Bug 8843 - Kdeaddons 3.1beta2 compile fails on fresh 1.4rc1 install
Bug#: 8843 Product:  Gentoo Linux Version: 1.4_rc1 Platform: x86
OS/Version: Linux Status: RESOLVED Severity: normal Priority: P1
Resolution: FIXED Assigned To: danarmak@gentoo.org Reported By: gyula@graphite-works.com
Component: KDE
URL: 
Summary: Kdeaddons 3.1beta2 compile fails on fresh 1.4rc1 install
Keywords:  
Status Whiteboard: 
Opened: 2002-10-07 05:37 0000
Description:   Opened: 2002-10-07 05:37 0000
I have recently installed a fresh 1.4rc1 on my P4 box. Everything emerged fine
except kdeaddons 3.1beta2 which is producing an odd error, it ends up with:

i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I/usr/kde/3.1/include
-I/usr/qt/3/include -I/usr/X11R6/include  -I/usr/include/SDL -D_REENTRANT 
-DQT_THREAD_SUPPORT  -D_REENTRANT  -ansi -W -Wall -pedantic -Wshadow
-Wpointer-arith -Wmissing-prototypes -Wwrite-strings -D_XOPEN_SOURCE=500
-D_BSD_SOURCE -DNDEBUG -O2  -march=pentium4 -mcpu=pentium4 -O3 -pipe
-fomit-frame-pointer -c `test -f compute.c || echo './'`compute.c
compute.c:26: warning: no previous prototype for `fct'
compute.c: In function `generate_sector':
compute.c:175: unable to find a register to spill in class `FLOAT_REGS'
compute.c:175: this is the insn:
(insn 149 292 153 (set (reg:DF 9 st(1) [103])
        (float_extend:DF (subreg:SF (reg/v:DI 21 rxmm0 [79]) 0))) 133
{*extendsfdf2_1} (nil)
    (nil))
compute.c:175: confused by earlier errors, bailing out
make[3]: *** [compute.o] Error 1
make[3]: *** Waiting for unfinished jobs....
file.cpp: In function `void load_effects()':
file.cpp:65: warning: comparison between signed and unsigned integer
   expressions
file.cpp: In function `void load_random_effect(t_effect*)':
file.cpp:77: warning: comparison between signed and unsigned integer
   expressions
make[3]: Leaving directory
`/var/tmp/portage/kdeaddons-3.1_beta2/work/kdeaddons-3.0.8/noatun-plugins/tyler'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory
`/var/tmp/portage/kdeaddons-3.1_beta2/work/kdeaddons-3.0.8/noatun-plugins'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory
`/var/tmp/portage/kdeaddons-3.1_beta2/work/kdeaddons-3.0.8'
make: *** [all] Error 2

!!! ERROR: The ebuild did not complete successfully.
!!! Function kde_src_compile, Line -6658, Exitcode 2
!!! died running emake, kde_src_compile:make

I am using gcc 3.2 (extracted from the Pentium4 stage3 ball), my flags are as
follows:

CHOST='i686-pc-linux-gnu'
CFLAGS='-march=pentium4 -mcpu=pentium4 -O3 -pipe -fomit-frame-pointer'
CXXFLAGS='-march=pentium4 -mcpu=pentium4 -O3 -pipe -fomit-frame-pointer'

------- Comment #1 From Gyula Feher 2002-10-07 05:39:36 0000 -------
Created an attachment (id=4472) [details]
config.log file generated by the ebuild

------- Comment #2 From Dan Armak (RETIRED) 2002-10-08 02:08:23 0000 -------
This is the same as bug #7625, only that was reported for kdeaddons 3.0.3. It  
was fixed by not allowing -march=pentium4 and -sse2 to be used together. 
However the fix was only added to the 3.0.x ebuilds because noone confirmed it 
for the 3.1.x. tree at the time. Now that you have, I'll add the fix to those 
ebuilds as well. Will commit in a few minutes, please tell me if it helps. 

------- Comment #3 From Gyula Feher 2002-10-08 06:06:55 0000 -------
Thanks, it is making somehow further now and fails at a later stage. Please see
below what it ends up with, I am attaching the lates config.log as well.

i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I/usr/kde/3.1/include
-I/usr/qt/3/include -I/usr/X11R6/include  -I/usr/include/SDL -D_REENTRANT 
-DQT_THREAD_SUPPORT  -D_REENTRANT  -ansi -W -Wall -pedantic -Wshadow
-Wpointer-arith -Wmissing-prototypes -Wwrite-strings -D_XOPEN_SOURCE=500
-D_BSD_SOURCE -DNDEBUG -O2  -march=pentium4 -mcpu=pentium4 -O3 -pipe
-fomit-frame-pointer -c `test -f compute.c || echo './'`compute.c
compute.c:26: warning: no previous prototype for `fct'
compute.c: In function `generate_sector':
compute.c:175: unable to find a register to spill in class `FLOAT_REGS'
compute.c:175: this is the insn:
(insn 149 292 153 (set (reg:DF 9 st(1) [103])
        (float_extend:DF (subreg:SF (reg/v:DI 21 rxmm0 [79]) 0))) 133
{*extendsfdf2_1} (nil)
    (nil))
compute.c:175: confused by earlier errors, bailing out
make[3]: *** [compute.o] Error 1
make[3]: *** Waiting for unfinished jobs....
file.cpp: In function `void load_effects()':
file.cpp:65: warning: comparison between signed and unsigned integer
   expressions
file.cpp: In function `void load_random_effect(t_effect*)':
file.cpp:77: warning: comparison between signed and unsigned integer
   expressions
make[3]: Leaving directory
`/var/tmp/portage/kdeaddons-3.1_beta2/work/kdeaddons-3.0.8/noatun-plugins/tyler'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory
`/var/tmp/portage/kdeaddons-3.1_beta2/work/kdeaddons-3.0.8/noatun-plugins'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory
`/var/tmp/portage/kdeaddons-3.1_beta2/work/kdeaddons-3.0.8'
make: *** [all] Error 2

!!! ERROR: The ebuild did not complete successfully.
!!! Function kde_src_compile, Line -6658, Exitcode 2
!!! died running emake, kde_src_compile:make

------- Comment #4 From Gyula Feher 2002-10-08 06:09:26 0000 -------
Created an attachment (id=4509) [details]
config log with updated ebuild

------- Comment #5 From Dan Armak (RETIRED) 2002-10-08 07:40:01 0000 -------
Um, your second error log is identical to the first one AFAICS... Are you sure
you posted the right thing? 

------- Comment #6 From Gyula Feher 2002-10-08 08:19:10 0000 -------
Yikes, you are right. I did emerge rsync, commented out the mask, deleted
/var/tmp/portage/ to make sure I am using the newest ebiuld. Same result. 
I tried another thing, changed my flags to '-mcpu=i686 -O3 -pipe' to see if it
compiles with these. Unfortunately configure failed because it could not find Qt
(using my original p4 flags it finds it). Quite strange.


------- Comment #7 From Dan Armak (RETIRED) 2002-10-08 16:14:50 0000 -------
The answer to that (configure not finding qt) should be in the config.log for  
 that configuration.        I found the reason for it failing in the same way
again. I forgot to add  inherit flag-o-matic to the changed ebuilds, so their
call to is-flag() didn't do anything. I've fixed it now, please test once the
new version hits rsync. 

------- Comment #8 From Gyula Feher 2002-10-08 20:26:48 0000 -------
The new ebuild compiled smoothly :) Thanks a lot Dan!