ICA support status

El Viernes, 31 de Marzo de 2006 19:28, M.Canales.es
escribió:

> Thinking more on that, seem to me that may not be
needed to call make from
> inside the Makefile. If RUN_ICA=1, we could to create
an additional
> "rebuild_system" target including all
chapter6 packages targets as their
> dependencies (plus the *-strippingagain one, if
STRIP=1).
>
> Then, just after finished the first full chapter06
build, add:
>
> iteration1:  chapter6
>   tar -cjf /tools-$BOOK_VERSION.tar.bz2 /tools
>   rm -rf /tools
>   prepare_ica
>   rm -f {list of touched files that need be redone for
rebuild_system
> target} touch iteration1

Well, that don't work :-(

The touched files are useful only when re-run make after a
crash,  but not 
honoured into the current make session. Make refuses to
re-do targets already 
finished.

We will need a separate set of package targets for each
iteration. The issue 
is how to create they in a no very code-intrusive way :-/


-- 
Manuel Canales Esparcia
Usuario de LFS nº2886:       http://www.linuxfroms
cratch.org
LFS en castellano: http://www.escomp
oslinux.org/lfs-es http://www.lfs-es.com
TLDP-ES:                           http://es.tldp.org
-- 
http://linuxfromscratch.org/mailman/listinfo/alfs-discu
ss
FAQ: http://www.linux
fromscratch.org/faq/
Unsubscribe: See the above information page

El Martes, 11 de Abril de 2006 19:00, M.Canales.es escribió:

> We will need a separate set of package targets for each
iteration. The
> issue is how to create they in a no very code-intrusive
way :-/

Well, the new func_compare.sh is almost ready. I will now to
see what changes 
are required into the  chapter6_Makefiles function to
support that new 
method.

There is another issue that need be addressed. Now there is
a new 
configuration option, called COMPARE, that controls if the
ICA/farce 
framework must be used or not.

I would that func_validate_configs.sh will do the following:

- If COMPARE=0, then ITERATIONS, RUN_ICA and RUN_FARCE are
ignored. The 
"settings confirmation" output should to show

COMPARE:  0

but not the others related settings.

- If COMPARE=1, then first must to verify that at least one
of RUN_ICA or 
RUN_FARCE is set to 1, and that ITERATIONS >= 2 (the
current limit to 5 could 
be removed).  The "settings confirmation" output
should to show

RUN_ICA:  {0,1}
RUN_FARCE:  {0,1}
ITERATION:  {N >=2}

but not COMPARE.

George, could you to handle that?


-- 
Manuel Canales Esparcia
Usuario de LFS nº2886:       http://www.linuxfroms
cratch.org
LFS en castellano: http://www.escomp
oslinux.org/lfs-es http://www.lfs-es.com
TLDP-ES:                           http://es.tldp.org
-- 
http://linuxfromscratch.org/mailman/listinfo/alfs-discu
ss
FAQ: http://www.linux
fromscratch.org/faq/
Unsubscribe: See the above information page

M.Canales.es wrote:
> El Martes, 11 de Abril de 2006 19:00, M.Canales.es
escribió:


> 
> I would that func_validate_configs.sh will do the
following:
> 
> - If COMPARE=0, then ITERATIONS, RUN_ICA and RUN_FARCE
are ignored. The 
> "settings confirmation" output should to
show
> 
> COMPARE:  0
> 
> but not the others related settings.
> 
> - If COMPARE=1, then first must to verify that at least
one of RUN_ICA or 
> RUN_FARCE is set to 1, and that ITERATIONS >= 2 (the
current limit to 5 could 
> be removed).  The "settings confirmation"
output should to show
> 
> RUN_ICA:  {0,1}
> RUN_FARCE:  {0,1}
> ITERATION:  {N >=2}
> 
> but not COMPARE.
> 
> George, could you to handle that?

   give me a few hours to ponder on a clean solution.

> 
-- 
http://linuxfromscratch.org/mailman/listinfo/alfs-discu
ss
FAQ: http://www.linux
fromscratch.org/faq/
Unsubscribe: See the above information page

El Miércoles, 12 de Abril de 2006 20:19, George Boudreau
escribió:

>
>    give me a few hours to ponder on a clean solution.

Take your time, I'm yet trying to figure out how to
generate the proper sets 
of targets without very intrusive changes in the
{C,H}LFS/master.sh files.

Then, several builds will be needed to verify the code.
Having that validate 
stuff ready before do the merge to trunk will be fine.

Plus, the new functions and the scripts under extras/ may
need a review for 
error handling. 

-- 
Manuel Canales Esparcia
Usuario de LFS nº2886:       http://www.linuxfroms
cratch.org
LFS en castellano: http://www.escomp
oslinux.org/lfs-es http://www.lfs-es.com
TLDP-ES:                           http://es.tldp.org
-- 
http://linuxfromscratch.org/mailman/listinfo/alfs-discu
ss
FAQ: http://www.linux
fromscratch.org/faq/
Unsubscribe: See the above information page

Manuel,
   you wrote earlier,

" If COMPARE=1, then first must to verify that at
least one of RUN_ICA 
or RUN_FARCE is set to 1, and that ITERATIONS >= 2 (the
current limit to 
5 could be removed).  The "settings
confirmation" output should to show

RUN_ICA:  {0,1}
RUN_FARCE:  {0,1}
ITERATION:  {N >=2}
"

    Are you going to allow settings where both RUN_ICA and
RUN_FARCE are 
enabled or is the analysis limited to either ICA or FARCE.

  George
-- 
http://linuxfromscratch.org/mailman/listinfo/alfs-discu
ss
FAQ: http://www.linux
fromscratch.org/faq/
Unsubscribe: See the above information page

El Miércoles, 12 de Abril de 2006 21:17, George Boudreau
escribió:

>     Are you going to allow settings where both RUN_ICA
and RUN_FARCE are
> enabled or is the analysis limited to either ICA or
FARCE.

Both must can be enabled at the same time. Thy are different
test-tools and 
will use separate copies of the build systems.

Of course, the user will need a lot of free space if both
are enabled and 
ITERATIONS is set to a big number 

-- 
Manuel Canales Esparcia
Usuario de LFS nº2886:       http://www.linuxfroms
cratch.org
LFS en castellano: http://www.escomp
oslinux.org/lfs-es http://www.lfs-es.com
TLDP-ES:                           http://es.tldp.org
-- 
http://linuxfromscratch.org/mailman/listinfo/alfs-discu
ss
FAQ: http://www.linux
fromscratch.org/faq/
Unsubscribe: See the above information page

On 4/12/06, M.Canales.es <manuellinuxfromscratch.org>
wrote:
>
> Of course, the user will need a lot of free space if
both are enabled and
> ITERATIONS is set to a big number 

Yes, this must go in the README or equivalent.   Just in
case it
isn't clear, $ITERATIONS systems for ICA and farce, plus
the one still
in /.  For a 3 iteration build with FARCE and ICA, that's 7
LFS
systems!  I filled an 8GB partition easily the other day
with this
situation when I was keeping all the toolchain build/source
directories, too.

--
Dan
-- 
http://linuxfromscratch.org/mailman/listinfo/alfs-discu
ss
FAQ: http://www.linux
fromscratch.org/faq/
Unsubscribe: See the above information page

M.Canales.es wrote:
> El Miércoles, 12 de Abril de 2006 21:17, George
Boudreau escribió:
> 
>>     Are you going to allow settings where both
RUN_ICA and RUN_FARCE are
>> enabled or is the analysis limited to either ICA or
FARCE.
> 
> Both must can be enabled at the same time. Thy are
different test-tools and 
> will use separate copies of the build systems.

   Display tests inserted into func_validate_configs.sh
   Not the best coding in the world and I 'may' rewrite
the validate 
function to allow for greater flexibility. (but not tonight)

> Of course, the user will need a lot of free space if
both are enabled and 
> ITERATIONS is set to a big number 
> 
-- 
http://linuxfromscratch.org/mailman/listinfo/alfs-discu
ss
FAQ: http://www.linux
fromscratch.org/faq/
Unsubscribe: See the above information page

El Miércoles, 12 de Abril de 2006 22:36, George Boudreau
escribió:

>    Display tests inserted into func_validate_configs.sh
>    Not the best coding in the world and I 'may'
rewrite the validate
> function to allow for greater flexibility. (but not
tonight)

Many thanks 

Also, the Makefile generation look sane now. I will start a
new build in few 
minutes to see what happens 


-- 
Manuel Canales Esparcia
Usuario de LFS nº2886:       http://www.linuxfroms
cratch.org
LFS en castellano: http://www.escomp
oslinux.org/lfs-es http://www.lfs-es.com
TLDP-ES:                           http://es.tldp.org
-- 
http://linuxfromscratch.org/mailman/listinfo/alfs-discu
ss
FAQ: http://www.linux
fromscratch.org/faq/
Unsubscribe: See the above information page

El Miércoles, 12 de Abril de 2006 22:51, M.Canales.es
escribió:

> Also, the Makefile generation look sane now. I will
start a new build in
> few minutes to see what happens 

The build bomb at the end of 069-gcc-build_2 with:

ln: creating symbolic link `/lib/cpp' to `../usr/bin/cpp':
File exists

We will need also a separate set of chapter6 scripts with -f
option added to 
ln, cp, and similar commands :-/

Trying to solve that now ...

-- 
Manuel Canales Esparcia
Usuario de LFS nº2886:       http://www.linuxfroms
cratch.org
LFS en castellano: http://www.escomp
oslinux.org/lfs-es http://www.lfs-es.com
TLDP-ES:                           http://es.tldp.org
-- 
http://linuxfromscratch.org/mailman/listinfo/alfs-discu
ss
FAQ: http://www.linux
fromscratch.org/faq/
Unsubscribe: See the above information page