auto-update problem

[Olivier Mattelaer]

Dear mg/me team,

I stupidly okayed the installation of the updated version of MG/ME over a working version, but with the wrong configuration, and my madevent is now broken! (how, you ask? I logged in to a random machine on lxplus, rather than directly to my desktop, and updated in my home directory, and I guess it installed something compatible with the config of the local machine over a version that was compatible with the architecture of my physical desktop. Or at least I'm guessing this is what the problem was). Is there a way to redo the update using the machine with the original config? Or roll back the version without screwing anything else up?

Desperately seeking a solution :(

cheers,

Rakhi

[Olivier Mattelaer (olivier-mattelaer)]

Hi rakhi,

I convert your question to a bug, and now you can attach a file directly from the launchpad website

Cheers,

Olivier

[Olivier Mattelaer (olivier-mattelaer)]

Hi Rakhi,

Looks like that the modification of one file didn't go through...

Do you have the lines:
subroutine boostm(p,q,m, pboost)

In the following files:
HELAS/boostx.F
and
aloha/template_files/aloha_functions.f

Did you modify one of this files by hand at a point?

Cheers,

Olivier

[Rakhi Mahbubani (r-mahbubani)]

Hi Olivier,

No I don't recall making any changes to those files.

Cheers,

Rakhi

=========================
Chemin de l'Esplanade 26
1214 Vernier
Suisse

Tel: [phone]
Mob: [phone]
==========================
==========================

"One does not discover new lands without consenting to lose
sight of the shore for a very long time." - Andre Gide

________________________________
 From: Olivier Mattelaer <email address hidden>
To: <email address hidden>
Sent: Monday, January 6, 2014 10:24 AM
Subject: [Bug 1261931] Re: auto-update problem

Hi Rakhi,

Looks like that the modification of one file didn't go through...

Do you have the lines:
subroutine boostm(p,q,m, pboost)

In the following files:
HELAS/boostx.F
and
aloha/template_files/aloha_functions.f

Did you modify one of this files by hand at a point?

Cheers,

Olivier

--
You received this bug notification because you are subscribed to the bug
report.
https://bugs.launchpad.net/bugs/1261931

Title:
  auto-update problem

Status in MadGraph5_aMC@NLO Generator:
  New

Bug description:
  Dear mg/me team,

  I stupidly okayed the installation of the updated version of MG/ME
  over a working version, but with the wrong configuration, and my
  madevent is now broken!  (how, you ask?  I logged in to a random
  machine on lxplus, rather than directly to my desktop, and updated in
  my home directory, and I guess it installed something compatible with
  the config of the local machine over a version that was compatible
  with the architecture of my physical desktop.  Or at least I'm
  guessing this is what the problem was).  Is there a way to redo the
  update using the machine with the original config?  Or roll back the
  version without screwing anything else up?

  Desperately seeking a solution :(

  cheers,

  Rakhi

To manage notifications about this bug go to:
https://bugs.launchpad.net/madgraph5/+bug/1261931/+subscriptions

[Olivier Mattelaer (olivier-mattelaer)]

So do you have now the boostm line/function in those files?

Cheers,

Olivier

On Jan 6, 2014, at 1:48 PM, Rakhi Mahbubani <email address hidden> wrote:

> Hi Olivier,
>
> No I don't recall making any changes to those files.
>
> Cheers,
>
> Rakhi
>
>
>
> =========================
> Chemin de l'Esplanade 26
> 1214 Vernier
> Suisse
>
> Tel: [phone]
> Mob: [phone]
> ==========================
> ==========================
>
> "One does not discover new lands without consenting to lose
> sight of the shore for a very long time." - Andre Gide
>
>
> ________________________________
> From: Olivier Mattelaer <email address hidden>
> To: <email address hidden>
> Sent: Monday, January 6, 2014 10:24 AM
> Subject: [Bug 1261931] Re: auto-update problem
>
>
> Hi Rakhi,
>
> Looks like that the modification of one file didn't go through...
>
>
> Do you have the lines:
> subroutine boostm(p,q,m, pboost)
>
> In the following files:
> HELAS/boostx.F
> and
> aloha/template_files/aloha_functions.f
>
> Did you modify one of this files by hand at a point?
>
> Cheers,
>
> Olivier
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1261931
>
> Title:
> auto-update problem
>
> Status in MadGraph5_aMC@NLO Generator:
> New
>
> Bug description:
> Dear mg/me team,
>
> I stupidly okayed the installation of the updated version of MG/ME
> over a working version, but with the wrong configuration, and my
> madevent is now broken! (how, you ask? I logged in to a random
> machine on lxplus, rather than directly to my desktop, and updated in
> my home directory, and I guess it installed something compatible with
> the config of the local machine over a version that was compatible
> with the architecture of my physical desktop. Or at least I'm
> guessing this is what the problem was). Is there a way to redo the
> update using the machine with the original config? Or roll back the
> version without screwing anything else up?
>
> Desperately seeking a solution :(
>
> cheers,
>
> Rakhi
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/madgraph5/+bug/1261931/+subscriptions
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1261931
>
> Title:
> auto-update problem
>
> Status in MadGraph5_aMC@NLO Generator:
> New
>
> Bug description:
> Dear mg/me team,
>
> I stupidly okayed the installation of the updated version of MG/ME
> over a working version, but with the wrong configuration, and my
> madevent is now broken! (how, you ask? I logged in to a random
> machine on lxplus, rather than directly to my desktop, and updated in
> my home directory, and I guess it installed something compatible with
> the config of the local machine over a version that was compatible
> with the architecture of my physical desktop. Or at least I'm
> guessing this is what the problem was). Is there a way to redo the
> update using the machine with the original config? Or roll back the
> version withou...

[Rakhi Mahbubani (r-mahbubani)]

Yes, the function is in both files.

R

=========================
Chemin de l'Esplanade 26
1214 Vernier
Suisse

Tel: [phone]
Mob: [phone]
==========================
==========================

"One does not discover new lands without consenting to lose
sight of the shore for a very long time." - Andre Gide

________________________________
 From: Olivier Mattelaer <email address hidden>
To: <email address hidden>
Sent: Monday, January 6, 2014 2:08 PM
Subject: Re: [Bug 1261931] Re: auto-update problem

So do you have now the boostm line/function in those files?

Cheers,

Olivier

On Jan 6, 2014, at 1:48 PM, Rakhi Mahbubani <email address hidden>
wrote:

> Hi Olivier,
>
> No I don't recall making any changes to those files.
>
> Cheers,
>
> Rakhi
>
>
>
> =========================
> Chemin de l'Esplanade 26
> 1214 Vernier
> Suisse
>
> Tel: [phone]
> Mob: [phone]
> ==========================
> ==========================
>
> "One does not discover new lands without consenting to lose
> sight of the shore for a very long time."    - Andre Gide
>
>
> ________________________________
> From: Olivier Mattelaer <email address hidden>
> To: <email address hidden>
> Sent: Monday, January 6, 2014 10:24 AM
> Subject: [Bug 1261931] Re: auto-update problem
>
>
> Hi Rakhi,
>
> Looks like that the modification of one file didn't go through...
>
>
> Do you have the lines:
> subroutine boostm(p,q,m, pboost)
>
> In the following files:
> HELAS/boostx.F
> and
> aloha/template_files/aloha_functions.f
>
> Did you modify one of this files by hand at a point?
>
> Cheers,
>
> Olivier
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1261931
>
> Title:
>  auto-update problem
>
> Status in MadGraph5_aMC@NLO Generator:
>  New
>
> Bug description:
>  Dear mg/me team,
>
>  I stupidly okayed the installation of the updated version of MG/ME
>  over a working version, but with the wrong configuration, and my
>  madevent is now broken!  (how, you ask?  I logged in to a random
>  machine on lxplus, rather than directly to my desktop, and updated in
>  my home directory, and I guess it installed something compatible with
>  the config of the local machine over a version that was compatible
>  with the architecture of my physical desktop.  Or at least I'm
>  guessing this is what the problem was).  Is there a way to redo the
>  update using the machine with the original config?  Or roll back the
>  version without screwing anything else up?
>
>  Desperately seeking a solution :(
>
>  cheers,
>
>  Rakhi
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/madgraph5/+bug/1261931/+subscriptions
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1261931
>
> Title:
>  auto-update problem
>
> Status in MadGraph5_aMC@NLO Generator:
>  New
>
> Bug description:
>  Dear mg/me team,
>
>  I stupidly okayed the installation of the updated version of MG/ME
>  over a working ve...

[Olivier Mattelaer (olivier-mattelaer)]

Then the update works correctly…

So the problem is probably due to a wrong use of the Template.
I guess that you do manual cp of the Template directory? (the v4 way to generate directory, which is not supported anymore since v2.0)
and then use the ./bin/newprocess_mg5 script.
If so, I guess that the problem is that your copy of the Template was done before the update. Is this possible?

I would advice to do the following:

cp PROCESSNAME/Cards/proc_card_mg5.dat .
emacs proc_card_mg5.dat
   #-> replace the line output by output PROCESSNAME
./bin/mg5 proc_card_mg5.dat

Cheers,

Olivier

On Jan 6, 2014, at 3:15 PM, Rakhi Mahbubani <email address hidden> wrote:

> Yes, the function is in both files.
>
> R
>
>
>
> =========================
> Chemin de l'Esplanade 26
> 1214 Vernier
> Suisse
>
> Tel: [phone]
> Mob: [phone]
> ==========================
> ==========================
>
> "One does not discover new lands without consenting to lose
> sight of the shore for a very long time." - Andre Gide
>
>
> ________________________________
> From: Olivier Mattelaer <email address hidden>
> To: <email address hidden>
> Sent: Monday, January 6, 2014 2:08 PM
> Subject: Re: [Bug 1261931] Re: auto-update problem
>
>
> So do you have now the boostm line/function in those files?
>
> Cheers,
>
> Olivier
>
> On Jan 6, 2014, at 1:48 PM, Rakhi Mahbubani <email address hidden>
> wrote:
>
>> Hi Olivier,
>>
>> No I don't recall making any changes to those files.
>>
>> Cheers,
>>
>> Rakhi
>>
>>
>>
>> =========================
>> Chemin de l'Esplanade 26
>> 1214 Vernier
>> Suisse
>>
>> Tel: [phone]
>> Mob: [phone]
>> ==========================
>> ==========================
>>
>> "One does not discover new lands without consenting to lose
>> sight of the shore for a very long time." - Andre Gide
>>
>>
>> ________________________________
>> From: Olivier Mattelaer <email address hidden>
>> To: <email address hidden>
>> Sent: Monday, January 6, 2014 10:24 AM
>> Subject: [Bug 1261931] Re: auto-update problem
>>
>>
>> Hi Rakhi,
>>
>> Looks like that the modification of one file didn't go through...
>>
>>
>> Do you have the lines:
>> subroutine boostm(p,q,m, pboost)
>>
>> In the following files:
>> HELAS/boostx.F
>> and
>> aloha/template_files/aloha_functions.f
>>
>> Did you modify one of this files by hand at a point?
>>
>> Cheers,
>>
>> Olivier
>>
>> --
>> You received this bug notification because you are subscribed to the bug
>> report.
>> https://bugs.launchpad.net/bugs/1261931
>>
>> Title:
>> auto-update problem
>>
>> Status in MadGraph5_aMC@NLO Generator:
>> New
>>
>> Bug description:
>> Dear mg/me team,
>>
>> I stupidly okayed the installation of the updated version of MG/ME
>> over a working version, but with the wrong configuration, and my
>> madevent is now broken! (how, you ask? I logged in to a random
>> machine on lxplus, rather than directly to my desktop, and updated in
>> my home directory, and I guess it installed something compatible with
>> the config of the local machine over a version th...

[Rakhi Mahbubani (r-mahbubani)]

Dear Olivier,

No, I don't do a manual cp of the Template directory, but generate the process (using an existing model directory) using the python interface of madgraph.  I think I have also had this problem before when trying to use a model folder created with an older version of madgraph, to generate a process in a newer version...is there some issue with backwards-compatibility of the v4 model folders?

Cheers,

Rakhi

=========================
Chemin de l'Esplanade 26
1214 Vernier
Suisse

Tel: [phone]
Mob: [phone]
==========================
==========================

"One does not discover new lands without consenting to lose
sight of the shore for a very long time." - Andre Gide

________________________________
 From: Olivier Mattelaer <email address hidden>
To: <email address hidden>
Sent: Monday, January 6, 2014 3:42 PM
Subject: Re: [Bug 1261931] auto-update problem

Then the update works correctly…

So the problem is probably due to a wrong use of the Template.
I guess that you do manual cp of the Template directory? (the v4 way to generate directory, which is not supported anymore since v2.0)
and then use the ./bin/newprocess_mg5 script.
If so, I guess that the problem is that your copy of the Template was done before the update. Is this possible?

I would advice to do the following:

cp PROCESSNAME/Cards/proc_card_mg5.dat .
emacs proc_card_mg5.dat
   #-> replace the line output by output  PROCESSNAME
./bin/mg5 proc_card_mg5.dat

Cheers,

Olivier

On Jan 6, 2014, at 3:15 PM, Rakhi Mahbubani <email address hidden>
wrote:

> Yes, the function is in both files.
>
> R
>
>
>
> =========================
> Chemin de l'Esplanade 26
> 1214 Vernier
> Suisse
>
> Tel: [phone]
> Mob: [phone]
> ==========================
> ==========================
>
> "One does not discover new lands without consenting to lose
> sight of the shore for a very long time."    - Andre Gide
>
>
> ________________________________
> From: Olivier Mattelaer <email address hidden>
> To: <email address hidden>
> Sent: Monday, January 6, 2014 2:08 PM
> Subject: Re: [Bug 1261931] Re: auto-update problem
>
>
> So do you have now the boostm line/function in those files?
>
> Cheers,
>
> Olivier
>
> On Jan 6, 2014, at 1:48 PM, Rakhi Mahbubani <email address hidden>
> wrote:
>
>> Hi Olivier,
>>
>> No I don't recall making any changes to those files.
>>
>> Cheers,
>>
>> Rakhi
>>
>>
>>
>> =========================
>> Chemin de l'Esplanade 26
>> 1214 Vernier
>> Suisse
>>
>> Tel: [phone]
>> Mob: [phone]
>> ==========================
>> ==========================
>>
>> "One does not discover new lands without consenting to lose
>> sight of the shore for a very long time."    - Andre Gide
>>
>>
>> ________________________________
>> From: Olivier Mattelaer <email address hidden>
>> To: <email address hidden>
>> Sent: Monday, January 6, 2014 10:24 AM
>> Subject: [Bug 1261931] Re: auto-update problem
>>
>>
>> Hi Rakhi,
>>
>> Looks like that the modification of one file didn't go through...
>>
>>
>> Do you have the line...

[Olivier Mattelaer (olivier-mattelaer)]

Hi Rakhi,

I'm confuse about this statement:

> I think I have also had this problem before when trying to
> use a model folder created with an older version of madgraph, to
> generate a process in a newer version…

MadGraph didn't create model… So what do you mean by that?

> is there some issue with
> backwards-compatibility of the v4 model folders?

Not that I know.

The bug that you mention, is doesn't seems to be related to that.
In this case, we have add one function the boostm and looks like this one is not found.
For v4 model this function is in the HELAS/boostx.F file which is automatically copy to
Source/DHELAS

I might have one "idea", do you have some .o file in HELAS directory?
If yes, please remove those, this might be the reason of the crash.

Cheers,

Olivier

On Jan 8, 2014, at 10:57 AM, Rakhi Mahbubani <email address hidden> wrote:

> Dear Olivier,
>
> No, I don't do a manual cp of the Template directory, but generate the
> process (using an existing model directory) using the python interface
> of madgraph. I think I have also had this problem before when trying to
> use a model folder created with an older version of madgraph, to
> generate a process in a newer version...is there some issue with
> backwards-compatibility of the v4 model folders?
>
> Cheers,
>
> Rakhi
>
>
>
> =========================
> Chemin de l'Esplanade 26
> 1214 Vernier
> Suisse
>
> Tel: [phone]
> Mob: [phone]
> ==========================
> ==========================
>
> "One does not discover new lands without consenting to lose
> sight of the shore for a very long time." - Andre Gide
>
>
> ________________________________
> From: Olivier Mattelaer <email address hidden>
> To: <email address hidden>
> Sent: Monday, January 6, 2014 3:42 PM
> Subject: Re: [Bug 1261931] auto-update problem
>
>
> Then the update works correctly…
>
> So the problem is probably due to a wrong use of the Template.
> I guess that you do manual cp of the Template directory? (the v4 way to generate directory, which is not supported anymore since v2.0)
> and then use the ./bin/newprocess_mg5 script.
> If so, I guess that the problem is that your copy of the Template was done before the update. Is this possible?
>
> I would advice to do the following:
>
> cp PROCESSNAME/Cards/proc_card_mg5.dat .
> emacs proc_card_mg5.dat
> #-> replace the line output by output PROCESSNAME
> ./bin/mg5 proc_card_mg5.dat
>
> Cheers,
>
> Olivier
>
> On Jan 6, 2014, at 3:15 PM, Rakhi Mahbubani <email address hidden>
> wrote:
>
>> Yes, the function is in both files.
>>
>> R
>>
>>
>>
>> =========================
>> Chemin de l'Esplanade 26
>> 1214 Vernier
>> Suisse
>>
>> Tel: [phone]
>> Mob: [phone]
>> ==========================
>> ==========================
>>
>> "One does not discover new lands without consenting to lose
>> sight of the shore for a very long time." - Andre Gide
>>
>>
>> ________________________________
>> From: Olivier Mattelaer <email address hidden>
>> To: <email address hidden>
>> Sent: Monday, January 6, 2014 2:08 PM
>> Subject: Re: [Bug 1261931] Re: a...

[Rakhi Mahbubani (r-mahbubani)]

Fantastic! I got rid of the .o files in the HELAS directory and everything now works!!  I'm v happy :)

Thanks so much for your help Olivier!

R

=========================
Chemin de l'Esplanade 26
1214 Vernier
Suisse

Tel: [phone]
Mob: [phone]
==========================
==========================

"One does not discover new lands without consenting to lose
sight of the shore for a very long time." - Andre Gide

________________________________
 From: Olivier Mattelaer <email address hidden>
To: <email address hidden>
Sent: Wednesday, January 8, 2014 1:41 PM
Subject: Re: [Bug 1261931] auto-update problem

Hi Rakhi,

I'm confuse about this statement:

>  I think I have also had this problem before when trying to
> use a model folder created with an older version of madgraph, to
> generate a process in a newer version…

MadGraph didn't create model… So what do you mean by that?

> is there some issue with
> backwards-compatibility of the v4 model folders?

Not that I know.

The bug that you mention, is doesn't seems to be related to that.
In this case, we have add one function the boostm and looks like this one is not found.
For v4 model this function is in the HELAS/boostx.F file which is automatically copy to
Source/DHELAS

I might have one "idea", do you have some .o file in HELAS directory?
If yes, please remove those, this might be the reason of the crash.

Cheers,

Olivier

On Jan 8, 2014, at 10:57 AM, Rakhi Mahbubani <email address hidden> wrote:

> Dear Olivier,
>
> No, I don't do a manual cp of the Template directory, but generate the
> process (using an existing model directory) using the python interface
> of madgraph.  I think I have also had this problem before when trying to
> use a model folder created with an older version of madgraph, to
> generate a process in a newer version...is there some issue with
> backwards-compatibility of the v4 model folders?
>
> Cheers,
>
> Rakhi
>
>
>
> =========================
> Chemin de l'Esplanade 26
> 1214 Vernier
> Suisse
>
> Tel: [phone]
> Mob: [phone]
> ==========================
> ==========================
>
> "One does not discover new lands without consenting to lose
> sight of the shore for a very long time."    - Andre Gide
>
>
> ________________________________
> From: Olivier Mattelaer <email address hidden>
> To: <email address hidden>
> Sent: Monday, January 6, 2014 3:42 PM
> Subject: Re: [Bug 1261931] auto-update problem
>
>
> Then the update works correctly…
>
> So the problem is probably due to a wrong use of the Template.
> I guess that you do manual cp of the Template directory? (the v4 way to generate directory, which is not supported anymore since v2.0)
> and then use the ./bin/newprocess_mg5 script.
> If so, I guess that the problem is that your copy of the Template was done before the update. Is this possible?
>
> I would advice to do the following:
>
> cp PROCESSNAME/Cards/proc_card_mg5.dat .
> emacs proc_card_mg5.dat
>    #-> replace the line output by output  PROCESSNAME
> ./bin/mg5 proc_card_mg5.dat
>
> Cheers,
>
> Olivier
>
> On Jan 6, 2014, at 3:15 PM, ...