Miscellaneous updates/Convenience script

The merge request suggests the following changes:

- Removal of trailing spaces in visit.ml.
- Adjustment of the `stdio.c` test to prevent it from writing a test file into the project's root directory.
- Added a convenience script for runs of E-ACSL/GCC. Functionality to `install` and `uninstall` executable files from the `scripts` directory has also been added. 

See merge request !19

   add an option to pass specific flags directly to frama-c

   added option to redirect all output to a given file

errors or warnings

  Used frama-c-config tool instead of frama-c executable to retrieve share path

    - Renamed to e-acsl-gcc.sh
    - BUGFIX: Clean abort on frama-c/gcc failures
    - Use an explicit list of disabled warnings instead of plain '-w'
    for gcc runs

using E-ACSL plugin and their subsequent compilation using GCC.

into the project's root

Kostyantyn/bugfix/global initialisers

This is the fix for the issue `0001817`: `Literal strings in global arrays with compound initializers are not correctly initialized`. The change is as we have discussed during our meeting on the 7th of December. The gist of the fix as follows:

- Initialisers of all global variables are kept (instead of being removed), so as the code used for initialisations within the body of the `__e_acsl_memory_init`. 
- Literal strings used in initialisations are recorded via special global variables. 

The fix relies on the fact that a compiler will maintain only a single instance of a literal string, so recording a literal string once will cover all its uses. 

For example, strings pointed to by some global variables `A` and `B`:

```
...
const char *A = "X";
const char *B = "X";
...
```

will be captured by:

```
...
__e_acsl_literal_string = "X";
__store_block((void *)__e_acsl_literal_string,sizeof("X"));
__full_init((void *)__e_acsl_literal_string);
__literal_string((void *)__e_acsl_literal_string);
...
```

There are some test failures, as some code generation has changed. I have intentionally left them in for the fist time so you can see the tests that fail. If this patch is accepted I am happy to fix the test suite and add a couple of new tests to cover the issue. I have also attached a file with some examples of global initialisers for different cases which you may find useful.

[compound_initialiser.c](https://git.frama-c.com/frama-c/e-acsl/uploads/3599d1cbca9c790d6dd231d052ccf2f0/compound_initialiser.c)


See merge request !18

    doc/Changelog

- Updated CI configuration

    initializers
- Changes to the oracles in order to accommodate the changes to code
    generation

variables (which point to literal strings) generated by E-ACSL

times

global initialisers or not) with a boolean switch called
is_initialiser used to detect and filter out compound initialisers
and update [global_vars] mapping

during global initialisation

declarations

variables (which point to literal strings) generated by E-ACSL

times