989240@upchuck>
References: <4.2.0.58.19991026064911.00c6ad40@garfield.dis.com>
X-SW-Source: 1999-10/msg00668.html

  In message < 4.2.0.58.19991026064911.00c6ad40@garfield.dis.com >you write:
  > >Seems to me that we can avoid #ifdefs in the actual funtions by providing
  > >a default value for PRI_FG_MIN_NP (and possibly others if they are not
  > >defined by your version of DCE threads).  If possible it helps to avoid
  > >#ifdefing executable C code just from a readability standpoint.
  > >
  > >What happens if you provide a default value of zero for those priority
  > >values if they're not already defined?
  > 
  > Actually, what's really missing is the pthread_setprio() call. That would
  > also need to be provided as a dummy procedure too. At that point, it
  > wouldn't matter what value were provided for PRI_FG_MIN_NP.
Ah.  Given this new information, it seems to me like this stuff (the use of
pthread_setprio) should be autoconf'd instead of ifdefing away pieces of
it based on the existence of PRI_FG_MIN_NP.

jeff

From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jeffrey A Law <law@cygnus.com>
To: Mark Klein <mklein@dis.com>
Cc: gcc@gcc.gnu.org
Subject: Re: MPE Port 
Date: Sun, 31 Oct 1999 23:35:00 -0000
Message-ID: <18403.940989240@upchuck>
References: <4.2.0.58.19991026064911.00c6ad40@garfield.dis.com>
X-SW-Source: 1999-10n/msg00666.html
Message-ID: <19991031233500.lT5ar8ZOa3l-n3vS5kG1jG3ouMG44ihHAquhsXWePEg@z>

  In message < 4.2.0.58.19991026064911.00c6ad40@garfield.dis.com >you write:
  > >Seems to me that we can avoid #ifdefs in the actual funtions by providing
  > >a default value for PRI_FG_MIN_NP (and possibly others if they are not
  > >defined by your version of DCE threads).  If possible it helps to avoid
  > >#ifdefing executable C code just from a readability standpoint.
  > >
  > >What happens if you provide a default value of zero for those priority
  > >values if they're not already defined?
  > 
  > Actually, what's really missing is the pthread_setprio() call. That would
  > also need to be provided as a dummy procedure too. At that point, it
  > wouldn't matter what value were provided for PRI_FG_MIN_NP.
Ah.  Given this new information, it seems to me like this stuff (the use of
pthread_setprio) should be autoconf'd instead of ifdefing away pieces of
it based on the existence of PRI_FG_MIN_NP.

jeff