2.2.7-STABLE bug?

2.2.7-STABLE bug?

Post by Matthias Schuendehuett » Mon, 17 Aug 1998 04:00:00


when Im doing a make world with a cvsuped 2.2.7-STABLE as of
Aug.13,1998 I found the following problems:


In the Makefile "$Id: Makefile,v 1997/08/08 02:17:57 jdp Exp"
Line 6 a file cplus-dem.c is mentioned which doesnt exist. Its
referenced from
warnings.c, ID "warnings.c,v 1997/08/08 02:18:03 jdp Exp",
function "demangle(name)", line 73.

I commented these things out, but its possibly not a solution....


the target install doesnt work for ./fonts and ./keymaps (but for

Workaround for me is to put a "cd ${.CURDIR} ; " before the
install-comand in the
Makefiles of the above mentioned subdirs, but its annoying to do these
edits after
every run of cvsup.

So: whats the problem and why does ./scrnmaps work?

Ciao/BSD -Matthias


2.2.7-STABLE bug?

Post by Tony Griffith » Tue, 18 Aug 1998 04:00:00

> Hello,

> when Im doing a make world with a cvsuped 2.2.7-STABLE as of
> Aug.13,1998 I found the following problems:

Well I did a CVSup + 'cvs update' + 'make world' yesterday (Sunday) morning
and it worked, so whatever problem you have/had has been resolved!  Since
2.2.7-STABLE is 'not so stable' (ie. there are checkins going on), it is
possible to get the respository in a slightly inconsistent state.  This is
generally fixed within the hour (or at most day) but can make an automated
build (mine is a cron job that kicks off at 1 AM every Sunday) fail!



1. config bug on FreeBSD 2.2.8-STABLE

As I posted to  gnu.utils.bug  FYI ...

I just tried building the new 2.5.4 version of patch(1) on a FreeBSD
2.2.8-STABLE platform, and ran into this error when make'ing it:

 gcc -c  -DHAVE_CONFIG_H -Ded_PROGRAM=\"/bin/ed\"  -I. -I. -O3 patch.c
 gcc -c  -DHAVE_CONFIG_H -Ded_PROGRAM=\"/bin/ed\"  -I. -I. -O3 pch.c
 pch.c: In function `open_patch_file':
 pch.c:113: too many arguments to function `setmode'
 *** Error code 1


Seems that on FreeBSD 2.2.8-STABLE, setmode(3) only takes a single arg,
per its man page (excerpted):

 SETMODE(3)           FreeBSD Library Functions Manual         SETMODE(3)

     getmode, setmode - modify mode bits

     #include <unistd.h>

     getmode(const void *set, mode_t mode)

     void *
     setmode(const char *mode_str)

     The setmode() function takes an absolute (octal) or symbolic value, as
     described in chmod(1),  as an argument and returns a pointer to mode val-
     ues to be supplied to getmode().  Because some of the symbolic values are
     relative to the file creation mask, setmode() may call umask(2).  If this
     occurs, the file creation mask will be restored before setmode() returns.
     If the calling program changes the value of its file creation mask after
     calling setmode(), setmode() must be called again if getmode() is to mod-
     ify future file modes correctly.

     If the mode passed to setmode() is invalid, setmode() returns NULL.

     The setmode() function may fail and set errno for any of the errors spec-
     ified for the library routine malloc(3).

The obvious quick-fix is to undef/comment-out the HAVE_SETMODE #define
in  config.h  before make'ing it (though I didn't look through the code
in any depth to see if doing so has any ramifications on functionality,
or performance).

Nor do I have any idea if this problem occurs on FreeBSD 3.x/4.x systems.

[Also posted to  comp.unix.bsd.freebsd.misc  for informational purposes,


"Any technology distinguishable from magic is insufficiently advanced."
                                      --Cleon I

2. Terminal Emulation

3. FreeBSD 4.4-20011111-STABLE bug?

4. What is the advantage of a _real time_ OS?

5. Network bug in FreeBSD 4.4 STABLE

6. Netscape compiled for LINUX, WHERE?!?!?

7. NFS bug fixes since 3.1-19990416-STABLE?

8. IPC: InterProcess Comunication

9. Network bug - FreeBSD 4.4 STABLE

10. Upgrade from 3.4-STABLE to 4.x-STABLE

11. 3.2-STABLE to 3.4-STABLE device name change

12. 4.0-20000430-STABLE vs. 4.0-STABLE

13. 2.2.8 Stable to 3.0 Stable Migration Question