Bug 267013 – >=net-fs/nfs-utils-1.14 fail to due to FORTIFY_SOURCE

• Status: RESOLVED WORKSFORME
• Product: Gentoo Linux

[Description James Browning 2009-04-21 20:32:12 UTC]

A grotty hack (not included) to add "-U_FORTIFY_SOURCE" to the ebuilds CFLAGS allowed it to emerge normally.
Gengor said to block re: bug 259417


Reproducible: Always

Steps to Reproduce:
1. have gcc-4.2.4-r1 from the xake toolchain overlay
2. emerge nfs-utils 1.1.4 or 1.1.5

Actual Results:  
compile fail to complete, breaking the emerge (microblurb from 1.1.4)

Expected Results:  
nfs-utils should emerge normally

i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I../../support/include  -I../../support/include -D_GNU_SOURCE -I../../support/include -I../../support/export -D_GNU_SOURCE -Wall -Wstrict-prototypes  -pipe -march=pentium2 -mtune=prescott -O2 -pipe -g -ggdb -MT mountd-rmtab.o -MD -MP -MF .deps/mountd-rmtab.Tpo -c -o mountd-rmtab.o `test -f 'rmtab.c' || echo './'`rmtab.c           
In file included from rmtab.c:25:                                                                                            
../../support/include/ha-callout.h: In function 'ha_callout':                                                                
../../support/include/ha-callout.h:57: error: 'N_DEBUG' undeclared (first use in this function)                              
../../support/include/ha-callout.h:57: error: (Each undeclared identifier is reported only once                              
../../support/include/ha-callout.h:57: error: for each function it appears in.)                                              
make[2]: *** [mountd-rmtab.o] Error 1

[Comment 1 James Browning 2009-04-21 20:35:35 UTC]

Created attachment 189086 [details]
The fragment of the build log

I pastebined it then copied it back from the pastebin

[Comment 2 Gordon Malm (RETIRED) 2009-04-23 02:17:03 UTC]

You didn't mention you were using a toolchain overlay on IRC that I recall. ;)

Please post your emerge --info.  I can't reproduce this compile error on x86 with in-tree sys-devel/gcc-4.3.3-r2 (pie,relro,now,fortify).  Haven't tested on amd64.

$ objdump -d /usr/sbin/exportfs | grep call | grep _chk | wc -l
59

I've added a URL to a bug the overlay had marked closed/worksforme.  CCing Magnus as he may know if its due to a bug in your overlay gcc (4.2.4 - which I think they abandoned).

[Comment 3 Gordon Malm (RETIRED) 2009-04-23 02:20:44 UTC]

Missed adding CC.

[Comment 4 James Browning 2009-04-23 02:53:19 UTC]

Given that I built it in early November and it's not listed in my eix output, probably. I didn't mention it then.  attachments arriving VSN.

[Comment 5 James Browning 2009-04-23 02:54:25 UTC]

Created attachment 189201 [details]
emerge -info output

[Comment 6 James Browning 2009-04-23 03:21:40 UTC]

Created attachment 189202 [details]
package list for the system.

[Comment 7 Gordon Malm (RETIRED) 2009-04-23 03:35:30 UTC]

Seeing as you are also on x86 I don't believe this to be a problem with our gcc.  Going to close as WORKSFORME.  Anyone feel free to re-open if they find it a problem on in-tree toolchain.

[Comment 8 James Browning 2009-04-26 22:21:50 UTC]

I just check again w/ the 4.3.3-r2 GCC from that overlay. It worked, okay guys overload the bozo bit relating to me.