Re: [blfs-support] Firefox-26.0 (with Xulrunner) vs. Glib-2.38.2

[James Richard Tyrer]

On 01/06/2014 05:09 PM, Armin K. wrote:
> On 7.1.2014 0:43, James Richard Tyrer wrote:
>> On 01/03/2014 08:07 PM, Armin K. wrote:
>>> On 01/04/2014 12:37 AM, James Richard Tyrer wrote:
>>>> On 01/03/2014 09:56 AM, Armin K. wrote:
>>>>> On 01/03/2014 05:29 PM, James Richard Tyrer wrote:
>>>>>> I am getting the error message:
>>>>>>
>>>>>> GLib-CRITICAL **: g_slice_set_config: assertion 'sys_page_size == 0' 
>>>>>> failed
>>>>>>
>>>>>> and Firefox does not open a window.
>>>>>>
>>>>>> It appears that Mozilla bug 833117 has not been fixed.
>>>>>>
>>>>>> https://bugzilla.mozilla.org/show_bug.cgi?id=833117
>>>>>>
>>>>>> Can you add a work around such as installing an older version of Glib
>>>>>> and linking Firefox to it?
>>>>>>
>>>>>>
>>>>>
>>>>> I am getting that message for more than a year now, but never had any
>>>>> issues that I'd notice. What is exactly a problem you have besides that
>>>>> message?
>>>>>
>>>> Although Firefox starts a process, it fails to open a window.
>>>>
>>>
>>> Hm, looking on some bugs, you might have a problem with this
>>>
>>> ac_add_options --enable-shared-js
>>>
>>> If mozillajs package is installed, xulrunner might try to use it instead
>>> of its own. This option looks like copy/paste from my xulrunner config
>>> from the time when shared mozillajs from xulrunner was required to build
>>> gnome-shell.
>>>
>>> I don't have that switch in my latest builds so that's maybe the reason
>>> I don't hit the bug. Can you try without the switch?
>>>
>> Well, I didn't get this done for a while because my system quit.  It
>> appears that I had some memory that was dieing and was only still
>> working because it was ECC.
>>
>> So, now I have rebuilt with the above change and it made no difference. :-(
>>
>> I am also getting other messages besides the error:
>>
>> (process:14840): GLib-CRITICAL **: g_slice_set_config: assertion
>> 'sys_page_size == 0' failed
>>
>> (firefox:14840): GLib-GObject-WARNING **: Attempt to add property
>> GnomeProgram::sm-connect after class was initialised
>>
>> (firefox:14840): GLib-GObject-WARNING **: Attempt to add property
>> GnomeProgram::show-crash-dialog after class was initialised
>>
>> (firefox:14840): GLib-GObject-WARNING **: Attempt to add property
>> GnomeProgram::display after class was initialised
>>
>> (firefox:14840): GLib-GObject-WARNING **: Attempt to add property
>> GnomeProgram::default-icon after class was initialised
>>
>> It appears that Firefox and GLib are not working together.
>>
>> IIUC, you are using GLib-2.38.2 with no problems although you get the
>> error message.
>>
>
> Do you use binary AMD driver for your GPU? Some sites seem to mention
> that it might be causing the issue. Mozilla bug says that this issue
> (but doesn't mention that it's related to fglrx) is fixed in firefox
> nightly (28.0, at the time I believe).
>

No, I am using xf86-video-ati built from source.

Perhaps I screw up one of the many libraries.  I will go through them 
when I get around to it.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] Firefox-26.0 (with Xulrunner) vs. Glib-2.38.2

[James Richard Tyrer]

On 01/03/2014 08:07 PM, Armin K. wrote:
> On 01/04/2014 12:37 AM, James Richard Tyrer wrote:
>> On 01/03/2014 09:56 AM, Armin K. wrote:
>>> On 01/03/2014 05:29 PM, James Richard Tyrer wrote:
>>>> I am getting the error message:
>>>>
>>>> GLib-CRITICAL **: g_slice_set_config: assertion 'sys_page_size == 0' failed
>>>>
>>>> and Firefox does not open a window.
>>>>
>>>> It appears that Mozilla bug 833117 has not been fixed.
>>>>
>>>> https://bugzilla.mozilla.org/show_bug.cgi?id=833117
>>>>
>>>> Can you add a work around such as installing an older version of Glib
>>>> and linking Firefox to it?
>>>>
>>>>
>>>
>>> I am getting that message for more than a year now, but never had any
>>> issues that I'd notice. What is exactly a problem you have besides that
>>> message?
>>>
>> Although Firefox starts a process, it fails to open a window.
>>
>
> Hm, looking on some bugs, you might have a problem with this
>
> ac_add_options --enable-shared-js
>
> If mozillajs package is installed, xulrunner might try to use it instead
> of its own. This option looks like copy/paste from my xulrunner config
> from the time when shared mozillajs from xulrunner was required to build
> gnome-shell.
>
> I don't have that switch in my latest builds so that's maybe the reason
> I don't hit the bug. Can you try without the switch?
>
Well, I didn't get this done for a while because my system quit.  It 
appears that I had some memory that was dieing and was only still 
working because it was ECC.

So, now I have rebuilt with the above change and it made no difference. :-(

I am also getting other messages besides the error:

(process:14840): GLib-CRITICAL **: g_slice_set_config: assertion 
'sys_page_size == 0' failed

(firefox:14840): GLib-GObject-WARNING **: Attempt to add property 
GnomeProgram::sm-connect after class was initialised

(firefox:14840): GLib-GObject-WARNING **: Attempt to add property 
GnomeProgram::show-crash-dialog after class was initialised

(firefox:14840): GLib-GObject-WARNING **: Attempt to add property 
GnomeProgram::display after class was initialised

(firefox:14840): GLib-GObject-WARNING **: Attempt to add property 
GnomeProgram::default-icon after class was initialised

It appears that Firefox and GLib are not working together.

IIUC, you are using GLib-2.38.2 with no problems although you get the 
error message.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] Firefox-26.0 (with Xulrunner) vs. Glib-2.38.2

[James Richard Tyrer]

On 01/03/2014 09:56 AM, Armin K. wrote:
> On 01/03/2014 05:29 PM, James Richard Tyrer wrote:
>> I am getting the error message:
>>
>> GLib-CRITICAL **: g_slice_set_config: assertion 'sys_page_size == 0' failed
>>
>> and Firefox does not open a window.
>>
>> It appears that Mozilla bug 833117 has not been fixed.
>>
>> https://bugzilla.mozilla.org/show_bug.cgi?id=833117
>>
>> Can you add a work around such as installing an older version of Glib
>> and linking Firefox to it?
>>
>>
>
> I am getting that message for more than a year now, but never had any
> issues that I'd notice. What is exactly a problem you have besides that
> message?
>
Although Firefox starts a process, it fails to open a window.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-support] Firefox-26.0 (with Xulrunner) vs. Glib-2.38.2

[James Richard Tyrer]

I am getting the error message:

GLib-CRITICAL **: g_slice_set_config: assertion 'sys_page_size == 0' failed

and Firefox does not open a window.

It appears that Mozilla bug 833117 has not been fixed.

https://bugzilla.mozilla.org/show_bug.cgi?id=833117

Can you add a work around such as installing an older version of Glib 
and linking Firefox to it?


-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] WebkitGTK for GIMP Help

[James Richard Tyrer]

On 04/08/2013 02:20 AM, Armin K. wrote:
> On 04/08/2013 08:06 AM, James Richard Tyrer wrote:
>>
>> Gimp is probably the main reason to build WebKitGTK+ NOT-2 with GTK+-2
>> and it already says: "WebKitGTK+-1.10.2 (this must be compiled with the
>> GTK+ 2 backend for the Gimp to use it as its help browser)".  Actually,
>> just listing the necessary configure flags would be shorter.  But, the
>> point is that it does not mention the need for the --disable-webkit2
>> option or the WebKit vs. WebKit2 issue.
>>
>
> Wrong. Gimp lists WebKitGTK as optional dep whereas Midori lists it as
> required - both require GTK+2 versions (Midori can use GTK+3 version,
> too - but it doesn't use it by default in version present in BLFS).
> Another package which requires GTK+2 version of WebKitGTK is GnuCash ...
> As I said once, instructions for other packages should remain on their
> pages. Gimp page clearly says you need GTK+2 version of WebKitGTK+ and
> WebKitGTK+ page clearly says how to build it.

Then I suppose that you better remove "(this must be compiled with the
 >> GTK+ 2 backend for the Gimp to use it as its help browser)" from the 
GIMP instructions and just leave readers guessing when they have 
problems -- when it doesn't work.  That is the purpose of a tutorial, 
isn't it.

Seriously, you can insist on being super right about everything and not 
even consider the suggestions and thoughts of other people.  But, you 
may find that you have the job all to yourself.


-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] Problem compiling Kde-baseapps-4.10.2, BLFS Version 2013-04-05

[James Richard Tyrer]

On 04/07/2013 11:15 PM, Niels Terp wrote:
> Hi,
>
> During compilation of kde-baseapps, all needed and all optional packages are
> found, and the compilation starts. But then:
>
> [  6%] Building CXX object
> dolphin/src/CMakeFiles/dolphinprivate.dir/kitemviews/kfileitemmodelrolesupda
> ter.o
> /home/niels/kde/kde-baseapps-4.10.2/dolphin/src/kitemviews/kfileitemmodelrol
> esupdater.cpp: In member function 'bool
> KFileItemModelRolesUpdater::applyResolvedRoles(const KFileItem&,
> KFileItemModelRolesUpdater::ResolveHint)':
> /home/niels/kde/kde-baseapps-4.10.2/dolphin/src/kitemviews/kfileitemmodelrol
> esupdater.cpp:999:42: fejl: 'const class KFileItem' has no member named
> 'isFinalIconKnown'
> make[2]: ***
> [dolphin/src/CMakeFiles/dolphinprivate.dir/kitemviews/kfileitemmodelrolesupd
> ater.o] Fejl 1
> make[1]: *** [dolphin/src/CMakeFiles/dolphinprivate.dir/all] Fejl 2
> make: *** [all] Fejl 2
>
> Any sugestions ?

If you have installed any libraries or modules in nonstandard location, 
I suggest that you set these 3 path environment variables with their 
locations.  You might want to try it even if they are standard locations 
due to where you have your source files:

CMAKE_INCLUDE_PATH

CMAKE_LIBRARY_PATH

CMAKE_PREFIX_PATH

You will then need to go to the build directory and:

rm CMakeCache.txt

and run cmake again to reconfigure.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-support] WebKitGTK+ NOT 2 vs. the patch.

[James Richard Tyrer]

So, I have tested building WebKitGTK+ (NOT 2) built against GTK+-2.24.17.

I tried it with the patch.  It demanded Automake-1.11 and them failed. 
So, I tried it without the patch.  It built OK.  However, there was a 
quirk installing.  Even though it was not built with GTK-Doc-1.x, it 
failed to install completely.  Strange.  With it, it installed OK.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] WebkitGTK for GIMP Help

[James Richard Tyrer]

On 04/07/2013 01:04 AM, Armin K. wrote:
> On 04/07/2013 09:41 AM, James Richard Tyrer wrote:
>> On 04/06/2013 11:51 PM, Armin K. wrote:
>>> On 04/07/2013 08:28 AM, James Richard Tyrer wrote:
>>>> The instructions for The GIMP correctly state that it needs Webkit built
>>>> with GTK-2.0 for the help browser.
>>>>
>>>> The problem is that while the current webkitgtk-1.10.2 configure script
>>>> help lists the parameter to choose version 2.0 or 3.0, when I choose
>>>> 2.0, I get an error message:
>>>>
>>>> configure: error: WebKit2 requires GTK+ 3.x, use --with-gtk=3.0
>>>>
>>>> I presume that there is a maximum version number that will compile
>>>> against GTK+-2.x.
>>>>
>>>> Does anybody know what that might be?  It is handy to have the 'Help' in
>>>> The GIMP.
>>>>
>>>
>>> --with-gtk=2.0 --disable-webkit2: These parameters force WebKitGTK+ to
>>> compile against GTK+ 2, even if GTK+ 3 is also installed. With GTK+ 2,
>>> everything WebKitGTK+ installs is suffixed with 1.0. When it is compiled
>>> against GTK+ 3 everything it installs is suffixed with 3.0. Both
>>> versions can be installed alongside one another with no namespace
>>> conflicts.
>>>
>>> It's in command explanations ... Read book, book good!
>>>
>> Yes, However, I was reading the GIMP page and it makes no mention of the
>> WebKit2 issue.  Perhaps it should show the two configure parameters as
>> you have shown.
>>
>
> Gimp has nothing to do with WebKit ... You need to follow instructions
> of the package which you build, not some other ... Sorry, I won't add
> one instructions to several other packages. I can reword Gimp page to
> say "With GTK+2 support."

Gimp is probably the main reason to build WebKitGTK+ NOT-2 with GTK+-2 
and it already says: "WebKitGTK+-1.10.2 (this must be compiled with the 
GTK+ 2 backend for the Gimp to use it as its help browser)".  Actually, 
just listing the necessary configure flags would be shorter.  But, the 
point is that it does not mention the need for the --disable-webkit2 
option or the WebKit vs. WebKit2 issue.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] WebkitGTK for GIMP Help

[James Richard Tyrer]

On 04/06/2013 11:51 PM, Armin K. wrote:
> On 04/07/2013 08:28 AM, James Richard Tyrer wrote:
>> The instructions for The GIMP correctly state that it needs Webkit built
>> with GTK-2.0 for the help browser.
>>
>> The problem is that while the current webkitgtk-1.10.2 configure script
>> help lists the parameter to choose version 2.0 or 3.0, when I choose
>> 2.0, I get an error message:
>>
>> configure: error: WebKit2 requires GTK+ 3.x, use --with-gtk=3.0
>>
>> I presume that there is a maximum version number that will compile
>> against GTK+-2.x.
>>
>> Does anybody know what that might be?  It is handy to have the 'Help' in
>> The GIMP.
>>
>
> --with-gtk=2.0 --disable-webkit2: These parameters force WebKitGTK+ to
> compile against GTK+ 2, even if GTK+ 3 is also installed. With GTK+ 2,
> everything WebKitGTK+ installs is suffixed with 1.0. When it is compiled
> against GTK+ 3 everything it installs is suffixed with 3.0. Both
> versions can be installed alongside one another with no namespace
> conflicts.
>
> It's in command explanations ... Read book, book good!
>
Yes, However, I was reading the GIMP page and it makes no mention of the 
WebKit2 issue.  Perhaps it should show the two configure parameters as 
you have shown.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-support] WebkitGTK for GIMP Help

[James Richard Tyrer]

The instructions for The GIMP correctly state that it needs Webkit built 
with GTK-2.0 for the help browser.

The problem is that while the current webkitgtk-1.10.2 configure script 
help lists the parameter to choose version 2.0 or 3.0, when I choose 
2.0, I get an error message:

configure: error: WebKit2 requires GTK+ 3.x, use --with-gtk=3.0

I presume that there is a maximum version number that will compile 
against GTK+-2.x.

Does anybody know what that might be?  It is handy to have the 'Help' in 
The GIMP.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] Can not compile WebKitGTK+-1.10.2

[James Richard Tyrer]

On 04/04/2013 11:18 AM, Matt Burgess wrote:
> On Thu, 2013-04-04 at 18:37 +0200, Armin K. wrote:
>
>> I believe the more correct term is "autotools variables" since CMake
>> also produces Makefiles :)
>
> The Autotools don't actually have anything to do with them, other than
> the fact that I think they can be used to override them at configure
> time.
>
> http://www.gnu.org/software/make/manual/html_node/Implicit-Variables.html#Implicit-Variables
>  has a list of all the variables that Make can use in its implicit rules.
>
> I'm pretty certain that other, non-GNU, Make implementations honour a
> similar list of variables.  The SUSV4 spec states:
>
> "The default rules for make shall achieve results that are the same as
> if the following were used. Implementations that do not support the
> C-Language Development Utilities option may omit CC, CFLAGS, YACC,
> YFLAGS, LEX, LFLAGS, LDFLAGS, and the .c, .y, and .l inference rules.
> Implementations that do not support FORTRAN may omit FC, FFLAGS, and
> the .f inference rules. Implementations may provide additional macros
> and rules."
>
> That suggests to me, that conforming Make implementations that *do*
> support the C-Language Development Utilities *must* support that
> particular list of variables.
>
>>From a quick browse of the cmake docs, it looks like it's a project
> similar to the autotools, in that it will produce 'native' Makefiles
> that can be run using GNU Make, and therefore CFLAGS & CXXFLAGS can be
> used with cmake too.
>
> Where the OPs C_FLAGS & CXX_FLAGS came from, I'm not sure :-)

Yes, it is CMake.  It has the variables CMAKE_C_FLAGS and CMAKE_CXX_FLAGS.

The problem is called retroactive inhibition (or interference) in 
learning theory.

http://en.wikipedia.org/wiki/Interference_theory#Retroactive_interference

"Retroactive interference

Retroactive interference (RI) is a phenomenon that occurs when newly 
learned information interferes with and impedes the recall of previously 
learned information."

I should have remembered that.  It is just one of the failings of the 
way your brain works.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] Can not compile WebKitGTK+-1.10.2

[James Richard Tyrer]

On 04/04/2013 09:37 AM, Armin K. wrote:
> On 04/04/2013 06:35 PM, Matt Burgess wrote:
>> On Thu, 2013-04-04 at 09:26 -0700, James Richard Tyrer wrote:
>>> On 04/02/2013 04:20 AM, James Richard Tyrer wrote:
>>>
>>>
>>>>export CXX_FLAGS="-L/lib -lrt"
>>>
>>>
>>> This should be CXXFLAGS.
>>>
>>> Obviously I must have used the correct variable since I got it to
>>> compile but my memory seems to be failing me lately; am I confusing this
>>> with CMake, or has this variable been changed from with the '_' to
>>> without it?
>>
>> CFLAGS and CXXFLAGS are, and always have been, Makefile variables.
>>
>> Whether CMake has decided to invent/use C_FLAGS/CXX_FLAGS I don't know;
>> I avoid it at all costs.
>>
>> Regards,
>>
>> Matt.
>>
>
> I believe the more correct term is "autotools variables" since CMake
> also produces Makefiles :)
>
> There is no need for setting cxxflags to fix this, it should be fixed in
> latest blfs svn ...
>
Yes, nice to know that it has been fixed.  However, considering how long 
it takes to compile, I doubt that I will do it over just to do it the 
more correct way.


-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] Can not compile WebKitGTK+-1.10.2

[James Richard Tyrer]

On 04/02/2013 04:20 AM, James Richard Tyrer wrote:


>   export CXX_FLAGS="-L/lib -lrt"


This should be CXXFLAGS.

Obviously I must have used the correct variable since I got it to 
compile but my memory seems to be failing me lately; am I confusing this 
with CMake, or has this variable been changed from with the '_' to 
without it?


-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] Can not compile WebKitGTK+-1.10.2

[James Richard Tyrer]

On 04/01/2013 11:59 AM, Niels Terp wrote:
> Hi,
>
> I have compiled all the required and the recommended dependencies, and all
> the optional but MathML.
>
> I then compile WebKitGTK+ straight from the book. After compiling for
> several hours, it stops with this error:
>
>CXXLD  Programs/WebKitPluginProcess
> /usr/bin/ld:
> Source/WebKit2/Platform/unix/Programs_WebKitPluginProcess-SharedMemoryUnix.o
> : undefined reference to symbol 'shm_open@@GLIBC_2.2'
> /usr/bin/ld: note: 'shm_open@@GLIBC_2.2' is defined in DSO /lib/librt.so.1
> so try adding it to the linker command line
> /lib/librt.so.1: could not read symbols: Invalid operation
> collect2: error: ld returned 1 exit status
> make[1]: *** [Programs/WebKitPluginProcess] Fejl 1
> make[1]: Forlader katalog '/home/niels/xc/webkitgtk-1.10.2'
>
> Can anybody help with this ?

I pursued it to compile.

Sorry that I can't remember if it was C or C++ code that it failed on so 
you will have to check.  Assuming that it was C++, you need to do this:

export CXX_FLAGS="-L/lib -lrt"

and start over from 'configure'.  If it was C rather than C++, then it 
would be C_FLAGS.  This trick can often be used to "nudge" something 
into compiling.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-support] Vala problems.

[James Richard Tyrer]

Various GNOME applications fail to compile against Vala with errors 
stating that files can not be found such as:

error: Package `GLib-2.0' not found in specified Vala API directories or 
GObject-Introspection GIR directories
error: Package `GObject-2.0' not found in specified Vala API directories 
or GObject-Introspection GIR directories
error: Package `Gio-2.0' not found in specified Vala API directories or 
GObject-Introspection GIR directories

If what is needed are: "*.deps" files, they are not in: 
"/usr/share/val-0.18/vapi" however there are: "*.vapi" files for these 
there.  I note that they are lower case.

Adding the configure parameter: "--enable-vala=no" makes it compile so 
it is obviously a Vala issue.

A Google search finds references to this problem, but I found no 
definite solution.  It was suggested on the Gentoo forum 
http:// that 
gsettings-desktop-schemas and dconf were necessary.  It didn't help and 
I can report the same problem with the command:

# gsettings list-schemas
Segmentation fault

I am not much for running those tests until something goes wrong but I 
did run: "make check" on Vala-0.18.1.  It did not pass.  There is a 
string of OKs and then:

   /dbus/bug596862: OK
   /dbus/bug602003/server:  OK
   /gir/bug651773:  FAIL
error: Package `GLib-2.0' not found in specified Vala API directories or 
GObject-Introspection GIR directories
error: Package `GObject-2.0' not found in specified Vala API directories 
or GObject-Introspection GIR directories
Generation failed: 2 error(s), 0 warning(s)
1 of 156 tests failed
make[4]: *** [check-TESTS] Error 1

There appears to be a problem here.

https://bugzilla.gnome.org/show_bug.cgi?id=651773

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-support] Xine-Lib-1.2.2 didn't build

[James Richard Tyrer]

To get Xine-Lib-1.2.2 to build, I had to use a file from the Mercurial 
repository on SF.


I attach a patch.

It is possible that more advanced processors don't need this.  The 
comment on the commit said that it was to fix a problem when only 4 
registers were available.


--
James Tyrer

Linux (mostly) From Scratch

diff -Naur 
xine-lib-1.2.2.old/src/post/deinterlace/plugins/greedy2frame_template_sse2.c 
xine-lib-1.2.2/src/post/deinterlace/plugins/greedy2frame_template_sse2.c
--- 
xine-lib-1.2.2.old/src/post/deinterlace/plugins/greedy2frame_template_sse2.c
2012-05-26 13:49:59.0 -0700
+++ xine-lib-1.2.2/src/post/deinterlace/plugins/greedy2frame_template_sse2.c
2012-06-30 23:59:00.0 -0700
@@ -84,11 +84,12 @@
 **   B0| |B1   | |
 */
 
-
+#if defined(ARCH_X86) || defined(ARCH_X86_64)
 static const sse_t Mask128 = { uq: { 0x7f7f7f7f7f7f7f7fll, 
0x7f7f7f7f7f7f7f7fll} };
 #define TP GREEDYTWOFRAMETHRESHOLD, GREEDYTWOFRAMETHRESHOLD2
 static const sse_t GreedyTwoFrameThreshold128 = { ub: {TP, TP, TP, TP, TP, TP, 
TP, TP} };
 #undef TP
+#endif
 
 static void DeinterlaceGreedy2Frame_SSE2(uint8_t *output, int outstride,
  deinterlace_frame_data_t *data,
@@ -170,10 +171,10 @@
 * See above for a description of the algorithm.
 * weave if (weave(M) AND (weave(T) OR weave(B)))
 */
-"movdqa  (%3), %%xmm1  \n\t" /* xmm1 = T1 */
-"movdqa  (%4), %%xmm0  \n\t" /* xmm0 = T0 */
-"movdqa  (%q5,%3), %%xmm3  \n\t" /* xmm3 = B1 */
-"movdqa  (%q5,%4), %%xmm2  \n\t" /* xmm2 = B0 */
+"movdqa  (%2), %%xmm1  \n\t" /* xmm1 = T1 */
+"movdqa  (%3), %%xmm0  \n\t" /* xmm0 = T0 */
+"movdqa  (%q4,%2), %%xmm3  \n\t" /* xmm3 = B1 */
+"movdqa  (%q4,%3), %%xmm2  \n\t" /* xmm2 = B0 */
 
 /* calculate |T1-T0| keep T1 put result in xmm5 */
 "movdqa  %%xmm1, %%xmm5\n\t"
@@ -181,20 +182,19 @@
 "psubusb %%xmm1, %%xmm0\n\t"
 "por %%xmm0, %%xmm5\n\t"
 
-"movdqa  (%0), %%xmm0  \n\t" /* xmm0 = M1 */
 /* T1 is data for line to copy */
-"movntdq  %%xmm1, %2   \n\t"
+"movntdq  %%xmm1, %1   \n\t"
 
 /* if |T1-T0| > Threshold we want 0 else dword minus one */
 "psrlw   $1, %%xmm5\n\t"
 "pand%%xmm6, %%xmm5\n\t"
-"pcmpgtb %1, %%xmm5\n\t"
+"pcmpgtb %0, %%xmm5\n\t"
 "pcmpeqd %%xmm7, %%xmm5\n\t"
 
-"prefetcht0  64(%q5,%3)\n\t"
-"prefetcht0  64(%q5,%4)\n\t"
+"prefetcht0  64(%q4,%2)\n\t"
+"prefetcht0  64(%q4,%3)\n\t"
   :
-  : "r" (M1), "m" (GreedyTwoFrameThreshold128),
+  : "m" (GreedyTwoFrameThreshold128),
 "m" (*Destc), "r" (T1), "r" (T0), "r" (Pitch) );
 
   asm volatile (
@@ -204,6 +204,7 @@
 "psubusb %%xmm3, %%xmm2\n\t"
 "por %%xmm2, %%xmm4\n\t"
 
+"movdqa  (%0), %%xmm0  \n\t" /* xmm0 = M1 */
 "movdqa  (%1), %%xmm2  \n\t" /* xmm2 = M0 */
 
 /* if |B1-B0| > Threshold we want 0 else dword minus one */
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-support] Open LDAP configuration warning

[James Richard Tyrer]

When I run 'configure' on OpenLDAP-2.4.31 I get this warning:

WARNING: Use of --without-threads is recommended with back-shell

Is this significant? and if so, what action is best?

I.E. Do I need the shell back-end?  Not using threads doesn't sound like 
a good solution.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] llvm in BLFS SVN

[James Richard Tyrer]

On 04/19/2012 04:56 PM, James Richard Tyrer wrote:

> I managed to get it to compile by making links.  So, it would seem
> reasonable to presume that the problem is Make files with the long
> locations for files.
>
Better make that: "wrong location for files".

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] llvm in BLFS SVN

[James Richard Tyrer]

On 04/14/2012 02:54 AM, Andrew Benton wrote:
> On Fri, 13 Apr 2012 20:58:55 +0100 James Richard
> Tyrer  wrote:
>
>> Do you have OCaml installed?
>
> No, I didn't, but since installing it I now get the same error as
> you. I got past that with these commands. I configure LLVM
> --with-ocaml-libdir=/usr/lib (ocaml has its .cmi files in
> /usr/lib/ocaml), after I run configure I changed one of the Makefiles
> with a sed and then copied META.llvm into the Release folder:
>
> sed -i 's#/ocaml$##;s#(OCAML_LIBDIR)$#&/ocaml#'
> bindings/ocaml/Makefile.ocaml mkdir bindings/ocaml/llvm/Release cp
> bindings/ocaml/llvm/{,Release/}META.llvm
>
> Basically the sed moves /ocaml from the end of one line onto the end
> of the previous line and it seems the make process doesn't copy
> META.llvm into the right place. It looks like buggy, unmaintained
> code to me, but what do I know ;)
>
> It compiled and installed Ok. I'm not sure if it works as I don't
> use ocaml. I only need LLVM as a dependency for Mesa.
>
I managed to get it to compile by making links.  So, it would seem
reasonable to presume that the problem is Make files with the long
locations for files.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-support] CLucene new release: 2.3.3.4

[James Richard Tyrer]

I had a problem building Strigi against CLucene-2.3.3.4.

After some research, I found that the problem was the PackageConfig 
file.  LibCLucene installs TWO SO libraries, but only listed one in the 
PC file. :-(


I found that adding it [attached] to the PC file that after a 
re-"configure" that it works.  So perhaps with this fix, it is not 
necessary to use the old version.

--
James Tyrer

Linux (mostly) From Scratch
prefix=/usr
exec_prefix=${prefix}/bin
libdir=${prefix}/lib
includedir=${prefix}/include:${prefix}/include/CLucene/ext

Name: libclucene
Description: CLucene - a C++ search engine, ported from the popular Apache 
Lucene
Version: 2.3.3.4
Libs: -L${prefix}/lib/ -lclucene-core -lclucene-shared
Cflags: -I${prefix}/include -I${prefix}/include/CLucene/ext
~
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] hicolor-icon-theme/MPlayer-svn_20120315 documentation Beyond Linux® From Scratch - Version 2012-04-11/Chapter 38. Video Utilities

[James Richard Tyrer]

On 04/18/2012 03:19 PM, rhubarbpie...@gmail.com wrote:
>
> Should hicolor-icon-theme be mentioned as a requirement in the "Skin
> Installation (Optional)" section?  If I compile the Blue skin (as
> per instructions) and access gmplayer I receive the following:
>
> Gtk-WARNING **: Could not find the icon 'mplayer.'  The 'hicolor'
> theme was not found either, perhaps you need to install it.
>
> I know little of MPlayer and installed only the Blue skin, but
> installing hicolor-icon-theme fixed the problem and gmplayer
> executed well.  hicolor-icon-theme is listed as Recommended in the
> gtk+ documentation.
>
You need to install the so-called hicolor-icon-theme as various
applications use it for icons even though it isn't really an icon theme.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] llvm in BLFS SVN

[James Richard Tyrer]

On 04/13/2012 09:04 AM, Andrew Benton wrote:
> On Fri, 13 Apr 2012 06:51:16 +0100
> James Richard Tyrer  wrote:
>
>> I was unable to build llvm-3.0.
>>
>> It failed with:
>>
>> make[3]: Entering directory `/usr/src/llvm-3.0.src/bindings/ocaml/llvm'
>> make[3]: *** No rule to make target
>> `/usr/src/llvm-3.0.src/Release/lib/ocaml/libLLVMCore.so', needed by
>> `build-deplibs'.  Stop.
>> make[3]: Leaving directory `/usr/src/llvm-3.0.src/bindings/ocaml/llvm'
>> make[2]: *** [all] Error 1
>> make[2]: Leaving directory `/usr/src/llvm-3.0.src/bindings/ocaml'
>>
>> This appears to be an error in a Makefine somewhere since the file
>> exists but it is in a different directory:
>>
>> /usr/src/llvm-3.0.src/Release/lib/libLLVMCore.so
>>
>> The same is true for other SO files so I can only speculate that it is
>> using:
>>
>>  ./Release/lib/ocaml
>>
>> instead of:
>>
>>  .Release/lib
>>
>> somewhere.  But, these Make files are too clever and I couldn't figure
>> it out.
>>
>> I also note that the: "configure" option: "--with-ocaml-libdir=" is
>> necessary if building OCaml bindings.  Should this be where OCaml is
>> installed?  The default appears to be for where LLVM will be installed:
>> "/opt/llvm/lib/ocaml".  "--prefix=" does not appear to change this.
>
> How did you configure llvm?

./configure --prefix=/opt/llvm --enable-targets=x86,x86_64 --enable-libffi

Then I ran the script code with the: "sed" instructions in the BLFS SVN 
page.

I decided to install in: "/opt/llvm" so I didn't use: 
"--with-ocaml-libdir=".

> If I grep through my build log for
> 'bindings/ocaml/llvm' the only mention of it is when it's creating a
> file called META.llvm. Ie, for me, it doesn't run make in the directory
> at all.
>
Do you have OCaml installed?

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] llvm in BLFS SVN

[James Richard Tyrer]

On 04/12/2012 10:50 PM, James Richard Tyrer wrote:

Addendum:

> I also note that the: "configure" option: "--with-ocaml-libdir=" is
> necessary if building OCaml bindings.  Should this be where OCaml is
> installed?  The default appears to be for where LLVM will be installed:
> "/opt/llvm/lib/ocaml".  "--prefix=" does not appear to change this.
>
Or, is there another reason that it needs to be installed in: 
"/opt/llvm".  Guess that I will play it safe and install it there.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-support] llvm in BLFS SVN

[James Richard Tyrer]

I was unable to build llvm-3.0.

It failed with:

make[3]: Entering directory `/usr/src/llvm-3.0.src/bindings/ocaml/llvm'
make[3]: *** No rule to make target 
`/usr/src/llvm-3.0.src/Release/lib/ocaml/libLLVMCore.so', needed by 
`build-deplibs'.  Stop.
make[3]: Leaving directory `/usr/src/llvm-3.0.src/bindings/ocaml/llvm'
make[2]: *** [all] Error 1
make[2]: Leaving directory `/usr/src/llvm-3.0.src/bindings/ocaml'

This appears to be an error in a Makefine somewhere since the file 
exists but it is in a different directory:

/usr/src/llvm-3.0.src/Release/lib/libLLVMCore.so

The same is true for other SO files so I can only speculate that it is 
using:

./Release/lib/ocaml

instead of:

.Release/lib

somewhere.  But, these Make files are too clever and I couldn't figure 
it out.

I also note that the: "configure" option: "--with-ocaml-libdir=" is 
necessary if building OCaml bindings.  Should this be where OCaml is 
installed?  The default appears to be for where LLVM will be installed: 
"/opt/llvm/lib/ocaml".  "--prefix=" does not appear to change this.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] DHCP-4.2.3-P2 Client Configuration

[James Richard Tyrer]

On 03/31/2012 02:07 PM, Bruce Dubbs wrote:
> James Richard Tyrer wrote:
>> On 03/29/2012 03:38 PM, Bruce Dubbs wrote:
>>> James Richard Tyrer wrote:
>>
>>>>>>>> I find that when I start : "dhclient" with the: "network" init.d script
>>>>>>>> that the host-name of my computer is changed to: "dhcppc".  I have
>>>>>>>> looked through the documentation and can not seem to find out how to
>>>>>>>> prevent this.
>>> You need to set up /etc/dhclient.conf.  Do not request a host-name.
>>>
>> OK, that sounds like a reasonable idea, but that is where I started.  I
>> have the file from the BLFS page: "DHCP-4.2.3-P2 Client Configuration"
>> installed and need to modify it.  I thought that all I needed to do was
>> remove: "host-name" and "domain-name" from the: "request" list.  The
>> problem is that this does not fix the issue.
>>
>>>>>>>> It would be nice if this page of the book addressed this problem.
>
> I can't get to this in a while.  Would you please create a ticket in
> BLFS fo rthis problem.
>
I had the time to run this down today.

The problem is occurring in the: "sbin/dhclient-script" script which has 
a conditional expression to reset the hostname.  Part of it is:

  [ x$current_hostname = xlocalhost ] || \

which results in the hostname being reset if it is the default: "localhost".

A good question is why am I using that hostname.  IIRC, there was a bug 
in WINE that was worked around by using that.

Is there a reason not to use: "localhost"? -- security?

Thanks for the help.  I have had health problems and I don't seem to be 
as sharp on these things as I usually am.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] DHCP-4.2.3-P2 Client Configuration

[James Richard Tyrer]

On 03/29/2012 03:38 PM, Bruce Dubbs wrote:
> James Richard Tyrer wrote:

>>>>>> I find that when I start : "dhclient" with the: "network" init.d script
>>>>>> that the host-name of my computer is changed to: "dhcppc".  I have
>>>>>> looked through the documentation and can not seem to find out how to
>>>>>> prevent this.
>
> You need to set up /etc/dhclient.conf.  Do not request a host-name.
>
OK, that sounds like a reasonable idea, but that is where I started.  I 
have the file from the BLFS page: "DHCP-4.2.3-P2 Client Configuration" 
installed and need to modify it.  I thought that all I needed to do was 
remove: "host-name" and "domain-name" from the: "request" list.  The 
problem is that this does not fix the issue.

>>>>>> It would be nice if this page of the book addressed this problem.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] DHCP-4.2.3-P2 Client Configuration

[James Richard Tyrer]

On 03/27/2012 05:40 AM, Ken Moffat wrote:
> On Tue, Mar 27, 2012 at 05:06:37AM -0700, James Richard Tyrer wrote:
>> On 03/24/2012 09:08 AM, Ken Moffat wrote:
>>> On Sat, Mar 24, 2012 at 12:33:05AM -0700, James Richard Tyrer wrote:
>>>> Small things can be very annoying department.
>>>>
>>>> I find that when I start : "dhclient" with the: "network" init.d script
>>>> that the host-name of my computer is changed to: "dhcppc".  I have
>>>> looked through the documentation and can not seem to find out how to
>>>> prevent this.
>>>>
>>>> It would be nice if this page of the book addressed this problem.
>>>
>>>I haven't seen this problem.  I use dhclient with the following in
>>> /etc/sysconfig/ifconfig.eth0:
>>>
>>> ONBOOT=yes
>>> IFACE=eth0
>>> SERVICE="dhclient"
>>> DHCP_START="-1 eth0"
>>> DHCP_STOP="-r"
>>>
>>>and my hostname is what I set in /etc/sysconfig/network.
>>>
>> Yes, I have:
>>
>> HOSTNAME=`cat /etc/hostname`
>>
>> in /etc/sysconfig/network.  But, dhclient isn't using that.
>>
>> Is it possible that something like:
>>
>>  hostname $HOSTNAME
>>
>> is being executed somewhere on your system?
>>
>   Well, the first difference is that I don't have /etc/hostname.

That appears to be an artifact from the fact that my system was formerly 
Fedora.  it is useful to only have the hostname in one file.

> But
> yes, /etc/rcS.d/S08localnet sources the network file and then runs
> hostname.  That's on an LFS-7.0 system, but this has worked for me
> for years and years (since at least LFS-6.3).
>
OK, I will look at that and see why it isn't setting the hostname for me.

>   Coincidentally, while I was googling for something else to do with
> dhcp yesterday, I came upon a page which claimed that dhclient does
> NOT use a default hostname if one is offered, in the way that e.g.
> dhcpcd apparently does.
>
>   From my current boot.log:
> 
> Mar 27 13:21:26 +01:00 (none)  Mounting virtual file systems: /run
> /proc /sys OK
> Mar 27 13:21:26 +01:00 (none)  Loading modules: i2c-viapro OK
> Mar 27 13:21:26 +01:00 (none)  Bringing up the loopback interface...
> OK
> Mar 27 13:21:26 +01:00 (none)  Setting hostname to bluesbreaker...
> OK
> Mar 27 13:21:26 +01:00 bluesbreaker  Populating /dev with device
> nodes...  OK
> Mar 27 13:21:29 +01:00 bluesbreaker  Activating all swap
> files/partitions... OK
> Mar 27 13:21:29 +01:00 bluesbreaker  Mounting root file system in
> read-only mode...  OK
> Mar 27 13:21:29 +01:00 bluesbreaker  Checking file systems... OK
> Mar 27 13:21:30 +01:00 bluesbreaker  Remounting root file system in
> read-write mode... OK
> Mar 27 13:21:30 +01:00 bluesbreaker  Recording existing mounts in
> /etc/mtab... OK
> Mar 27 13:21:30 +01:00 bluesbreaker  Mounting remaining file
> systems... OK
> Mar 27 13:21:30 +01:00 bluesbreaker  Cleaning file systems: /tmp OK
> Mar 27 13:21:30 +01:00 bluesbreaker  Creating files and
> directories...  OK
> Enabling ondemand cpu frequency
> Mar 27 13:21:30 +01:00 bluesbreaker  Retrying failed uevents, if
> any... OK
> Mar 27 13:21:30 +01:00 bluesbreaker  Setting up Linux console... OK
> Mar 27 13:21:31 +01:00 bluesbreaker  Starting system log daemon...
> OK
> Mar 27 13:21:31 +01:00 bluesbreaker  Starting kernel log daemon...
> OK
> Mar 27 13:21:32 +01:00 bluesbreaker  Starting the D-Bus Messagebus
> Daemon... OK
> Mar 27 13:21:32 +01:00 bluesbreaker  Bringing up the eth0 interface:
> Mar 27 13:21:32 +01:00 bluesbreaker
> Starting dhclient on the eth0 interface... OK
> Mar 27 13:21:36 +01:00 bluesbreaker  Starting rpcbind OK
> (and so forth)
>
This is odd.  It is setting the hostname before running dhclient.

I have a CD and it works when I boot from it so I think that I will have 
a look at the scripts on it.
-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] DHCP-4.2.3-P2 Client Configuration

[James Richard Tyrer]

On 03/24/2012 08:54 AM, Bruce Dubbs wrote:
> James Richard Tyrer wrote:
>> Small things can be very annoying department.
>>
>> I find that when I start : "dhclient" with the: "network" init.d script
>> that the host-name of my computer is changed to: "dhcppc".  I have
>> looked through the documentation and can not seem to find out how to
>> prevent this.
>>
>> It would be nice if this page of the book addressed this problem.
>
> Isn't that covered in the dhclient man page using th e-F or -H options?
>
I tried the -H option and I get an error message containing:

Usage: dhclient [-4|-6] [-SNTP1dvrx] [-nw] [-p ] [-D LL|LLT]
 [-s server-addr] [-cf config-file] [-lf lease-file]
 [-pf pid-file] [--no-pid] [-e VAR=val]
 [-sf script-file] [interface]

and it fail to start.  It appears from the error message that the -H 
option isn't supported.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-support] DHCP-4.2.3-P2 Client Configuration

[James Richard Tyrer]

On 03/24/2012 09:08 AM, Ken Moffat wrote:
> On Sat, Mar 24, 2012 at 12:33:05AM -0700, James Richard Tyrer wrote:
>> Small things can be very annoying department.
>>
>> I find that when I start : "dhclient" with the: "network" init.d script
>> that the host-name of my computer is changed to: "dhcppc".  I have
>> looked through the documentation and can not seem to find out how to
>> prevent this.
>>
>> It would be nice if this page of the book addressed this problem.
>
>   I haven't seen this problem.  I use dhclient with the following in
> /etc/sysconfig/ifconfig.eth0:
>
> ONBOOT=yes
> IFACE=eth0
> SERVICE="dhclient"
> DHCP_START="-1 eth0"
> DHCP_STOP="-r"
>
>   and my hostname is what I set in /etc/sysconfig/network.
>
Yes, I have:

HOSTNAME=`cat /etc/hostname`

in /etc/sysconfig/network.  But, dhclient isn't using that.

Is it possible that something like:

hostname $HOSTNAME

is being executed somewhere on your system?

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-support] DHCP-4.2.3-P2 Client Configuration

[James Richard Tyrer]

Small things can be very annoying department.

I find that when I start : "dhclient" with the: "network" init.d script 
that the host-name of my computer is changed to: "dhcppc".  I have 
looked through the documentation and can not seem to find out how to 
prevent this.

It would be nice if this page of the book addressed this problem.
-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

exiv2 vs. expat

[James Richard Tyrer]

Using BLFS  - Version svn-20110417

Attempt to configure exiv2-0.21.1 ends with:

checking whether to enable XMP support (requires expat)... yes
checking expat.h usability... no
checking expat.h presence... yes
configure: WARNING: expat.h: present but cannot be compiled
configure: WARNING: expat.h: check for missing prerequisite headers?
configure: WARNING: expat.h: see the Autoconf documentation
configure: WARNING: expat.h: section "Present But Cannot Be Compiled"
configure: WARNING: expat.h: proceeding with the compiler's result
configure: WARNING: ## -- ##
configure: WARNING: ## Report this to ahug...@gmx.net ##
configure: WARNING: ## -- ##
checking for expat.h... no

I checked and: "expat.h" is in: "/usr/include".

I have expat-2.0.1 installed.

-- 
James Tyrer

Linux (mostly) From Scratch

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Session-end KDE-4.5.3 ?

[James Richard Tyrer]

On 11/24/2010 02:29 PM, Dr.-Ing. Edgar Alwers wrote:
> Hi,
>
> probably a trivial question. I have built KDE-4.5.3 and I am starting the 
> system through SLIM. Start is OK, but when I want to finish my work and enter 
> the application launcher menu, I get leave->  End Session ->  ...and the 
> launcher menu closes. Nothing more happens. The same more or less as user: 
> launcher menu->leave and nothing more.
> I think, I have missed something during the building/installing process. But 
> what ? I am looking to end the session and get the Slim menu again, as it 
> does in my old KDE system. Any hint ?
>
The problem is probably that your process chain is broken on the return 
side.  What ever the front end, it should run a script: "Xsession" which 
should contain the command:

 exec /opt/KDE-4.5/bin/startkde

{with the path for your system}

Note that you must use the full path.

Then: "startkde" will replace the: "Xsession" process and when it ends 
it should be the same as a return from: "Xsession"

If you are not using the: "exec" command, you should have:

 /opt/KDE-4.5/bin/startkde
 end

IAC, the problem is probably in the: "Xsession" script.

-- 

James Tyrer

Linux (mostly) From Scratch

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: computer wont mumble

[James Richard Tyrer]

  On 09/04/10 14:51, luxInteg wrote:
> but I cant get  even a mumble
>
> It also has  pulse-audio, and kde-4.4.5
> but all I get  in the latter  is a  sound device called dummy.
>
> which is quite strange as I disabled   'dummy' for alsa in the kernel
> compilation
>
> a little advice would be appreciated
There have been reports of KDE-4.x.y no being compatible with PulseAudio 
dispute that fact that it is supposed to be able to use it as the sound 
server.

IAC, if PulseAudio is being used, it must be running.  Open System 
Monitor and look for it.

Then open system settings and go to the section for sound configuration 
and look at the window showing "Output Device Preference".  Select 
Music.  If you want to try PulseAudio, it should be at the top.  OSS 
should always be at the bottom.  It it isn't working, kill the 
PulseAudio daemon and more ALSA to the top.  To use the same settings 
for everything, use the: "Apply Device List To ..." button.

-- 
James Tyrer

Linux (mostly) From Scratch

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: BLFS SVN: Ed-0.8

[James Richard Tyrer]

  On 08/28/10 15:34, Bruce Dubbs wrote:
> James Richard Tyrer wrote:
>>The FTP link to ed-0.8.tar.gz is dead.  All that is available in that
>> directory is>=1.0.  Is there any reason not to use version 1.4?
> To answer your question with a question:  Is there any reason to use ed?
>
> The answer is to satisfy a build requirement of some ancient package.
> If the new version will satisfy that package's requirement, then yes,
> it's OK.
>
> Doing a quick grep of the book, the only references I can see are tetex
> and Xorg libraries.  Taking a look at Xorg libraries, a quick check
> indicates that sed is now being used and not ed, so ed there may just be
>a holdover from xfree86.
>
> BTW, the version of ed in BLFS svn 1.4.  You shouldn't be using the
> "stable" version.  It's just too far out of date.
>
Interesting.  I thought that I had SVN open, but it appears that I have 
the stable version.  How, I managed that is something that I am not 
clear on.

-- 
James Tyrer

Linux (mostly) From Scratch

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: polkit docs vs. DocBook

[James Richard Tyrer]

 On 08/28/10 03:43, Andrew Benton wrote:

On 28/08/10 01:35, James Richard Tyrer wrote:

I had to reinstall version 4.2 of:

   DocBook SGML DTD
   DocBook XML DTD

to build KDE-4.5 and now I seem to be missing something.  When building:
"polkit-0.97" I have an error stating that it can't find the necessary
XSL file:

I/O error : Attempt to load network entity
http://docbook.sourceforge.net/release/xsl/current/manpages/docbook.xsl

I installed: docbook-xsl-1.75.2 according to the BLFS instructions.

I checked and the file:

/usr/share/xml/docbook/xsl-stylesheets-1.75.2/manpages/docbook.xsl

exists.  I checked the catalog:

/etc/xml/catalog

and there is a line:

http://docbook.sourceforge.net/release/xsl/current";
rewritePrefix="/usr/share/xml/docbook/xsl-stylesheets-1.75.2"/>

which appears to be correct.

The command which is failing is:

/usr/bin/xsltproc -nonet --stringparam man.base.url.for.relative.links
/usr/share/gtk-doc/html/polkit-1/ --xinclude
http://docbook.sourceforge.net/release/xsl/current/manpages/docbook.xsl
polkit.xml

If I remove: "-nonet", then it uses the file from the internet and works
correctly.  I have checked and the local file is identical.

Obviously, I am missing something.


Try:

xmlcatalog --noout --add "delegateSystem" \
   "http://docbook.sourceforge.net/release/xsl/current/"; \
   "/usr/share/xml/docbook/docbook-xsl-1.75.2/catalog.xml" /etc/xml/catalog
xmlcatalog --noout --add "delegateURI" \
   "http://docbook.sourceforge.net/release/xsl/current/"; \
   "/usr/share/xml/docbook/docbook-xsl-1.75.2/catalog.xml" /etc/xml/catalog

I can't remember why I started adding those "delegate" entries (lost in
the time before I started using git). If that doesn't help, can we see
the whole of /etc/xml/catalog please?

Andy

See attached.

--
James Tyrer

Linux (mostly) From Scratch


http://www.oasis-open.org/committees/entity/release/1.0/catalog.dtd";>

  
  
  http://www.oasis-open.org/docbook/"; catalog="file:///etc/xml/docbook"/>
  http://www.oasis-open.org/docbook/"; catalog="file:///etc/xml/docbook"/>
  http://oasis-open.org/docbook/xml/4.1.2/"; catalog="file:///etc/xml/docbook"/>
  http://oasis-open.org/docbook/xml/4.1.2/"; catalog="file:///etc/xml/docbook"/>
  http://www.oasis-open.org/docbook/xml/4.3/"; catalog="file:///etc/xml/docbook"/>
  http://www.oasis-open.org/docbook/xml/4.3/"; catalog="file:///etc/xml/docbook"/>
  http://www.oasis-open.org/docbook/xml/4.4/"; catalog="file:///etc/xml/docbook"/>
  http://www.oasis-open.org/docbook/xml/4.4/"; catalog="file:///etc/xml/docbook"/>
  http://docbook.sourceforge.net/release/xsl/1.75.2"; rewritePrefix="/usr/share/xml/docbook/xsl-stylesheets-1.75.2"/>
  http://docbook.sourceforge.net/release/xsl/1.75.2"; rewritePrefix="/usr/share/xml/docbook/xsl-stylesheets-1.75.2"/>
  http://docbook.sourceforge.net/release/xsl/current"; rewritePrefix="/usr/share/xml/docbook/xsl-stylesheets-1.75.2"/>
  http://docbook.sourceforge.net/release/xsl/current"; rewritePrefix="/usr/share/xml/docbook/xsl-stylesheets-1.75.2"/>

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

BLFS SVN: Ed-0.8

[James Richard Tyrer]

  The FTP link to ed-0.8.tar.gz is dead.  All that is available in that 
directory is >=1.0.  Is there any reason not to use version 1.4?

-- 
James Tyrer

Linux (mostly) From Scratch

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: polkit docs vs. DocBook

[James Richard Tyrer]

  On 08/28/10 03:43, Andrew Benton wrote:
> On 28/08/10 01:35, James Richard Tyrer wrote:
>> I had to reinstall version 4.2 of:
>>
>>DocBook SGML DTD
>>DocBook XML DTD
>>
>> to build KDE-4.5 and now I seem to be missing something.  When building:
>> "polkit-0.97" I have an error stating that it can't find the necessary
>> XSL file:
>>
>> I/O error : Attempt to load network entity
>> http://docbook.sourceforge.net/release/xsl/current/manpages/docbook.xsl
>>
>> I installed: docbook-xsl-1.75.2 according to the BLFS instructions.
>>
>> I checked and the file:
>>
>> /usr/share/xml/docbook/xsl-stylesheets-1.75.2/manpages/docbook.xsl
>>
>> exists.  I checked the catalog:
>>
>> /etc/xml/catalog
>>
>> and there is a line:
>>
>> > uriStartString="http://docbook.sourceforge.net/release/xsl/current";
>> rewritePrefix="/usr/share/xml/docbook/xsl-stylesheets-1.75.2"/>
>>
>> which appears to be correct.
>>
>> The command which is failing is:
>>
>> /usr/bin/xsltproc -nonet --stringparam man.base.url.for.relative.links
>> /usr/share/gtk-doc/html/polkit-1/ --xinclude
>> http://docbook.sourceforge.net/release/xsl/current/manpages/docbook.xsl
>> polkit.xml
>>
>> If I remove: "-nonet", then it uses the file from the internet and works
>> correctly.  I have checked and the local file is identical.
>>
>> Obviously, I am missing something.
>>
> Try:
>
> xmlcatalog --noout --add "delegateSystem" \
>"http://docbook.sourceforge.net/release/xsl/current/"; \
>"/usr/share/xml/docbook/docbook-xsl-1.75.2/catalog.xml" /etc/xml/catalog
> xmlcatalog --noout --add "delegateURI" \
>"http://docbook.sourceforge.net/release/xsl/current/"; \
>"/usr/share/xml/docbook/docbook-xsl-1.75.2/catalog.xml" /etc/xml/catalog
>
> I can't remember why I started adding those "delegate" entries (lost in
> the time before I started using git). If that doesn't help, can we see
> the whole of /etc/xml/catalog please?
Thanks Andy.  I found that I did make some errors trying to install both 
4.2 and 4.5 of the DTDs.  So, I am doing a reinstall.  So, I will get 
back to you in a while.

-- 
James Tyrer

Linux (mostly) From Scratch

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: polkit docs vs. DocBook

[James Richard Tyrer]

  On 08/27/10 21:17, Ken Moffat wrote:
> On 28 August 2010 01:35, James Richard Tyrer  wrote:
>>   I had to reinstall version 4.2 of:
>>
>>  DocBook SGML DTD
>>  DocBook XML DTD
>>
>> to build KDE-4.5 and now I seem to be missing something.  When building:
>> "polkit-0.97" I have an error stating that it can't find the necessary
>> XSL file:
>>
>> I/O error : Attempt to load network entity
>> http://docbook.sourceforge.net/release/xsl/current/manpages/docbook.xsl
>>
>> I installed: docbook-xsl-1.75.2 according to the BLFS instructions.
>>
>> I checked and the file:
>>
>> /usr/share/xml/docbook/xsl-stylesheets-1.75.2/manpages/docbook.xsl
>>
>> exists.  I checked the catalog:
>>
>> /etc/xml/catalog
>>
>> and there is a line:
>>
>> > uriStartString="http://docbook.sourceforge.net/release/xsl/current";
>> rewritePrefix="/usr/share/xml/docbook/xsl-stylesheets-1.75.2"/>
>>
>> which appears to be correct.
>>
>> The command which is failing is:
>>
>> /usr/bin/xsltproc -nonet --stringparam man.base.url.for.relative.links
>> /usr/share/gtk-doc/html/polkit-1/ --xinclude
>> http://docbook.sourceforge.net/release/xsl/current/manpages/docbook.xsl
>> polkit.xml
>>
>> If I remove: "-nonet", then it uses the file from the internet and works
>> correctly.  I have checked and the local file is identical.
>>
>> Obviously, I am missing something.
>>
>> --
>   Dunno, docbook is a pain if it isn't set up correctly.
Yes!  And I have to say that I don't really understand it.

> But in my xml
> catalog here, on the line before that I have
>
>systemIdStartString="http://docbook.sourceforge.net/release/xsl/current";
> rewritePrefix="/usr/share/xml/docbook/xsl-stylesheets-1.73.2"/>
>
> (so, I haven't updated my stylesheets) - maybe you missed the
> rewrite System line (the third of the additions to xmlcatalog ?
Yes, I have that one to.  I was guessing that last one was the one that 
should be used.  I ran the configuration script at the bottom of the 
page is SVN BLFS.

My best guess is that the catalog isn't the problem since it was copied 
from the BLFS page.

-- 
James Tyrer

Linux (mostly) From Scratch

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

polkit docs vs. DocBook

[James Richard Tyrer]

  I had to reinstall version 4.2 of:

 DocBook SGML DTD
 DocBook XML DTD

to build KDE-4.5 and now I seem to be missing something.  When building: 
"polkit-0.97" I have an error stating that it can't find the necessary 
XSL file:

I/O error : Attempt to load network entity 
http://docbook.sourceforge.net/release/xsl/current/manpages/docbook.xsl

I installed: docbook-xsl-1.75.2 according to the BLFS instructions.

I checked and the file:

/usr/share/xml/docbook/xsl-stylesheets-1.75.2/manpages/docbook.xsl

exists.  I checked the catalog:

/etc/xml/catalog

and there is a line:

http://docbook.sourceforge.net/release/xsl/current"; 
rewritePrefix="/usr/share/xml/docbook/xsl-stylesheets-1.75.2"/>

which appears to be correct.

The command which is failing is:

/usr/bin/xsltproc -nonet --stringparam man.base.url.for.relative.links 
/usr/share/gtk-doc/html/polkit-1/ --xinclude 
http://docbook.sourceforge.net/release/xsl/current/manpages/docbook.xsl 
polkit.xml

If I remove: "-nonet", then it uses the file from the internet and works 
correctly.  I have checked and the local file is identical.

Obviously, I am missing something.

-- 
James Tyrer

Linux (mostly) From Scratch

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: ATLAS build problems

[James Richard Tyrer]

On 05/05/10 17:29, James Richard Tyrer wrote:

> So, I will try it with "--nof77" first.  If that works, I will know that
> it is a FORTRAN issue.  Will report in a day or so.
>
Nope, didn't work.  It finished with the same error message as before 
about not finding "main".

So, I guess I will stick with 3.8 and wait for 3.10.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: ATLAS build problems

[James Richard Tyrer]

On 05/05/10 15:50, David Jensen wrote:
> On Wed, 05 May 2010 13:55:49 -0700 James Richard
> Tyrer  wrote:
>
>> I have a strange error when trying to build ATLAS 3.9.24:
>>
>> make[4]: Entering directory `/usr/src/ATLAS/LINUX-ELF/tune/sysinfo'
>> gcc -c -DL2SIZE=4194304 -I/usr/src/ATLAS/LINUX-ELF/include
>> -I/usr/src/ATLAS/LINUX-ELF/..//include
>> -I/usr/src/ATLAS/LINUX-ELF/..//include/contrib -DAdd_
>> -DF77_INTEGER=int -DStringSunStyle -DATL_OS_Linux -DATL_ARCH_K7
>> -DATL_CPUMHZ=1100 -DATL_SSE1 -DATL_3DNow -DATL_GAS_x8632
>> -fomit-frame-pointer -mfpmath=387 -O2 -falign-loops=4 -fPIC -m32
>> /usr/src/ATLAS/LINUX-ELF/..//tune/sysinfo/findNT.c gcc
>> -DL2SIZE=4194304 -I/usr/src/ATLAS/LINUX-ELF/include
>> -I/usr/src/ATLAS/LINUX-ELF/..//include
>> -I/usr/src/ATLAS/LINUX-ELF/..//include/contrib -DAdd_
>> -DF77_INTEGER=int -DStringSunStyle -DATL_OS_Linux -DATL_ARCH_K7
>> -DATL_CPUMHZ=1100 -DATL_SSE1 -DATL_3DNow -DATL_GAS_x8632
>> -fomit-frame-pointer -mfpmath=387 -O2 -falign-loops=4 -fPIC -m32 -o
>> xfindNT findNT.o ATL_walltime.o -lm
>> /usr/lib/gcc/i686-pc-linux-gnu/4.4.4/../../../crt1.o: In function
>> `_start': (.text+0x18): undefined reference to `main' collect2: ld
>> returned 1 exit status make[4]: *** [xfindNT] Error 1 make[4]:
>> Leaving directory `/usr/src/ATLAS/LINUX-ELF/tune/sysinfo' make[3]:
>> *** [/usr/src/ATLAS/LINUX-ELF/include/atlas_pthreads.h] Error 2
>> make[3]: Leaving directory `/usr/src/ATLAS/LINUX-ELF/src/auxil'
>> make[2]: *** [IBuildPtlibs0] Error 2 make[2]: Leaving directory
>> `/usr/src/ATLAS/LINUX-ELF/bin'
>>
>> Any suggestions will be appreciated.
>>
Thanks.
>
> main is in findNT.c,

Right, that is why the error is so strange.

> it compliled above without errors.  main will not build if ATL_NCPU
> is undefined.  try: grep -n ATL_NCPU
> LINUX-ELF/include/atlas_buildinfo.h if it's there, I fold.
>
I don't have a file: "atlas_buildinfo.h". :-(

> Maybe corrupt download, maybe Fortran, never used it?
>
Have tried downloading again twice.  Have read bad things about 
gfortran.  But ATLAS now used it instead of f77.

> Builds fine here, x86_64, gcc 4.4.3, 'No Fortran' ../configure
> --nof77 make

So, I will try it with "--nof77" first.  If that works, I will know that 
it is a FORTRAN issue.  Will report in a day or so.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

ATLAS build problems

[James Richard Tyrer]

I have a strange error when trying to build ATLAS 3.9.24:

make[4]: Entering directory `/usr/src/ATLAS/LINUX-ELF/tune/sysinfo'
gcc -c -DL2SIZE=4194304 -I/usr/src/ATLAS/LINUX-ELF/include 
-I/usr/src/ATLAS/LINUX-ELF/..//include 
-I/usr/src/ATLAS/LINUX-ELF/..//include/contrib -DAdd_ -DF77_INTEGER=int 
-DStringSunStyle -DATL_OS_Linux -DATL_ARCH_K7 -DATL_CPUMHZ=1100 
-DATL_SSE1 -DATL_3DNow -DATL_GAS_x8632  -fomit-frame-pointer 
-mfpmath=387 -O2 -falign-loops=4 -fPIC -m32 
/usr/src/ATLAS/LINUX-ELF/..//tune/sysinfo/findNT.c
gcc -DL2SIZE=4194304 -I/usr/src/ATLAS/LINUX-ELF/include 
-I/usr/src/ATLAS/LINUX-ELF/..//include 
-I/usr/src/ATLAS/LINUX-ELF/..//include/contrib -DAdd_ -DF77_INTEGER=int 
-DStringSunStyle -DATL_OS_Linux -DATL_ARCH_K7 -DATL_CPUMHZ=1100 
-DATL_SSE1 -DATL_3DNow -DATL_GAS_x8632  -fomit-frame-pointer 
-mfpmath=387 -O2 -falign-loops=4 -fPIC -m32 -o xfindNT findNT.o 
ATL_walltime.o -lm
/usr/lib/gcc/i686-pc-linux-gnu/4.4.4/../../../crt1.o: In function `_start':
(.text+0x18): undefined reference to `main'
collect2: ld returned 1 exit status
make[4]: *** [xfindNT] Error 1
make[4]: Leaving directory `/usr/src/ATLAS/LINUX-ELF/tune/sysinfo'
make[3]: *** [/usr/src/ATLAS/LINUX-ELF/include/atlas_pthreads.h] Error 2
make[3]: Leaving directory `/usr/src/ATLAS/LINUX-ELF/src/auxil'
make[2]: *** [IBuildPtlibs0] Error 2
make[2]: Leaving directory `/usr/src/ATLAS/LINUX-ELF/bin'

Any suggestions will be appreciated.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: help required kdeedu-4.4.2

[James Richard Tyrer]

On 04/27/10 10:25, lux-integ wrote:
> On Tuesday 27 April 2010 09:44:43 am lux-integ wrote:
>
>> /bin/sh: --slave: command not found
>> make[2]: *** [cantor/src/backends/R/rserver/CMakeFiles/rautoloads] Error
>> 127 make[2]: Leaving directory `$SOURCES/kdeedu-4.4.2/build'
>> make[1]: *** [cantor/src/backends/R/rserver/CMakeFiles/rautoloads.dir/all]
>> Error 2
>> make[1]: Leaving directory `$SOURCES/kdeedu-4.4.2/build'
>> make: *** [all] Error 2
>>
> I had  nudgings the above could be due to not finding the binary file R.
> There are two ~/bin directories for the installation of R I am using
> (/usr/local/R-2.10.1/) and (/use/local/R-2.10.1/lib64/R/bin); see below:-
>
> poet [ ~ ]$ ls /usr/local/R-2.10.1/bin
> R  Rscript
> poet [ ~ ]$
> poet [ ~ ]$
> poet [ ~ ]$ ls /use/local/R-2.10.1/lib64/R/bin
> BATCHINSTALL  R   RcmdRdconv  RprofSHLIB  Stangle  build
> config  f77_f2c libtoolpager
> COMPILE  LINK REMOVE  Rd2dvi  Rdiff   Rscript  Sd2Rd  Sweave   check  exec
> javareconf  mkinstalldirs  rtags
> poet [ ~ ]$
> poet [ ~ ]$
> (exec is a directory containing the R binary ~7Mbytes in size)
>
>   I targeted the latter (set as R_HOME)  from  where  my FindR.cmake script
> found R, Rscript and   the file (R/a-compiled-binary) in  the directory exe).
> Running cmake gives:-
>
> 
> -- Building with gpsd position provider
> -- Found Readline: /usr/lib64/libncurses.so
> -- Found:/usr/local/R-2.10.1/lib64/R/bin/R
> -- Found:/usr/local/R-2.10.1/lib64/R/bin/Rscript
> -- Found:/usr/local/R-2.10.1/lib64/R/bin/exec/R
> -- Found:/usr/local/R-2.10.1/lib64/R/include/R.h
> -- Found:/usr/local/R-2.10.1/lib64/R/modules/R_X11.so
> -- Found:/usr/local/R-2.10.1/lib64/R/modules/internet.so
> -- Found:/usr/local/R-2.10.1/lib64/R/modules/vfonts.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/cluster/libs/cluster.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/foreign/libs/foreign.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/grDevices/libs/grDevices.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/grid/libs/grid.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/KernSmooth/libs/KernSmooth.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/lattice/libs/lattice.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/MASS/libs/MASS.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/Matrix/libs/Matrix.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/methods/libs/methods.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/mgcv/libs/mgcv.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/nlme/libs/nlme.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/nnet/libs/nnet.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/spatial/libs/spatial.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/splines/libs/splines.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/stats/libs/stats.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/survival/libs/survival.so
> -- Found:/usr/local/R-2.10.1/lib64/R/library/tools/libs/tools.so
> -- Found:/usr/lib64/libgfortran.a
> -- Found:/usr/local/R-2.10.1/lib64/R/lib/libRlapack.so
> -- Found:/usr/local/R-2.10.1/lib64/R/modules/lapack.so
> -- Found:/usr/local/R-2.10.1/lib64/R/include/R_ext/Lapack.h
> Yuw should be chuffed matey;  R2 has been compiled wiv LAPAC
> -- Found:/usr/local/R-2.10.1/lib64/R/lib/libRblas.so
> -- Found:/usr/local/R-2.10.1/lib64/R/include/R_ext/BLAS.h
> YIPPI;  R2 has been compiled with BLAS
> -- Found R: /usr/local/R-2.10.1/lib64/R:
> libs: /usr/local/R-2.10.1/lib64/R/lib/libRblas.so
> and /usr/local/R-2.10.1/lib64/R/library/tools/libs/tools.so
> -- WARNING: you are using the obsolete 'PKGCONFIG' macro use FindPkgConfig
> -- Found LibSpectre: -lspectre
> -- Found Eigen2: /opt/kde-4.4.2/include/eigen2
> -- Found OCaml: /usr/bin/ocamlc
>
> #
> all executable files targeted are  'found'   but  'make' still fails at 79%
>
> If I am targeting the wrong R's  ( or indeed the right ones) I would be
> grateful if someone on list could let me know.

If you have a 64 bit machine, then the lib64 libraries should be the 
correct ones.

I presume that CMake found the: " .../pkgconfig/libR.pc" file, and then 
it choose which libraries to use.

I am starting to think that there is -- that you have found -- a build 
bug with 64 bit libraries.  Or, it is possible that the problem is due 
to the fact that you have installed ATLAS or LAPAC.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: help required kdeedu-4.4.2

[James Richard Tyrer]

On 04/26/10 17:03, lux-integ wrote:
> On Monday 26 April 2010 11:05:03 pm lux-integ wrote:
>> [ 79%] Built target cantor_rbackend
>> Scanning dependencies of target renvvars
>> /bin/sh: --slave: command not found
>> make[2]: *** [cantor/src/backends/R/rserver/CMakeFiles/renvvars] Error 127
>> make[1]: *** [cantor/src/backends/R/rserver/CMakeFiles/renvvars.dir/all]
>> Error 2
>
> there does not seem to be a CMakeFiles directory:-

That will be in your build directory.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: kde4 and CMAKE_PREFIX_PATH

[James Richard Tyrer]

On 04/25/10 15:00, lux-integ wrote:
> Greeings,
>
> I am trying to use cmake to find R2- (an optional program used in
> kdeedu-4.4.2).  I can find libraries and headers but I dont seem to be able
> set CMAKE_PREFIX_PATH  to find  $R_HOME/bin/R.
>
And I don't think that you are going to be able to because R doesn't use 
the regular *NIX install path structure.

I suggest that you set:

CMAKE_INCLUDE_PATH
CMAKE_LIBRARY_PATH

to point to the "include" and "lib" paths.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Problem in cmake on kdeadmin-4.4.2

[James Richard Tyrer]

On 04/25/10 15:03, cliffhan...@gardener.com wrote:
>
>
> -Original Message-
> From: James Richard Tyrer 
> To: BLFS Support List 
> Cc: cliffhan...@gardener.com
> Sent: Sun, Apr 25, 2010 9:38 pm
> Subject: Re: Problem in cmake on kdeadmin-4.4.2
>
> On 04/25/10 09:03, cliffhan...@gardener.com wrote:
>>
>>
>> -----Original Message-
>> From: James Richard Tyrer 
>> To: BLFS Support List 
>> Sent: Sun, Apr 25, 2010 4:37 pm
>> Subject: Re: Problem in cmake on kdeadmin-4.4.2
>>
>> On 04/25/10 07:35, cliffhan...@gardener.com wrote:
>>> -Original Message- From: James Richard Tyrer
>>
>>>> The actual finding is done by a Python script: "FindPyCups.py" in:
>>>
>>>> .../kdeadmin/system-config-printer-kde/cmake-modules
>>>
>>>> I would try running that manually. Change to that directory and:
>>>
>>>> python FindPyCups.py
>>>
>>>> If I read it correctly, it prints: "Groovy" to standard output if
>>>> it
>>> is
>>>> found. If not, the easiest way to check to see that PyCups is
>>> installed
>>>> correctly is to install it again.
>>>
>>> Doing this returns me to prompt, with no response
>>>
>>> OK, that is the problem. PyCups is not installed correctly
>>
>> Maybe, but see below. echo $? returned 1 by the way.
>>
>>> Maybe i need to install Pycups elsewhere as suggested by lux-integ?
>>>
>>> AFAIK, you can't do that. It will automatically be installed with the
>>> same prefix as Python.
>>
>>> I've tried editing FindSystemConfigPrinter.py with
>>> path/to/python/site-packages/cupshelpers - this doesn't help with
>>> system-config-printer either.
>>>
>>> That isn't going to help if the Python script doesn't find (py)cups.
>>
>>> So, I would try installing PyCups again
>>
>> Thanks James, done that twice now and the same two files, cups.so and
>> cups-1.0-py2.6.egg-info are installed under
>> /usr/lib/python2.6/site-packages. These are the only two expected I
>> assume? A file list on the web says yes.
>>
>> This makes little sense. It may be related to having Python installed
>> in: "/usr/local" but Python should know where it is installed. But,
>> perhaps, some symbolic links to: "/usr" could help
>
> No luck here, even tried a new FindPyCups.cmake script with advice from
> lux-integ but no go
>
>  >/usr/bin/python -> /usr/local/bin/python
>  >/usr/lib/python -> /usr/local/lib/python
>  >/usr/include/python -> /usr/local/include/python
>
>> IIRC, I had to do that when I still had Python installed in:
> "/usr/local".
>
>> If it was CMake that couldn't find something, you can set the two
> CMake
>> environment variables that very roughly correspond to the normal ones
>> for AutoTools (but ONLY for specifying paths!):
>
>  > CPPFLAGS CMAKE_INCLUDE_PATH
>  > LD_FLAGS CMAKE_LIBRARY_PATH
>
> Also tried this, but cmake will not find cups.so
>
I am running out of ideas. :-(

Is it possible that PyCups requires other Python addons that you don't 
have installed?

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Problem in cmake on kdeadmin-4.4.2

[James Richard Tyrer]

On 04/25/10 09:03, cliffhan...@gardener.com wrote:
>
>
> -Original Message-
> From: James Richard Tyrer 
> To: BLFS Support List 
> Sent: Sun, Apr 25, 2010 4:37 pm
> Subject: Re: Problem in cmake on kdeadmin-4.4.2
>
> On 04/25/10 07:35, cliffhan...@gardener.com wrote:
>> -Original Message- From: James Richard Tyrer
>
>>> The actual finding is done by a Python script: "FindPyCups.py" in:
>>
>>> .../kdeadmin/system-config-printer-kde/cmake-modules
>>
>>> I would try running that manually. Change to that directory and:
>>
>>> python FindPyCups.py
>>
>>> If I read it correctly, it prints: "Groovy" to standard output if
>>> it
>> is
>>> found. If not, the easiest way to check to see that PyCups is
>> installed
>>> correctly is to install it again.
>>
>> Doing this returns me to prompt, with no response
>>
>> OK, that is the problem. PyCups is not installed correctly
>
> Maybe, but see below. echo $? returned 1 by the way.
>
>> Maybe i need to install Pycups elsewhere as suggested by lux-integ?
>>
>> AFAIK, you can't do that. It will automatically be installed with the
>> same prefix as Python.
>
>> I've tried editing FindSystemConfigPrinter.py with
>> path/to/python/site-packages/cupshelpers - this doesn't help with
>> system-config-printer either.
>>
>> That isn't going to help if the Python script doesn't find (py)cups.
>
>> So, I would try installing PyCups again
>
> Thanks James, done that twice now and the same two files, cups.so and
> cups-1.0-py2.6.egg-info are installed under
> /usr/lib/python2.6/site-packages. These are the only two expected I
> assume? A file list on the web says yes.
>
This makes little sense.  It may be related to having Python installed 
in: "/usr/local" but Python should know where it is installed.  But, 
perhaps, some symbolic links to: "/usr" could help.

/usr/bin/python -> /usr/local/bin/python
/usr/lib/python -> /usr/local/lib/python
/usr/include/python -> /usr/local/include/python

IIRC, I had to do that when I still had Python installed in: "/usr/local".

If it was CMake that couldn't find something, you can set the two CMake 
environment variables that very roughly correspond to the normal ones 
for AutoTools (but ONLY for specifying paths!):

CPPFLAGS CMAKE_INCLUDE_PATH
LD_FLAGS CMAKE_LIBRARY_PATH

and they use the same syntax as other PATH environment variables.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Problem in cmake on kdeadmin-4.4.2

[James Richard Tyrer]

On 04/25/10 07:35, cliffhan...@gardener.com wrote:
> -Original Message- From: James Richard Tyrer

>> The actual finding is done by a Python script: "FindPyCups.py" in:
>
>> .../kdeadmin/system-config-printer-kde/cmake-modules
>
>> I would try running that manually.  Change to that directory and:
>
>> python FindPyCups.py
>
>> If I read it correctly, it prints: "Groovy" to standard output if
>> it
> is
>> found.  If not, the easiest way to check to see that PyCups is
> installed
>> correctly is to install it again.
>
> Doing this returns me to prompt, with no response
>
OK, that is the problem.  PyCups is not installed correctly.

> Maybe i need to install Pycups elsewhere as suggested by lux-integ?
>
AFAIK, you can't do that.  It will automatically be installed with the
same prefix as Python.

> I've tried editing FindSystemConfigPrinter.py with
> path/to/python/site-packages/cupshelpers - this doesn't help with
> system-config-printer either.
>
That isn't going to help if the Python script doesn't find (py)cups.

So, I would try installing PyCups again.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Problem in cmake on kdeadmin-4.4.2

[James Richard Tyrer]

On 04/24/10 10:14, cliffhan...@gardener.com wrote:
> Hi
>
> I encountered this problem when running cmake on kdeadmin-4.4.2:-
>
>
> -
> --
> -- -- The following OPTIONAL packages could NOT be located on your
> system.
>   -- Consider installing them to enable more features from this software.
>
> -
> --
> -- * PyCups
> PyCups was not found.  It is needed by system-config-printer-kde
> to
>   run. (Use -DINSTALL_SYSTEM_CONFIG_PRINTER=TRUE to install anyway)
>* system-config-printer>
> 
> system-config-printer was not found.  Some of its modules
> (cupshelpers.py, config.py, smburi.py and debug.py) are required by
>   system-config-printer-kde.
>
>
> -
> --
>
> But these are installed:-
>
>   ###  for pycups  (installed is version 1.9.49) - a recent version
>
>   root [ /usr/lib/python2.6/site-packages ]#
>   root [ /usr/lib/python2.6/site-packages ]#
>   root [ /usr/lib/python2.6/site-packages ]# ls  cups*
>   cups-1.0-py2.6.egg-info  cups.so
>   root [ /usr/lib/python2.6/site-packages ]#
>
> for system config printer  (installed is version 1.2.1- ) -the
> latest
>   available
>
>   [ /usr/share/system-config-printer ]# ls -l
>   total 924
>   -rw-r--r-- 1 root root  19177 Apr 17 21:29 applet.glade
>   -rwxr-xr-x 1 root root  11891 Apr 17 21:29 applet.py
>   -rw-r--r-- 1 root root   8756 Apr 17 21:29 authconn.py
>   -rw-r--r-- 1 root root   1766 Apr 17 21:29 config.py
>   -rw-r--r-- 1 root root   6427 Apr 17 21:29 contextmenu.py
>   -rw-r--r-- 1 root root  25074 Apr 17 21:29 cupshelpers.py
>   -rw-r--r-- 1 root root   1703 Apr 17 21:29 debug.py
>   -rw-r--r-- 1 root root   3374 Apr 17 21:29 errordialogs.py
>   -rw-r--r-- 1 root root   1651 Apr 17 21:29 gtk_label_autowrap.py
>   -rw-r--r-- 1 root root   8046 Apr 17 21:29 gtk_treeviewtooltips.py
>   drwxr-xr-x 2 root root   4096 Apr 17 21:29 icons
>   -rw-r--r-- 1 root root  39096 Apr 17 21:29 jobviewer.py
>   -rw-r--r-- 1 root root  21724 Apr 17 21:29 monitor.py
>   -rwxr-xr-x 1 root root   9259 Apr 17 21:29 my-default-printer.py
>   -rw-r--r-- 1 root root  14601 Apr 17 21:29 openprinting.py
>   -rw-r--r-- 1 root root  12898 Apr 17 21:29 options.py
>   -rw-r--r-- 1 root root   7424 Apr 17 21:29 optionwidgets.py
>   -rwxr-xr-x 1 root root  31054 Apr 17 21:29 ppds.py
>   -rw-r--r-- 1 root root   3678 Apr 17 21:29 probe_printer.py
>   -rwxr-xr-x 1 root root  13780 Apr 17 21:29 pysmb.py
>   -rw-r--r-- 1 root root   3027 Apr 17 21:29 smburi.py
>   -rw-r--r-- 1 root root   5064 Apr 17 21:29 statereason.py
>   -rw-r--r-- 1 root root 443717 Apr 17 21:29 system-config-printer.glade
>   -rwxr-xr-x 1 root root 189264 Apr 17 21:29 system-config-printer.py
>   drwxr-xr-x 2 root root   4096 Apr 17 21:29 troubleshoot
>   root [ /usr/share/system-config-printer ]#
>
> Any ideas here, tried changing .py files etc
>
This is strange.  Probably can't be much help since I can't reproduce 
the problem.

I have that version of PyCups installed and using the KDE-4.4 SVN BRANCH 
I get:

-- Found PythonInterp: /usr/bin/python2.6
-- Found Python executable: /usr/bin/python2.6
-- Found Python version: 2.6.4
-- Found Python library: /usr/lib/libpython2.6.so
-- Found SIP version: 4.10
-- Found PyQt4 version: 4.7
-- Found PyKDE4 version 4.4.2 /usr/share/sip/PyKDE4
-- Found PyCups

Does your CMake output list the previous Python packages as found?

The actual finding is done by a Python script: "FindPyCups.py" in:

.../kdeadmin/system-config-printer-kde/cmake-modules

I would try running that manually.  Change to that directory and:

python FindPyCups.py

If I read it correctly, it prints: "Groovy" to standard output if it is 
found.  If not, the easiest way to check to see that PyCups is installed 
correctly is to install it again.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: gst-plugins-good vs. TabLib; compile issue

[James Richard Tyrer]

On 04/23/10 22:47, Trent Shea wrote:
> On April 23, 2010 11:24:21 pm Trent Shea wrote:
>> On a LFS-6.6 32bit host I don't see a difference in the output of
>> ./configure HAVE_CXX="yes" and ./configure.
> Actually looking back, is this variable for taglibs or gst-plugins-good?
>
It is to build the TabLib plugin in gst-plugins-good.  It appears to be 
to override the failure of the: 'configure' script to properly detect 
that you have a C++ compiler installed.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: gst-plugins-good vs. TabLib; compile issue

[James Richard Tyrer]

On 04/23/10 16:10, Ken Moffat wrote:
> On 23 April 2010 22:41, James Richard Tyrer  wrote:
>> I find, as I noted here:
>>
>> http://techbase.kde.org/Getting_Started/Build/KDE4/LFS#TagLib
>>
>> that GST Plugins Good will only build the TagLib plugin if I set:
>>
>> HAVE_CXX="yes".  I was wondering if other people have this issue and if
>> it should be added to the BLFS SVN.
>>
>
>   As with many things in wikis, it's hard to know when it was added,
> and whether it is still correct (or, indeed, if it ever was - I don't doubt
> that whoever noted it had a problem which it solved, but I question
> if it was the right fix).
>
>   I've given up on kde4, but I do still build taglib-1.5 and in my log for
> gst-plugins-good-0.10.16 (which is what I'm currently using) I see
> configure: *** checking feature: taglib tagging library ***
> configure: *** for plug-ins: taglib ***
> checking for TAGLIB... yes
> checking to see if c++ compiler understands -Wno-attributes... yes
> configure: *** These plugins will be built: taglib
>
> FWIW, HAVE_CXX in gst-plugins-good overrides the tests that for me
> give the following output:
> checking for g++... g++
> checking whether we are using the GNU C++ compiler... yes
> checking whether g++ accepts -g... yes
> checking dependency style of g++... gcc3
> checking whether we are using the GNU C++ compiler... (cached) yes
> checking whether g++ accepts -g... (cached) yes
> checking dependency style of g++... (cached) gcc3
> checking how to run the C++ preprocessor... g++ -E
> checking for ld used by g++... /usr/bin/ld -m elf_x86_64
> checking if the linker (/usr/bin/ld -m elf_x86_64) is GNU ld... yes
> checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports
> shared libraries... yes
> checking for g++ option to produce PIC... -fPIC -DPIC
> checking if g++ PIC flag -fPIC -DPIC works... yes
> checking if g++ static flag -static works... yes
> checking if g++ supports -c -o file.o... yes
> checking if g++ supports -c -o file.o... (cached) yes
> checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports
> shared libraries... yes
> checking dynamic linker characteristics... GNU/Linux ld.so
> checking how to hardcode library paths into programs... immediate
> checking for g++... yes
> checking how to run the C++ preprocessor... g++ -E
>
>   - that makes me think that any LFS system needing HAVE_CXX would
> need it in many, or most, c++ packages which use autotools (and I
> would expect cmake to be equally problematic, but I neither know nor
> care about that  any more).
>
> OTOH, if *your* system actually needs that variable to be set, then
> I stand corrected.  If that is the case, which version of LFS, and have
> you upgraded gcc or added a newer version of gcc after you built it ?
>
I have recently upgraded to LFS SVN.  I am using GCC-4.3.4.  The plugin 
for TagLib isn't built unless I set HAVE_CXX to "yes".

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

gst-plugins-good vs. TabLib; compile issue

[James Richard Tyrer]

I find, as I noted here:

http://techbase.kde.org/Getting_Started/Build/KDE4/LFS#TagLib

that GST Plugins Good will only build the TagLib plugin if I set:

HAVE_CXX="yes".  I was wondering if other people have this issue and if 
it should be added to the BLFS SVN.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: NetworkManager working on Linux

[James Richard Tyrer]

On 04/21/10 03:55, Simon Geard wrote:
> On Tue, 2010-04-20 at 14:57 -0700, James Richard Tyrer wrote:
>> As I mentioned elsewhere, I had an issue building it, but I now have
>> version 0.8 of it and the applet installed.
>>
>> To work, you need to add a basic configuration file:
>>
>>
>> /etc/NetworkManager/nm-system-settings.conf
>
> The file's not actually required - NM will run just fine without it,

Do you have version 0.8 installed?  Previous versions didn't seem to 
need it, but I couldn't get them to work. :-|  But, perhaps it isn't 
always needed, although the documentation at GNOME Live states that it is.

> though it's probably necessary in order to define system connections
> (i.e ones not tied to a logged-in user).

It is necessary to define which plugin(s) you are using.  And, I have 
previously had issues with NetworkManager changing my hostname.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: NetworkManager working on Linux

[James Richard Tyrer]

On 04/20/10 14:57, James Richard Tyrer wrote:

> There is an issue with the applet. The default policy means that it will
> only run as root. Some Googling and I found how to set this up so that
> the applet will be on the system tray and will work for user accounts.
> These two patches [attached] based on having a: "network" group and
> making the needed accounts members of it work for me. Files to patch are
> in "/etc.../dbus-1/system.d".
>
Also, the applet will not start unless the icon:

/usr/share/icons/hicolor/22x22/apps/nm-secure-lock.png

exists.  So, if you installed with a different prefix, make a link.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

NetworkManager working on Linux

[James Richard Tyrer]

As I mentioned elsewhere, I had an issue building it, but I now have 
version 0.8 of it and the applet installed.


To work, you need to add a basic configuration file:


/etc/NetworkManager/nm-system-settings.conf
--8<--8<--8<--8<--8<--8<--8<--8<--
[main]
plugins=keyfile

no-auto-default=00:12:17:5a:5e:35,

[keyfile]
hostname=

--8<--8<--8<--8<--8<--8<--8<--8<--

I note that the documentation says that this file name will be changed 
in the future.


If running KDE, you also need to copy: "nm-applet.desktop" from your 
GNOME autostart directory to your KDE autostart directory.


You need to setup your init.d scripts to start the NetworkManager daemon 
at boot.


After doing this, you need to reboot your system (or you can go to level 
1 and back).


Now, you should be able to run: "nm-connection-editor" as root.

If you have Ethernet, "Add" a "Wired" connection.  If you have a DHCP 
Client installed, all you should need to do is supply the: "MAC 
Address".  You can find this with the console application: "nm-tool". 
Be sure to check: "Connect automatically" and "Available to all users".


Now, your Ethernet should automatically startup when you plugin the 
cable as someone requested.  I don't do that, but I tried it and it 
worked.  What I do need is for it to starup at boot and to restart if my 
DSL connection is interrupted.


There is an issue with the applet.  The default policy means that it 
will only run as root.  Some Googling and I found how to set this up so 
that the applet will be on the system tray and will work for user 
accounts.  These two patches [attached] based on having a: "network" 
group and making the needed accounts members of it work for me.  Files 
to patch are in "/etc.../dbus-1/system.d".


Note that because these are system configurations that you will need to 
reboot again before the changes take effect.


I still have issues.  Stuff that should be logged is showing up on VTs. 
 I will have to read more of the documentation (which is terrible or 
nonexistent).


--
James Tyrer

Linux (mostly) From Scratch
--- NetworkManager.conf.old	2010-04-19 02:58:27.0 -0700
+++ NetworkManager.conf	2010-04-19 02:58:27.0 -0700
@@ -12,6 +12,10 @@
 
 
+	
+
+
+
 
 
 
--- nm-applet.conf.old	2010-04-19 03:00:48.0 -0700
+++ nm-applet.conf	2010-04-19 03:00:48.0 -0700
@@ -23,6 +23,12 @@
 
 
+	
+		
+		
+		
+		
+	
 
 
 
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Fwd: Re: Errors compiling kdebase-runtime 4.4(SOLVED)

[James Richard Tyrer]

On 04/20/10 12:13, cliffhan...@gardener.com wrote:
>
>
> -Original Message- From: cliffhan...@gardener.com To:
> onkel.j...@t-online.de Sent: Tue, Apr 20, 2010 8:12 pm Subject: Re:
> Errors compiling kdebase-runtime 4.4(SOLVED)
>
>
>
> -Original Message- From:
> onkel.j...@t-online.de To:
> cliffhan...@gardener.com Sent: Tue, Apr 20, 2010 12:19 pm Subject:
> Re: Errors compiling kdebase-runtime 4.4
>
> -Original-Nachricht- Subject: Re: Errors compiling
> kdebase-runtime 4.4 Date: Thu, 15 Apr 2010 18:30:08 +0200 From:
> cliffhan...@gardener.com To: onkel.j...@t-online.de
>
>
>
> -Original Message- From: Thomas
> Glaesser To:
> blfs-support@linuxfromscratch.org Sent: Thu, Apr 15, 2010 4:45 pm
> Subject: Re: Errors compiling kdebase-runtime 4.4
>
> > > Hi > > Has anyone seen this unusual error when
> compling kdebase-runtime 4.4 on > LFS6.5 > > Apparently gcc
> thinks it's allocating the actual abstract objects, > instead of
> an > array of pointers and reports the following error: "cannot
> allocate an > object > of abstract type '...'". > > Well,
> I'm not a programmer so how does one solve this? > >
> /home/sources/kdebase-runtime-4.4.2/attica/kdeplugin/kdeplatformdependent
>
>
>
>
> > > .cpp: > In function 'QObject* qt_plugin_instance()':
> >
> /home/sources/kdebase-runtime-4.4.2/attica/kdeplugin/kdeplatformdependent
>
>
>
>
> > > .cpp:225: > error: cannot allocate an object of abstract
> type > 'Attica::KdePlatformDependent' >
> /home/sources/kdebase-runtime-4.4.2/attica/kdeplugin/kdeplatformdependent
>
>
>
>
> > > .h:42: > note: because the following virtual functions
> are pure within > 'Attica::KdePlatformDependent': >
> /opt/kde/include/attica/platformdependent.h:47: note: virtual void
> > Attica::PlatformDependent::addDefaultProviderFile(const QUrl&)
> > /opt/kde/include/attica/platformdependent.h:48: note: virtual
> void > Attica::PlatformDependent::removeDefaultProviderFile(const
> QUrl&) > /opt/kde/include/attica/platformdependent.h:53: note:
> virtual void > Attica::PlatformDependent::enableProvider(const
> QUrl&, bool) const >
> /opt/kde/include/attica/platformdependent.h:54: note: virtual bool
> > Attica::PlatformDependent::isEnabled(const QUrl&) const >
> make[2]: *** >
> [attica/kdeplugin/CMakeFiles/attica_kde.dir/kdeplatformdependent.o]
> > Error 1 > make[1]: ***
> [attica/kdeplugin/CMakeFiles/attica_kde.dir/all] Error 2 > make:
> *** [all] Error 2 > > I'm running qt 4.6.1. Any ideas? >
> > thanks > > MAC >
>
>
> >run into the same problem. the root cause is, there is no
> provider. >I figured it out, you will need at least Soprano have
> installed.
>
> I have Soprano and the other files. Maybe i need a newer version?
> I'm running 1.2.9 I think
>
> regards
>
> MAC
>
>
>> Hi Mac, seems somewhat complicated, i work on compile kde4.4.2
>> since 2 weeks. I did download and compile half the internet, sorry,
>> i dont rememner
> exactly what i had done to get this particular problem goes away ...
>> But i remember something: I did resolve most of the "optional"
> dependencies.
>> Looks like each of them is optional but at least one is needed. And
>> btw: in most cases its not enough to have these components
> installed, you may need the development version of them often.
>
> Seems the only way to cure this problem was to use
> kdesupport-for-4.4 as opposed to kdesupport TRUNK.  TRUNK was okay
> for kde-4.4.1 but not for 4.4.2, not for me anyway and I don't know
> why.
>
> As an aside.  Polkit-qt-1 is not present in kdesupport-for-4.4 so I
> downloaded and installed the latest version from online.   Doing
> this caused problems with PolKit-KDE-1 which complained about
> missing libraries and version nos.   In the end I had to install
> polkit-qt-1 from kdesupport TRUNK!  This cured the problem.
>
That is odd.  What is your "latest" version, and where did you get it?

I am using: "polkit-kde-1-0.95.1" which I downloaded from a KDE FTP site.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: MP4 Players

[James Richard Tyrer]

On 04/19/10 18:50, Mike McCarty wrote:
> I've read through the BLFS sections related to video, but am still
> uncertain how to proceed toward the goal of playing MP4 (so called)
> format files. It looks like MPLAYER may be the way to go. I have
> an older XINE which I installed some time back, but it doesn't seem
> to be able to handle those.
>
> Anyone have recommendations or suggestions for an MP4 player?

First, MP4 is called a container file format.  In order to be able to 
play an MP4 file, you need to have the codecs for the audio and video 
which it contains.

Personally, I prefer Xine and Kaffine, but the issue will be the same no 
matter what you use.  You need to install not only Xine but a bunch of 
codec libraries as well as the Windows DLL files to be able to play all 
AV formats.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: NetworkManager-0.8; build problem

[James Richard Tyrer]

On 04/19/10 18:04, Mike McCarty wrote:
> James Richard Tyrer wrote:
>> Is there something wrong with this code:
>>
>> NetworkManager-0.8.build/src/NetworkManagerPolicy.c:
>>
>> #include
>>
>>  if (isblank (*p) || (*p == '\0')) {
>>
>> ??  GCC-4.3.4 doesn't like it:
>>
>> NetworkManagerPolicy.c:272: error: implicit declaration of function
>> 'isblank'
>
> Yep. The Standard C provides somehthing which is similar to what is
> meant, but the name is "isspace(.)". "isblank(.)" is a GNU
> extension.
>
And I am compiling against GLibC, so it should be OK.

Could this be the wrong error message?

> Are you compiling with strict ANSI compliance?

The command when it doesn't work is:

/usr/bin/gcc -DHAVE_CONFIG_H -I. -I.. -I.. -I../include -I../marshallers 
-I../src/named-manager -I../src/vpn-manager -I../src/dhcp-manager 
-I../src/ip6-manager -I../src/supplicant-manager 
-I../src/dnsmasq-manager -I../src/modem-manager -I../src/bluez-manager 
-I../src/system-settings -I../libnm-util -I../libnm-glib -I../callouts 
-pthread -I/usr/include/dbus-1.0 -I/usr/lib/dbus-1.0/include 
-I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include   -pthread 
-I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include-pthread 
-I/usr/include/gudev-1.0 -I/usr/include/glib-2.0 
-I/usr/lib/glib-2.0/include -pthread -I/usr/include/glib-2.0 
-I/usr/lib/glib-2.0/include   -DG_DISABLE_DEPRECATED 
-DBINDIR=\"/usr/bin\" -DSBINDIR=\"/usr/sbin\" 
-DLIBEXECDIR=\"/usr/lib/network-manager\" -DDATADIR=\"/usr/share\" 
-DSYSCONFDIR=\"/etc\" -DLOCALSTATEDIR=\"/var\" 
-DNM_RUN_DIR=\"/var/run/NetworkManager\" 
-DNMLOCALEDIR=\"/usr/share/locale\" -DARP_DEBUG -I/usr/gnome-2/include 
-I/usr/include/mysql -I/usr/lib/xulrunner-1.9.2.3   -Wall -Werror 
-std=gnu89 -march=athlon -mtune=athlon -O2 -m128bit-long-double -Wshadow 
-Wmissing-declarations -Wmissing-prototypes 
-Wdeclaration-after-statement -Wfloat-equal -Wno-unused-parameter 
-Wno-sign-compare -fno-strict-aliasing -MT 
NetworkManager-NetworkManagerPolicy.o -MD -MP -MF 
.deps/NetworkManager-NetworkManagerPolicy.Tpo -c -o 
NetworkManager-NetworkManagerPolicy.o `test -f 'NetworkManagerPolicy.c' 
|| echo './'`NetworkManagerPolicy.c

Doesn't look that way,

>
> isspace(.) looks for space, tab, vertical tab, newline, return,
> and formfeed.
> isblank(.) looks only for space and tab.
>
>> Any suggestions?
>>
>> Note that it does compile OK if I add the: "configure" paramater:
>>
>>   --disable-more-warnings
>
> Uh, it doesn't compile any differently, you just don't see the
> warning. That function has no prototype in scope.

The GNU: "ctype.h" header defines it as something else, apparently 
dependent on whether or not various things are defined, so there should 
be the function somewhere, but it is possible that the C headers are 
missing the prototype.  That is a warning, not an error, so it should 
compile correctly.

So, I added: "-Wno-error=implicit-function-declaration" to the above 
command and it compiles correctly.  Using: 
"-Wno-error=missing-prototypes" doesn't fix it, so I doubt that it is a 
missing prototype.

What is implicit about the definition unless there is a problem with the 
header?  OTOH, using: "isspace" instead of: "isblank" in the code fixes 
the problem cures the problem.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Automatically get ip with dhcp when cable is connected

[James Richard Tyrer]

On 04/16/10 11:16, Maginot Junior wrote:
> Hi.
>
>
>
> I'm wondering if anybody knows how to make my lfs system automatically
> get ip when we plug the ethernet cable. If the system boots with the
> cable connected then I get ip with dhcp, but after boot if I plug the
> cable I have to manually run into the terminal to configure the
> network. I notice that distros like Ubuntu detects when the cable is
> plugged and automatically connects, how can I implement that on my
> distro, is that anything to do with avahi?
>
NetworkManager (I have 0.8 installed) will do this.  See my new posting 
in the near future.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

NetworkManager-0.8; build problem

[James Richard Tyrer]

Is there something wrong with this code:

NetworkManager-0.8.build/src/NetworkManagerPolicy.c:

#include 

if (isblank (*p) || (*p == '\0')) {

??  GCC-4.3.4 doesn't like it:

NetworkManagerPolicy.c:272: error: implicit declaration of function 
'isblank'


Any suggestions?

Note that it does compile OK if I add the: "configure" paramater:

 --disable-more-warnings

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: kdebaseworkspace-4.4.2-help required

[James Richard Tyrer]

On 04/12/10 15:03, William Immendorf wrote:
> On Mon, Apr 12, 2010 at 2:37 PM, lux-integ  wrote:
>> CMakeFiles/kcm_keyboard_layout.dir/x11helper.o: In function
>> `X11Helper::registerForNewDeviceEvent(_XDisplay*)':
>> $SOURCES/kdebase-workspace-4.4.2/kcontrol/kxkb/x11helper.cpp:413: undefined
>> reference to `_XiGetDevicePresenceNotifyEvent'
>> $SOURCES/kdebase-workspace-4.4.2/kcontrol/kxkb/x11helper.cpp:414: undefined
>> reference to `XSelectExtensionEvent'
>> CMakeFiles/kcm_keyboard_layout.dir/x11helper.o: In function
>> `X11Helper::isNewDeviceEvent(_XEvent*)':
>> $SOURCES/kdebase-workspace-4.4.2/kcontrol/kxkb/x11helper.cpp:388: undefined
>> reference to `XListInputDevices'
>> $SOURCES/kdebase-workspace-4.4.2/kcontrol/kxkb/x11helper.cpp:399: undefined
>> reference to `XFreeDeviceList'
> Hmm... You might want to try this:
>
> ln -sfv $XORG_PREFIX/lib/libXi.so /usr/lib
>
> CMAKE, unfountrantly, always assumes that X.org is in /usr, and you
> have to satisfy it.
>
Actually, it is the FindX11.cmake (and probably other) file(s) that 
contains the search path:

   SET(X11_INC_SEARCH_PATH
 /usr/pkg/xorg/include
 /usr/X11R6/include
 /usr/X11R7/include
 /usr/include/X11
 /usr/openwin/include
 /usr/openwin/share/include
 /opt/graphics/OpenGL/include

You can modify them, or setting: CMAKE_INCLUDE_PATH & CMAKE_LIBRARY_PATH 
to include your Xorg 'include' and 'lib' directories should also work.

Note that this file should be patched so that everything else comes 
before X11R6 because most systems have a: "/usr/X11R6" link and CMake 
doesn't seem to use the actual directory when it found it with a link as 
Autotools does.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: kdebase-4.4.2 and xkb?

[James Richard Tyrer]

On 04/16/10 09:54, lux-integ wrote:
> Greetings,
>
>
>
> I am   trying to compile kde-4.4.2 for a 64-bit only linux machine (linuxCBLFS
> kernel-2.6.33 gcc-4.4.2 ) where neither kde nor Xorg are in /usr
> However when I reached   kdebase-4.4.2  cmake reported the following:
>
> -
> -- The following external packages were located on your system.
> -- This installation will have the extra features provided by these packages.
> -
> * Soprano - Semantic Desktop Storing
> * Shared desktop ontologies - Desktop ontologies
> * Nepomuk - Nepomuk
> * OpenGL - API for developing portable, interactive 2D and 3D graphics
> applications
> * PCIUTILS - PciUtils is a library for direct access to PCI slots
> * RAW1394 - library for direct access to IEEE 1394 bus
> * glib2 - Low-level core library for data structure handling, portability
> wrappers, etc.
>
> -
> -- The following OPTIONAL packages could NOT be located on your system.
> -- Consider installing them to enable more features from this software.
> -
> * XKB
>   X keyboard extension
>   Gives Konsole better keyboard support.
>
> -
>
> Help would be appreciated to explain if   this  (
> http://www.freedesktop.org/wiki/Software/XKeyboardConfig _ is what is being
> referred to or otherwise?

Have you tried setting:

CMAKE_INCLUDE_PATH=$XORG_PREFIX/include
CMAKE_LIBRARY_PATH=$XORG_PREFIX/lib

??  This works for me.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Fonts

[James Richard Tyrer]

On 04/10/10 05:13, Baho Utot wrote:
> I have completed installing KDE-3.5.10
>
> Anyone have some suggestions on installing or configuring fonts?
>
> Thunderbird and Firefox are terrible

This is possible.  The way this works is that Firefox and Thunderbird 
use the generic names as defaults (Sans, Serif, etc) and FontConfig 
determines which actual font to use based on the contents of the 
configuration files in:

/etc/fonts
/etc/fonts/conf.d

Also, by default it only looks for fonts in the sub-directories of:

/usr/share/fonts
~/.fonts

So, you need to either make links to your Xorg font sub-directories or 
add: "/etc/fonts/local.conf" with any other font locations that you want 
to use.  E.G.

--8<--8<--8<--8<--8<--8<--8<--8<--








/usr/X11R6/lib/X11/fonts



--8<--8<--8<--8<--8<--8<--8<--8<--

Also note that you need to link one of these:

/etc/fonts/conf.avail/70-no-bitmaps.conf
/etc/fonts/conf.avail/70-yes-bitmaps.conf

to:

/etc/fonts/conf.d/

depending on whether or not you want FontConfig to use bitmapped fonts.

That said, whether the default fonts look good depends on what you have 
installed with the worst case being re-sized bitmap fonts.

So, you want to install the Xorg Type1 and TrueType (ttf) packages.  You 
also should consider installing other available fonts:

http://dejavu-fonts.org/wiki/index.php?title=Main_Page
http://ftp.gnome.org/pub/GNOME/sources/ttf-bitstream-vera/1.10/
http://corefonts.sourceforge.net/
http://mirror.cs.wisc.edu/pub/mirrors/ghost/AFPL/GhostPCL/urwfonts-1.41.tar.bz2

And,or configure FontConfig to use the GhostScript fonts

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: help with-kde442.polkit-kde

[James Richard Tyrer]

On 04/09/10 06:37, lux-integ wrote:
> On Friday 09 April 2010 07:53:43 am James Richard Tyrer wrote:
>> Remember that if you do this, you will have to configure D-Bus:
>>
>> http://techbase.kde.org/Getting_Started/Tutorials/D-Bus/Configuration
> I did this
>>> QUESTION 2:
>>> my $KDEDIR is not /usr,  is is sensible to  change
>>> -DKDE4_AUTH_POLICY_FILES_INSTALL_DIR:STRING=/usr/share/PolicyKit/policy \
>>> to
>>> -DKDE4_AUTH_POLICY_FILES_INSTALL_DIR:STRING=$KDEDIR/share/PolicyKit/polic
>>> y \
>>
>> Very bad idea.  There is currently no way to configure this, so it will
>> not work.
>
> I have
> file1:
> $KDEDIR/share/PolicyKit/policy/org.kde.kcontrol.kcmremotewidgets.policy
>
> and the other policy files in /usr/share/PolicyKit/policy.
>
> Is it   therefore wise to  move   file1   to   /usr/share/PolicyKit/policy
> then make a symlink  from  $KDEDIR/share/PolicyKit/policy/ ?
>
> or OTHERWISE

Yeah that. :-|.  I don't know if it is a bug, or what, but it couldn't 
hurt to make a link to "/usr/share/PolicyKit/policy/" which is what I did.
-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: help with-kde442.polkit-kde

[James Richard Tyrer]

On 04/08/10 06:27, lux-integ wrote:

> -DCMAKE_INSTALL_PREFIX=$KDEDIR -DSYSCONF_INSTALL_DIR=/etc/kde

Probably better to use:

-DSYSCONF_INSTALL_DIR=/etc/kde-4.4.2

and make a link:

/etc/kde -> /etc/kde-4.4.2

And something similar for the PREFIX.

Remember that if you do this, you will have to configure D-Bus:

http://techbase.kde.org/Getting_Started/Tutorials/D-Bus/Configuration

> QUESTION 2:
> my $KDEDIR is not /usr,  is is sensible to  change
> -DKDE4_AUTH_POLICY_FILES_INSTALL_DIR:STRING=/usr/share/PolicyKit/policy \
> to
> -DKDE4_AUTH_POLICY_FILES_INSTALL_DIR:STRING=$KDEDIR/share/PolicyKit/policy \

Very bad idea.  There is currently no way to configure this, so it will 
not work.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

System Screwed [Was GLibc-2.11.1]

[James Richard Tyrer]

Well, I think that I have found the problem and it isn't the header files.

Somehow, my system got screwed, but that isn't a great problem since 
some things needed upgrading any way.  But, I can't build things.

Well actually, I built GLibC with no problem but I still get the failure 
when trying to build: libiberty in binutils:

/usr/bin/gcc -c -DHAVE_CONFIG_H -march=athlon -mtune=athlon -O2 
-m128bit-long-double -I/usr/gnome-2/include -I/usr/include/mysql 
-I/usr/lib/xulrunner-1.9.1.7  -I. 
-I../../binutils-2.20.1/libiberty/../include  -W -Wall -Wwrite-strings 
-Wc++-compat -Wstrict-prototypes -pedantic 
../../binutils-2.20.1/libiberty/floatformat.c -o floatformat.o
In file included from ../../binutils-2.20.1/libiberty/floatformat.c:31:
/usr/include/string.h:546: error: expected declaration specifiers or 
'...' before '__locale_t'
/usr/include/string.h:547: error: nonnull argument with out-of-range 
operand number (argument 1, operand 3)
/usr/include/string.h:550: error: expected declaration specifiers or 
'...' before '__locale_t'
/usr/include/string.h:551: error: nonnull argument with out-of-range 
operand number (argument 1, operand 4)

I find that all of the files in:

/usr/share/locale

have gone missing along with other files in: "/usr/share".  It appears 
that, my best guess, is that the: "make uninstall" command in R went rouge.

So, perhaps someone could tell me where the files in: 
"/usr/share/locale" come from and/or if there are other locale type 
dependencies to build: "libiberty".  If I can borrow the needed data 
from a binary, I should be on my way to fixing it.


-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: GLibc-2.11.1

[James Richard Tyrer]

On 04/05/10 12:30, Andrew Benton wrote:

On 05/04/10 03:40, James Richard Tyrer wrote:

I made the error of updating my GLibc to 2.11.1 and my GCC to 4.4.3 and
all HELL (DLL HELL that is) has broken loose. :-(

The major problem appears to be with various "locale" objects in the
headers for GLibc and possibly in the headers for LibStdC++-4.4.3.

Do we, I hope, have a patch in the works for this problem, or it it
possible that down grading to GCC-4.4.1 would improve the problem (I
consider this to be very low probability unless there is a specific
regression).



I would suggest that you start again from the beginning. Even if you do somehow
get your system functional again, there's no way of knowing what else is broken
and you just haven't found that brokenness yet. Everytime something doesn't work
you'll be thinking "Is that because I tried to reinstall glibc?"
Just start again and don't try and reinstall glibc again - it never makes 
anything
better.

While I appreciate your suggestion, I have upgraded GLibC many times 
over the past few years without any problems.


Now I have the specific issue that I can not build things due to the 
issues with the GLibC headers.  Unfortunately, this includes rebuilding 
GCC as well.


For example, building PCRE (now trying 8.0.2) I get specific errors [see 
attached] with a header file.


Don't know how this could be caused by BinUtils, but I will borrow a 
compiler and try reinstalling it.



--
James Tyrer

Linux (mostly) From Scratch
make  all-am
make[1]: Entering directory `/usr/src/pcre-8.02'
/bin/sh ./libtool --tag=CXX   --mode=compile /usr/bin/g++ -DHAVE_CONFIG_H -I.   
-I/usr/gnome-2/include -I/usr/include/mysql -I/usr/lib/xulrunner-1.9.1.7   
-march=athlon -mtune=athlon -O2 -m128bit-long-double -MT pcrecpp.lo -MD -MP -MF 
.deps/pcrecpp.Tpo -c -o pcrecpp.lo pcrecpp.cc
libtool: compile:  /usr/bin/g++ -DHAVE_CONFIG_H -I. -I/usr/gnome-2/include 
-I/usr/include/mysql -I/usr/lib/xulrunner-1.9.1.7 -march=athlon -mtune=athlon 
-O2 -m128bit-long-double -MT pcrecpp.lo -MD -MP -MF .deps/pcrecpp.Tpo -c 
pcrecpp.cc  -fPIC -DPIC -o .libs/pcrecpp.o
In file included from /usr/include/stdio.h:909,
 from pcrecpp.cc:37:
/usr/include/bits/stdio.h: In function '__ssize_t getline(char**, size_t*, 
FILE*)':
/usr/include/bits/stdio.h:118: error: '__getdelim' was not declared in this 
scope
In file included from /usr/include/c++/4.4.3/bits/localefwd.h:42,
 from /usr/include/c++/4.4.3/string:45,
 from pcrecpp.cc:42:
/usr/include/c++/4.4.3/i686-pld-linux/bits/c++locale.h: At global scope:
/usr/include/c++/4.4.3/i686-pld-linux/bits/c++locale.h:52: error: 'uselocale' 
was not declared in this scope
/usr/include/c++/4.4.3/i686-pld-linux/bits/c++locale.h:52: error: invalid type 
in declaration before ';' token
/usr/include/c++/4.4.3/i686-pld-linux/bits/c++locale.h: In function 'int 
std::__convert_from_v(__locale_struct* const&, char*, int, const char*, ...)':
/usr/include/c++/4.4.3/i686-pld-linux/bits/c++locale.h:72: error: 
'__gnu_cxx::__uselocale' cannot be used as a function
/usr/include/c++/4.4.3/i686-pld-linux/bits/c++locale.h:97: error: 
'__gnu_cxx::__uselocale' cannot be used as a function
make[1]: *** [pcrecpp.lo] Error 1
make[1]: Leaving directory `/usr/src/pcre-8.02'
make: *** [all] Error 2
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

GLibc-2.11.1

[James Richard Tyrer]

I made the error of updating my GLibc to 2.11.1 and my GCC to 4.4.3 and 
all HELL (DLL HELL that is) has broken loose. :-(

The major problem appears to be with various "locale" objects in the 
headers for GLibc and possibly in the headers for LibStdC++-4.4.3.

Do we, I hope, have a patch in the works for this problem, or it it 
possible that down grading to GCC-4.4.1 would improve the problem (I 
consider this to be very low probability unless there is a specific 
regression).


-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: D-Bus configuration for GNOME-2 (or KDE-4.4) TYPO ERRORS

[James Richard Tyrer]

On 04/01/10 12:43, James Richard Tyrer wrote:

> Then there is the question of the:
>
>   /opt/gnome/share/share/polkit-1/actions

Should be:

/opt/gnome/share/polkit-1/actions
>
> directory.  On KDE-4.4 this could only be addressed by installing the
> contents in:
>
>   /usr/share/share/polkit-1/actions   

Should be:

/usr/share/polkit-1/actions

Sorry about that.


-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

D-Bus configuration for GNOME-2 (or KDE-4.4)

[James Richard Tyrer]

Having fought with this for KDE-4.4 for a while, I think that I finally 
have if figured out and applying it to GNOME packages solved my GTK app 
issues as well.

I have posted a tutorial for KDE-4.4:

http://techbase.kde.org/Getting_Started/Tutorials/D-Bus/Configuration

which is also applicable (KDE-4.4 instructions too) to GNOME, especially 
if installed in:

/opt/gnome

Specifically, the configuration in:

http://www.linuxfromscratch.org/blfs/view/svn/gnome/GConf.html

is not sufficient.  I would also suggest that it be moved to:

http://www.linuxfromscratch.org/blfs/view/svn/gnome/pre-install-config.html

There are 4 directories:

/etc/gnome/2.28.2/dbus-1/session.d
/etc/gnome/2.28.2/dbus-1/system.d
/opt/gnome/share/dbus-1/services
/opt/gnome/share/dbus-1/system-services

that D-Bus needs to find to work correctly -- although I note that the 
first one is currently empty on my system, it is a drop in directory and 
it must be correctly configured.  These have a direct correspondence to 
the KDE-4.4 directories in the tutorial.

Then there is the question of the:

/opt/gnome/share/share/polkit-1/actions

directory.  On KDE-4.4 this could only be addressed by installing the 
contents in:

/usr/share/share/polkit-1/actions   

and KDE packages have a build option to do this.

I have not looked into that for GNOME yet.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: DHCP client vs. PAM

[James Richard Tyrer]

On 03/27/10 22:53, James Richard Tyrer wrote:

> But, the problem with the DHCP client persisted.  It appeared that the
> problem was that it didn't find: "eth0" despite the fact that it is in
> the configuration file (cated from the above page) -- it worked if I
> ran: "dhclient eth0" but not if I just ran: "dhclient".
>
> So, I did a reinstall of DHCP and now it works.

And now the problem has returned after I rebooted the system.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: DHCP client vs. PAM

[James Richard Tyrer]

On 03/27/10 19:50, Chris Staub wrote:
> On 03/27/2010 04:51 PM, James Richard Tyrer wrote:
>> I upgraded Linux-PAM using the instructions in BLFS SVN and my DHCP
>> client quit working.  If I run (as root):
>>
>> dhclient
>>
>> nothing happens.  If there is something in the BLFS SVN that covers
>> this, I have missed it.
>>
>> Have to admit that what I don't know about PAM would fill a book.
>>
>> Is there a PAM drop in file that I should add to configure this?
>>
>
> Far as I know, DHCP doesn't even use PAM, so there should be no
> relationship between them.

True, it doesn't use PAM, but it needs to find the network device, and
have access to it, to work.

> I have a feeling this could be related to your other thread about
> the DHCP instructions, confusing them with dhcpcd. How exactly is
> DHCP configured?

I'm only using the client and it is configured according to the:

DHCP-3.0.6 Client Configuration

page.

My system has been doing a lot of strange things after I updated
Linux-PAM, but I seem to have resolved most of them and I am not sure if
I made errors or what but there is no point in trying to reproduce them
as long as they work now.  Well, perhaps, not 100% as I am still using a
permissive: "other" file for PAM.

But, the problem with the DHCP client persisted.  It appeared that the
problem was that it didn't find: "eth0" despite the fact that it is in
the configuration file (cated from the above page) -- it worked if I
ran: "dhclient eth0" but not if I just ran: "dhclient".

So, I did a reinstall of DHCP and now it works.  Go figure.  Perhaps I
am just having a bad day or trying to figure out KDE-4.4 has taken its
toll on my concentration.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: BLFS SVN problem: DHCP-3.0.6 Client Configuration

[James Richard Tyrer]

On 03/27/10 18:39, Chris Staub wrote:
> On 03/27/2010 09:30 PM, James Richard Tyrer wrote:
>> This line:
>>
>> cat>   /etc/dhclient.conf<<   "EOF"
>>
>> should be:
>>
>> cat>   /etc/dhcpcd/dhclient.conf<<   "EOF"
>>
>> because in:
>>
>> dhcpcd-4.0.11
>>
>> this parameter is set:
>>
>> SYSCONFDIR=/etc/dhcpcd
>>
>
> I think you're a bit confused. /etc/dhcpcd is used for dhcpcd, not DHCP.
> Those are 2 different packages.

Yes, I read the wrong page.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

BLFS SVN problem: DHCP-3.0.6 Client Configuration

[James Richard Tyrer]

This line:

cat > /etc/dhclient.conf << "EOF"

should be:

cat > /etc/dhcpcd/dhclient.conf << "EOF"

because in:

dhcpcd-4.0.11

this parameter is set:

SYSCONFDIR=/etc/dhcpcd

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

DHCP client vs. PAM

[James Richard Tyrer]

I upgraded Linux-PAM using the instructions in BLFS SVN and my DHCP 
client quit working.  If I run (as root):

dhclient

nothing happens.  If there is something in the BLFS SVN that covers 
this, I have missed it.

Have to admit that what I don't know about PAM would fill a book.

Is there a PAM drop in file that I should add to configure this?

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[kde-linux] Help needed: KAuth and SystemSettings.

[James Richard Tyrer]

I have been told various things about installation of KAuth from various 
developers that may, or may not, know what they were talking about. 
Without any documentation, I have to proceed with trial and error which 
is ridiculous.

Edgar Alwers was told that he should build it against PolKit-Qt-1.  I 
did this and the result was that the KDE policies disappeared from 
"System Settings:: Advanced -> PolicyKit Authorization".  However, I 
used version 0.95.1 and someone said that 0.95.2 was needed (actually 
the TRUNK version).  Does anyone know?

Having now installed KDE-4.4.1+ in "/usr" along with KDESupport-4.4 and 
all of the D-Bus and Policy stuff, I find that it still doesn't work.

If you are running KDE-4.4.x, can you set the Date and Time Zone using 
"System Settings:: General -> Date & Time"?  If so, what did you do? 
Can you do this as user by doing an authentication with the root 
password or otherwise?

So, I was tired of the security that didn't work and went to the 
"/usr/share/PolicyKit/policy" and (as root) executed:

sed -i -e 's/auth_admin/yes/' org.kde.*

which should remove all security.

You might have to edit them by hand so that each instance of 
"" would have:

yes

So, now I should be able to do everything as any user.  Well I can NOT, 
so I need help to determine if this is a bug.  I can set the Date, but 
when I try to set the Time Zone, the "Apply" button is not enabled. This 
looks like a bug since if I change the Date too, I can change the Time Zone.

Using KSysGuard, I can now (as user) change the "nice" of a process, but 
when I try to set the "I/O Scheduler" all I get are error messages. Both 
do work if I open KSysGuard ad root, but that worked OK before. Can you 
use KSysGuard to change the I/O Scheduler as root?

I was told that my problems were caused by my installation and 
configuration -- falsely told that I couldn't build it (apparently the 
developer didn't know the difference between building and 
configuration).  And now it looks like this is just another instance of 
KDE releasing new features that are seriously broken.  Although I have 
to admit that it is difficult to tell if you have something installed 
and configured correctly when it is broken and wouldn't work correctly 
in any case.

-- 
James Tyrer

Linux (mostly) From Scratch


-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Building KDE-4.4.1

[James Richard Tyrer]

Matthew Burgess wrote:
> Hi all,
> 
> I know a couple of folks here have built KDE recently.  I think I 
> already know the answer to this, but I think I have to grab kdesupport 
> out of SVN in order to build KDELibs, right?  My conclusion was arrived 
> at because kdelibs wants 'automoc4' from kdesupport but there is no 
> kdesupport tarball available from KDE's download site.
> 
> If that's the case, how do I stop things like oxygen-icons being built 
> that already have their own tarballs available?  Preferably I don't even 
> want to check out the oxygen-icons area of svn but I'm not sure that's 
> possible...an 'svn update' on such a checkout would pull oxygen-icons 
> next time around anyway, I would have thought.
> 
> Why can't developers release a bloody tarball of their stuff?
> 
Good question to which I have no answer except to say that they seem to 
expect that you will get the dependencies from your distro.

So, you will need to use SVN.  This is really almost as simple as using 
a command line FTP client.  Just execute:

svn co svn://anonsvn.kde.org/home/kde/tags/kdesupport-for-4.4

where you want it downloaded.

If you only want "automoc" you can check out only that piece.

svn co svn://anonsvn.kde.org/home/kde/tags/kdesupport-for-4.4/automoc

There is some question as to whether the TAG has the correct version of: 
"PolKit-Qt-1".  So, you might want to get the newer version from TRUNK:

svn co svn://anonsvn.kde.org/home/kde/trunk/kdesupport/polkit-qt-1

To further confuse things, The TAG version does not have the: 
"ontologies" which are available here:

svn co https://oscaf.svn.sourceforge.net/svnroot/oscaf/trunk/ontologies

And, the TRUNK version does not contain: "Phonon" anymore but you can 
still get it by first checking it out and then:

cd /phonon
svn up -r1092912

There have been problems with KDEBindings not building against newer 
versions of Phonon so that is the latest version which I recommend.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

NetworkManager Applet

[James Richard Tyrer]

After I managed to configure D-Bus (I hope) correctly, I now have the 
NetworkManager Applet in my system tray.  It will start my network 
(which is only a DSL modem) OK, but when it does, it changes the host 
name.  I think that it is using the host name for the modem, but I'm not 
sure.

I am using DHClient and that works fine, no change in host name if I 
start it from the command line.

I would guess that there is some configuration file somewhere.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: DBus configuration issues: system.d

[James Richard Tyrer]

Simon Geard wrote:
> On Thu, 2010-02-04 at 16:27 -0700, James Richard Tyrer wrote:
>> Do you know if it is possible to configure DBus so that this
>> workaround is not needed?
> 
> You'd want to check docs for details, but the default dbus
> system.conf contains a line that includes the contents of system.d.

Unfortunately, the documentation is less than complete.  Will try the
FD.o D-Bus list.

> Potentially you could add additional includedir directives to support
> other directories?

Yes, that is what I need to do, but not sure that it will work.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: DBus configuration issues: system.d

[James Richard Tyrer]

Wayne Blaszczyk wrote:
> On 04/02/10 18:43, James Richard Tyrer wrote:
>> BLFS SVN says to install various GNOME packages with:
>>
>>  --sysconfdir=/etc/gnome/2.28.2
>>
>> Th9is results in files being installed in:
>>
>>  /etc/gnome/2.28.2/dbus-1/system.d/
>>
>> I presume that it is necessary to configure something somewhere so that 
>> those files will be found and used.  It appears that this is configured in:
>>
>>  /etc/dbus-1/system.conf
>>
>> Or, I should add a file:
>>
>>  /etc/dbus-1/system-local.conf
>>
>> I note that I haven't found any instruction in BLFS SVN on how to do 
>> this.  I haven't read the whole book, but I haven't seen anything.  If 
>> it is there, please point me to it.
>>
>> If not, the question is how to do this.  The "system.conf" file has this 
>> line:
>>
>>system.d
>>
>> This is somewhat confusing since it does not use the full path.
>>
>> So, what do I do so that the files in:
>>
>>  /etc/gnome/2.28.2/dbus-1/system.d/
>>
>> will be found and uses?
>>
>> Can I add additional "system.d" directories using a full path?  Is this 
>> necessary?
>>
> 
> See the BLFS SVN GConf-2.28.0 page, there is a section there that
> describes how to set this up using the /etc/dbus-1/system-local.conf file.

I would swear that I read that page.  But, apparently not well enough.

Thanks.  I am making progress -- actually trying to configure KDE-4.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: DBus configuration issues: system.d

[James Richard Tyrer]

Mike McCarty wrote:
> James Richard Tyrer wrote:
>> BLFS SVN says to install various GNOME packages with:
>> 
>> --sysconfdir=/etc/gnome/2.28.2
>> 
>> Th9is results in files being installed in:
>> 
>> /etc/gnome/2.28.2/dbus-1/system.d/
>> 
>> I presume that it is necessary to configure something somewhere so
>> that those files will be found and used.  It appears that this is
>> configured in:
>> 
>> /etc/dbus-1/system.conf
>> 
>> Or, I should add a file:
> 
> I thought the book instructed one to put in a soft link, so the stuff
> could be accessed in "standard" locations.
> 
Thanks Mike.

That is a work around which I am currently using.

Do you know if it is possible to configure DBus so that this workaround 
is not needed?

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

DBus configuration issues: system.d

[James Richard Tyrer]

BLFS SVN says to install various GNOME packages with:

--sysconfdir=/etc/gnome/2.28.2

Th9is results in files being installed in:

/etc/gnome/2.28.2/dbus-1/system.d/

I presume that it is necessary to configure something somewhere so that 
those files will be found and used.  It appears that this is configured in:

/etc/dbus-1/system.conf

Or, I should add a file:

/etc/dbus-1/system-local.conf

I note that I haven't found any instruction in BLFS SVN on how to do 
this.  I haven't read the whole book, but I haven't seen anything.  If 
it is there, please point me to it.

If not, the question is how to do this.  The "system.conf" file has this 
line:

  system.d

This is somewhat confusing since it does not use the full path.

So, what do I do so that the files in:

/etc/gnome/2.28.2/dbus-1/system.d/

will be found and uses?

Can I add additional "system.d" directories using a full path?  Is this 
necessary?

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: kde44x and webkit

[James Richard Tyrer]

lux-integ wrote:
> Greetings,
> 
> I am planning to start a build of kde4.4.x on a pure amd-64-bit clfs 
> linux-box.  I know one of the requirements is webkit.  I have some help from 
> the folowing url:-
> 
> http://trac.webkit.org/wiki/BuildingQtOnLinux
> 
> wowever it is unclear   if   the build script  using
> 
> "  WebKit/WebKitTools/Scripts/build-webkit --qt "
> 
> just  compiles webkit  or compiles and installs it/
> 
> 
> Q1:  if it just compiles it does anyone   where it put   the binaries, 
> libraries  so that one can do a manual install? 
> 
> Q2:  If it does a compile and install,does anyone know where it installs 
> the binaries etc?  
> 
> Q3:  If the latter is the script best run as root?
> 
> 
> advice will be appreciated

For Qt, use the webkit that comes with Qt-4.x.y.

For GNOME, use the BLFS snapshot.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Problem booting into KDE-4.3

[James Richard Tyrer]

Cliff McDiarmid wrote:

> No probs, its running.
> 
?? strange.

> Anyway, do people think dbus is the problem?
> 
I still think that that error message is indicating the problem.

When the user space message bus starts up, it should set
DBUS_SESSION_BUS_ADDRESS.  To see this, you are going to need to echo it
to a file since you don't have a console running.

I have this:

## test for an existing bus daemon, just to be safe
  if test -z "$DBUS_SESSION_BUS_ADDRESS" ; then
## if not found, launch a new one
eval `dbus-launch --sh-syntax --exit-with-session`
echo "D-BUS per-session daemon address is: $DBUS_SESSION_BUS_ADDRESS"
  fi

which I copied from the documentation.  But, it doesn't help if you
don't have a console for it to be echoed to.  So, direct the echo to a
file.  Remember to make sure that the file can be written to.  What I do
is create a file in /root/ and change the permissions to 666.

-- 
JRT



-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Problem booting into KDE-4.3

[James Richard Tyrer]

Cliff McDiarmid wrote:
>> - Original Message -
>> From: "James Richard Tyrer" 
>> To: "BLFS Support List" 
>> Subject: Re: Problem booting into KDE-4.3
>> Date: Wed, 16 Sep 2009 04:24:10 -0700
> 
>> That is not the command which I am using.  You might try changing it to:
>>
>> eval `dbus-launch --exit-with-session`
>> exec startkde
>>
>> --
>> James Tyrer
> 
> Thanks for the advice James, but the problem still persists.  Haven't had any 
> time to mess with it in the last week(I use 2nd LFS on another partition), 
> but I think It's a dbus problem, although would welcome advice to the 
> contrary.  I can get an xterm with clock etc, but when I startx (have 
> .Xinitrc in users account with the above in it) I'm thrown back to command 
> prompt.  End of log reads:
> 
> startkde: Starting up...
> Failed to start message bus: Element  is not allowed in this 
> contextEOF in dbus-launch reading address from bus daemon
> startkde: Could not start D-Bus. Can you call qdbus?
> 
That last line probably means that the D-Bus (a.k.a. message bus) daemon 
isn't running.

Try (as root):

/etc/rc.d/init.d/messagebus status

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Problem booting into KDE-4.3

[James Richard Tyrer]

Cliff McDiarmid wrote:
>> - Original Message -
>> From: "William Immendorf" 
> 
>>> /usr/bin/startx: line 159: xauth: command not found
>> I think you are missing xauth. Try running 'which xauth' (if you have
>> which). If it dosen't give anything, you are missing xauth or it's not
>> in your path.
>>
>> William
>  
> I had thought that too, but KDE has booted before without it - anyway I have 
> now installed it, but there's no change.
> 
> log file now reads: 
> 
> xauth:  creating new authority file /home/cliffys/.serverauth.2513
> 
> X Window System Version 7.2.0
> Release Date: 22 January 2007
> X Protocol Version 11, Revision 0, Release 7.2
> Build Operating System: Linux 2.6.24.4 i686 
> Current Operating System: Linux localhost.localdomain 2.6.27.10 #1 PREEMPT 
> Sun Jan 25 15:37:02 GMT 2009 i686
> Build Date: 03 October 2008
>   Before reporting problems, check http://wiki.x.org
>   to make sure that you have the latest version.
> Module Loader present
> Markers: (--) probed, (**) from config file, (==) default setting,
>   (++) from command line, (!!) notice, (II) informational,
>   (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
> (==) Log file: "/var/log/Xorg.0.log", Time: Sun Aug 30 22:01:46 2009
> (==) Using config file: "/etc/X11/xorg.conf"
> 
> expected keysym, got XF86KbdLightOnOff: line 70 of pc
> expected keysym, got XF86KbdBrightnessDown: line 71 of pc
> expected keysym, got XF86KbdBrightnessUp: line 72 of pc
> exec dbus-launch --exit-with-session startkde
> 
> waiting for X server to shut down FreeFontPath: FPE "built-ins" refcount is 
> 2, should be 1; fixing.
> 

That is not the command which I am using.  You might try changing it to:

eval `dbus-launch --exit-with-session`
exec startkde

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: starting kde-3 and kde-4 ( two users )

[James Richard Tyrer]

Dr. Edgar Alwers wrote:
> In order to give kde-4.3.0 a try, I have build kde-4 with an user "kde4", as 
> adviced ( Getting started/Build/KDE4.x ). 
> However: choosing the user "kde4" in kdm ( kdm continues to be the kde-3.5.10 
> one ), results in a kde-3.5 start, followed by a kde-4 one. I get a kde4 
> desktop mounted on a kde3 one. I do not think  this is a normal behaviour, 
> but I could not solve the issue up to now, inspite of intensive reading of 
> the kdm manual.
> It should be possible to start e.g. gnome, or kde4 without underlying a kde-3 
> session, isn'it ? Can somebody point me out of this problem ? 

I don't normally use KDM to start my KDE-4 session so I didn't include 
that.  Sorry about that.

The problem is that you have two "startkde" scripts on your system and 
how this should work is that the PATH should determine which one to use.

However, the KDE "session" desktop file: "kde.desktop" (which is in the 
directory: $KDEDIR/share/apps/kdm/sessions" {unless the default was 
changed}) might have the full path for the "startkde" command.  For this 
to work, you need to remove the full path so that these two lines are thus:

Exec=startkde
TryExec=startkde

and then it should work.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: Using phonon from kde-qt

[James Richard Tyrer]

Cliff McDiarmid wrote:
> Hello
> 
> I'm trying to build the latest kdebase-runtime 4.3  with phonon from
> qt-4.5 - at least thats what kde wants me to do
> 
No, KDE does not say that.  You should build Qt "-no-phonon" if you are 
using Qt only for KDE.

> cmake finds headers/libraries but then build stops with:
> 
> phonon/kded-module/phononserver.h:26:36: error:
> Phonon/ObjectDescription: No such file or directory
> 
> It seems to want to use phonon from qt in /opt rather than the other
> version I have installed in /usr. I understand this is right, is it?.
> 
> 
>> From the cmake configure output it is finding the version under qt,
>> but still fails to find the above file(it exists) when compiling.
> 
> Can anyone suggest a reason for this on LFS?  Has anyone had similar
> problems?
> 
It is possible that this will work if Qt was built with Phonon, but 
since Qt Phonon is not what KDE requires, you MUST install the KDE 
Phonon package:

ftp://ftp.kde.org/pub/kde/stable/phonon/4.3.0/phonon-4.3.0.tar.bz2

You should install this with the same: "-DCMAKE_INSTALL_PREFIX=" 
as you will use for KDE.

Since there are other packages required, I would just install KDESupport 
which is available only from SVN:

http://techbase.kde.org/Getting_Started/Build/KDE4/LFS#Library_Requirements

svn co svn://anonsvn.kde.org/home/kde/tags/kdesupport-for-4.3/kdesupport

You need the other stuff which is listed there as well!

You also need: "shared-mime-info" and "Boost".  Note the instructions 
for getting KDE packages to build against Boost built from source.

-- 
James Tyrer

Linux (mostly) From Scratch
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page