Linux version, little problem at installation

[temla]

Dear All,

I went to Japan for two years (from Nov. 2), and I have not so many
times for try RMCProfile until now. I have just started to install,
and a (little) problem occured at the installation: in the "exe"
library accessibility of all files have set to 644, but these files
are executable (they should be 755 or something similar).

I suggest two ways to solve the problem:
- 1st solution: distribution in ".tar.gz" format, which reserve the
accessibility of the files (as far as know it), or something possible
in zip (I do not know the possibility of "zip").
- 2nd solution: It should be put two additional line into the
RMCProfile_setup file after "export RMCProfile_PATH" (the second one
is not so important for the right working of the installation):
chmod 755 $RMCProfile_PATH/exe/*
chmod 644 $RMCProfile_PATH/exe/*.*

Best regards,

Laszlo

[Matt Tucker]

Dear Lazlo,

Thank you for pointing this out, I will fix this in the next Linux
distribution, I will probably go with option 1 if possible.
I suspect the zipping or unzipping cause the problem so I will
use .tar.gz as you suggest.

I hope you have fun in Japan.

Best wishes,
Matt

[temla]

Dear Matt,

First of all, thanks you for the wishes.

I successfully completed the Tutorial part of RMCProfile (I say sorry,
but nowadays I have not been so much free time..).

I found the following troubles in the Linux version:

1: After I have ran rmcprofile according to exercise 1.:
a.) Two rows are appearing with a number (1678 and 1693). I suppose,
they are not removed after the debug procedure):

>INFO> Number of coordination constraints assumed to be zero
> 1678
>INFO> Number of average coordination number constraints assumed to be zero
> 1693
>INFO> No magnetic scattering included

b.) The x component of the a axis vector of the sample configuration
file possesses a typing error: 8.83015 instead of 8.8315 . This little
error appears at that .cfg files, which are contained in the ex_1 and
ex_2 (the cfg file in ex_3 is correct) libraries.

c.) Just a comment for Linux users: if somebody uses Debian, at least
the "lib32g2c0" package installation is required to use the RMCPlot
properly.

2. Exercise 2:
a.) After I ran "get_gsas_bragg", I have got some interesting error
messages. The solution of the problem is the next:
- I installed gsas into a proper place (in my case:"/usr/local/gsas",
for Debian users: some "lib32...." libraries are required to work this
program properly)
- After that, the "get_gsas_bragg
[where_the_gsas_exe_programs_are]" (in my case: "get_gsas_bragg /usr/
local/gsas/exe/") command works properly.

I suggest to add some sentences to the tutorial at this step: please
install GSAS, before you try to run "get_gsas_bragg", and if GSAS were
not in "/Applications/gsas/exe/" library, then it would run as
"get_gsas_bragg where_the_gsas_exe_programs_are"

b.) At rescaling of the G(r), in contrast to the tutorial, the writing
of the
whole filename is needed (in the other case, program will give error
message), and the order of the prepared answers are not the same as in
the tutorial (e.g. shifted along x is missing). For instance, the next
output produced a correct output file:

>temla@nfo2ltp:~/RMCProfile_trial/ex_2/data$ data_rescale
> Input file > sf6190k_all_gr.dat
> contanst to subtract
>1.0
> constant to multiple by > 0.27693
> Output file > sf6190k_all_gr_rmc.dat

c.) If somebody uses the provided .cif file, then (s)he will get
opposite order of the atoms (F S) instead of (S F).

d.) The Linux version package does not contains the "rmc_to_atomeye"
program.

OK, that's all.

Best regards,

Laszlo

[Martin Dove]

A quick comment on some bits, leaving Matt to address the bits he
knows more about ...

On Nov 18, 6:01 am, temla <te...@szfki.hu> wrote:
>
> 1: After I have ran rmcprofile according to exercise 1.:
>   a.) Two rows are appearing with a number (1678 and 1693). I suppose,
> they are not removed after the debug procedure):

Yes, these are old debug statements, and bit by bit I try to remove
them! There are probably a few more lurking.

> b.) The x component of the a axis vector of the sample configuration
> file possesses a typing error: 8.83015 instead of 8.8315 . This little
> error appears at that .cfg files, which are contained in the ex_1 and
> ex_2 (the cfg file in ex_3 is correct) libraries.

Support for .cfg files may cause some problems sometimes. The idea is
that users use the more modern .rmc6f format, and we have provided
tools to convert from .cfg to .rmc6f format.

Further action is for Matt is who responsible for ensuring backward
support.

> c.)  If somebody uses the provided .cif file, then (s)he will get
> opposite order of the atoms (F S) instead of (S F).

Quite possibly, because there is an ordering by element number. The
next versions of rmcprofile and data2config have a keyword to avoid
this. I am not sure whether you have the latest version of data2config
(you won't have the latest rmcprofile because only I have this).
However, I would remark that by using element names the exact order
need not matter, and I would like to ensure that we remove absolutely
all dependence on order of atoms in the files.

Best wishes
Martin