pcp
[Top] [All Lists]

Re: [pcp] pcp updates - #shell for pmcpp, gentoo fixes

To: Nathan Scott <nathans@xxxxxxxxxx>
Subject: Re: [pcp] pcp updates - #shell for pmcpp, gentoo fixes
From: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu, 13 Aug 2015 17:37:46 +1000
Cc: pcp@xxxxxxxxxxx
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <127430246.9646957.1439430874257.JavaMail.zimbra@xxxxxxxxxx>
References: <55CB2A34.2040903@xxxxxxxxxxxxxxxx> <127430246.9646957.1439430874257.JavaMail.zimbra@xxxxxxxxxx>
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.8.0
On 13/08/15 11:54, Nathan Scott wrote:
> Hi Ken,
> 
> ----- Original Message -----
>> Changes committed to git://git.pcp.io/kenj/pcp master
>>
>> Ken McDonell (8):
>>        qa/581: add another network interface pattern to the filter
>>        qa/188: add glibc-2.20 as OK for Gentoo
>>        qa/551: hide unhelpful chatter from kill
>>        qa/admin/check-vm: teach script about Gentoo packages
>>        configure.ac: make sure some optional gcc flags make it into the build
> 
> This one seems to have broken the Windows build:
> http://buildbot.pcp.io/builders/fedora-mingw64/builds/3/steps/shell/logs/stdio

Could be the configure.ac change, the others are not implicated.

I'm not sure where this first goes wrong ...

Here?
 
== Configuring pcp, log is in /home/bb/pcp-slave/fedora-mingw64/build/Logs/pcp 
(--prefix=/usr --sysconfdir=/etc --localstatedir=/var mingw64)
configure: WARNING: you should use --build, --host, --target

or here?

x86_64-w64-mingw32-gcc  -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions 
--param=ssp-buffer-size=4 -fno-strict-aliasing -fstack-protector-all 
-U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=2 -DPCP_INTERNAL '-DEXEC_SUFFIX=".exe"' 
'-DDSO_SUFFIX="dll"' -Wall -O2 -g -DPCP_DEBUG -DPCP_VERSION=\"3.10.7\" 
-I../../../src/include -I../../../src/include/pcp   -c -o connect.o connect.c

pulling this apart ....

-O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions 
--param=ssp-buffer-size=4 ... no clue where this comes from

-fno-strict-aliasing ... initial CFLAGS setting in configure for target_os = 
mingw?

-fstack-protector-all -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=2 ... from configure 
in the place I made a change

-O2 -g .... from $CFLAGS_OPTS

and the rest is sort of expected

or here?

I../../../src/include/pcp   -c -o p_result.o p_result.c
p_result.c: In function '__pmEncodeResult':
p_result.c:140:24: warning: cast from pointer to integer of different size 
[-Wpointer-to-int-cast]
      vlp = (vlist_t *)((__psint_t)vlp + sizeof(*vlp) + 
(j-1)*sizeof(vlp->vlist[0]));
                        ^
this looks like some serious snarfoo in the header files perhaps HAVE___PSINT_T 
is defined when it should not be.

All I can suggest is comparing Logs/pcp before and after if that is possible to 
see where the configure selections deviate, else if PCP was previously 
installed on this system running

        $ qa/admin/myconfigure

ends with a diff listing of key files in src/include compared to 
/usr/include/pcp which may be insightful.

<Prev in Thread] Current Thread [Next in Thread>