Bug 7057 - ncurses segfaults during compile
: ncurses segfaults during compile
Status: CLOSED FIXED
Product: Codex
Classification: Unclassified
Component: libs
: test grimoire
: Other other
: P2 normal
Assigned To: Grimoire Bug List
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2004-06-27 00:23 UTC by David Kowis
Modified: 2007-04-01 01:47 UTC (History)
1 user (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description David Kowis 2004-06-27 00:23:24 UTC
I've had this happen to me on two different computers/architectures.
Same thing happens on both machines. Very wierd.

** Building terminfo database, please wait...
Running tic to install /usr/share/terminfo ...

        You may see messages regarding unknown capabilities, e.g., AX.
        These are extended terminal capabilities which can be compiled
        using
                tic -x
        Read the INSTALL document before doing this - it can cause
        problems for older ncurses applications.

"../misc/terminfo.src", line 402, col 3, terminal 'ecma+color': unknown
capability 'AX'
"../misc/terminfo.src", line 2920, col 31, terminal 'xterm-1002': unknown
capability 'XM'
"../misc/terminfo.src", line 2922, col 31, terminal 'xterm-1003': unknown
capability 'XM'
"../misc/terminfo.src", line 3640, col 40, terminal 'screen': unknown capability
'G0'
"../misc/terminfo.src", line 3661, col 31, terminal 'screen': unknown capability
'E0'
"../misc/terminfo.src", line 3661, col 44, terminal 'screen': unknown capability
'S0'
"../misc/terminfo.src", line 4594, col 13, terminal 'djgpp204': unknown
capability 'AX'
1486 entries written to /usr/share/terminfo
** built new /usr/share/terminfo
** sym-linked /usr/lib/terminfo for compatibility
installing std
installing stdcrt
installing vt100
installing vt300
make[2]: Leaving directory `/usr/src/ncurses-5.4/misc'
cd c++ && /usr/bin/make DESTDIR="" install
make[2]: Entering directory `/usr/src/ncurses-5.4/c++'
/bin/install -c ../lib/libncurses++.a /usr/lib/libncurses++.a
installing ./cursesapp.h in /usr/include
installing ./cursesf.h in /usr/include
installing ./cursesm.h in /usr/include
installing ./cursesp.h in /usr/include
installing ./cursesw.h in /usr/include
installing ./cursslk.h in /usr/include
installing etip.h in /usr/include
make[2]: Leaving directory `/usr/src/ncurses-5.4/c++'
make[1]: Leaving directory `/usr/src/ncurses-5.4'
make: *** [ncurses] Segmentation fault
make: *** No rule to make target `coreutils', needed by `basesystem'.
make: *** No rule to make target `gcc', needed by `basesystem'.
make: *** No rule to make target `textutils', needed by `basesystem'.
make: *** No rule to make target `termcap', needed by `psmisc'.
make: Target `all' not remade because of errors.
Comment 1 erics 2004-06-27 09:54:28 UTC

*** This bug has been marked as a duplicate of 7029 ***
Comment 2 erics 2004-06-27 09:55:09 UTC
fix I think is applied in test and devel grimoires... see other bug. 
 
erics 
Comment 3 David Kowis 2004-06-27 12:41:19 UTC
I'm not sure how these two bugs relate...
Ncurses segfaults for me on two different machines. The

make: *** No rule to make target `coreutils', needed by `basesystem'.
make: *** No rule to make target `gcc', needed by `basesystem'.
make: *** No rule to make target `textutils', needed by `basesystem'.
make: *** No rule to make target `termcap', needed by `psmisc'.

stuff doesn't cause it to fail, if that's what you're thinking.
For some strange reason ncurses segmentation faults during it's install:
/bin/install -c ../lib/libncurses++.a /usr/lib/libncurses++.a
installing ./cursesapp.h in /usr/include
installing ./cursesf.h in /usr/include
installing ./cursesm.h in /usr/include
installing ./cursesp.h in /usr/include
installing ./cursesw.h in /usr/include
installing ./cursslk.h in /usr/include
installing etip.h in /usr/include
make[2]: Leaving directory `/usr/src/ncurses-5.4/c++'
make[1]: Leaving directory `/usr/src/ncurses-5.4'
make: *** [ncurses] Segmentation fault


either that or it's make thats segfaulting. I know its not my computer because
it's happening on multiple (3) machines.
Comment 4 Eric Sandall 2004-06-27 22:56:16 UTC
David, if you recompile psmisc and then cast ncurses, does it still segfault?
Comment 5 I quit smgl. 2004-06-28 04:21:13 UTC
the libs guru needs to be more competent than to copy libraries to fix a simple
./configure problem.
If he had taken the time to read the commented lines in BUILD before deleting
them he would have seen that your conversion to API2 lost --libdir=/lib which
was how the API1 spell solved the problem of making the libs available at boot
before /usr was mounted.
It's a real shame all the problems in sorcery during his attempts to mislead
that team didn't teach him anything, and I can't believe he expected you guys to
give him stable grimoire access !
Perhaps the current libs guru should be reassigned to some section that doesn't
require common sense, deductive skills or general competence.
Of course as I don't use smgl any more and I have my own "grimoire" , whatever
you decide is your business, it just seems like the reasons I left are still
valid so I took the chance to say "I told you so"
Happy SMuGLing
Comment 6 Arwed v. Merkatz 2004-06-28 07:58:13 UTC
I can't reproduce the segfault, but try if the current spell in test works for
you David.

Hamish: Eric didn't do the BUILD_API conversion of this spell that lost the
--libdir switch and used copying instead, so no need for those remarks.
Comment 7 I quit smgl. 2004-06-28 08:30:59 UTC
Arwed,
 please read bug 7026 (not 7029 which, unfortunately, schabell incorrectly
marked this as a duplicate of) and you will see that while schabell wasn't
responsible for commenting out all the old ./configure code in BUILD he was the
one who removed it and added the rm of the /lib/* still leaving the cp -* line,
that caused the problems with API2.
It is your choice to allow him to continue making contributions to the grimoire. 
Honestly appraising his skills based on all the past experiences and only giving
him access to some non-volatile section (that won't stop the system from booting
and being recoverable by ordinary users) is the right thing to do for the smgl
user base. 
Refusing to hold people accountable for their mistakes and safeguard against
futher damage, either because of some good actions in the distant past or
because of their high profile is irresponsible towards those users.
Comment 8 Arwed v. Merkatz 2004-06-29 03:54:26 UTC
Got someone else with the segfault problem to test the fixed spell, seems to
work. I integrated it into stable, please check if it works for you when you get
back David.
Comment 9 David Kowis 2004-07-02 12:05:42 UTC
I've had a successful compile of it, so whatever you did works. Testing it on
another machine to verify...

Success! It works on both machines. wonderful. Even with the negative input from
the "doesn't use smgl" user. 

So I'm gonna close it and call it fixed. Thanks!
Comment 10 erics 2004-07-07 05:23:13 UTC
well, luckily for you all, being as incompetent as I am... I will just close 
this bug. 
 
erics 
Comment 11 Jeremy Blosser 2007-04-01 00:47:23 UTC
reassign to sm-grimoire-bugs