(fwd from russh347@charter.net) Re: [distcc] Yet another compile failure

[Martin Pool]

(forwarded for the archives)

----- Forwarded message from Russel Hill <russh347 at charter.net> -----

From: Russel Hill <russh347 at charter.net>
Subject: Re: [distcc] Yet another compile failure
Date: Tue, 16 Sep 2003 19:16:21 -0700
To: Martin Pool <mbp at sourcefrog.net>
User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.5b) Gecko/20030827
X-Spam-Status: No, hits=-4.0 required=3.5
	tests=AWL,BAYES_10,EMAIL_ATTRIBUTION,FROM_ENDS_IN_NUMS,IN_REP_TO,
	      QUOTED_EMAIL_TEXT,REFERENCES,REPLY_WITH_QUOTES,
	      USER_AGENT_MOZILLA_UA
	version=2.55

Martin Pool wrote:
>On Fri 2003-09-12, Russel Hill wrote:
>This seems to show that distcc just ran this command directly on
>localhost.  Can you try running this by hand to see if it works?  Please
>make sure there is not more than one version of m68k-elf-g++ installed
>on the machine.  distcc might be using one from a directory that you
>didn't expect.

This turned out to be an issue in our makefiles... it was fine (somehow) 
for a serial make... a parallel make broke.  Once we dug into it, we 
couldn't figure out how it had ever worked.  Sorry about the red herring.

>>The second failing compile is strange:
>>==============================================================
>>g++ -c  -g -Wall -I.. -DDESKTOP -I/usr/include/stlport \
>>-ftemplate-depth-31 CameraSimTest.cpp -o desktop/CameraSimTest.o
>>
>>It's output
>>==============================================================
>>g++: Internal error: Killed (program cc1plus)
>>Please submit a full bug report.
>>See <URL:http://gcc.gnu.org/bugs.html> for instructions.
>>For Debian GNU/Linux specific bugs,
>>please see /usr/share/doc/debian/bug-reporting.txt.
>>
>>make[2]: *** [desktop/CameraSimTest.o] Error 1
>>make[1]: *** [all] Error 1
>>make: *** [camera] Error 2
>
>
>This is a gcc bug, not distcc.  There should be nothing (within reason)
>that distcc can do to gcc that will make it fail this way.  Please go
>ahead and follow their bug-reporting instructions.

I haven't been able to reproduce this reliably.  It isn't tied to a 
specific host.  All the machines are using the same version of gcc.

Thanks for all the help.  Sorry for the noise in the channel.

Russel Hill


----- End forwarded message -----