This is necessary for Libreoffice 7, where uno from Libreoffice python needs
LD_LIBRARY_PATH to be set.

This uses psutil.process_iter to iterate on all processes, but in the meantime
some process might terminate.

As we could observe in the logs:

    2021-04-26 12:53:02 - Cloudooo - DEBUG - Stop Pid - 15816
    Exception in thread Thread-1:
    Traceback (most recent call last):
      File "python2.7/lib/python2.7/threading.py", line 801, in __bootstrap_inner
        self.run()
      File "cloudooo-repository/cloudooo/handler/ooo/monitor/request.py", line 60, in run
        self.openoffice.restart()
      File "cloudooo-repository/cloudooo/handler/ooo/application/application.py", line 79, in restart
        self.stop()
      File "cloudooo-repository/cloudooo/handler/ooo/application/openoffice.py", line 169, in stop
        self._releaseOpenOfficePort()
      File "cloudooo-repository/cloudooo/handler/ooo/application/openoffice.py", line 122, in _releaseOpenOfficePort
        if process.exe() == join(self.office_binary_path, self._bin_soffice):
      File "psutil-5.8.0-py2.7-linux-x86_64.egg/psutil/__init__.py", line 660, in exe
        exe = self._proc.exe()
      File "psutil-5.8.0-py2.7-linux-x86_64.egg/psutil/_pslinux.py", line 1688, in exe
        raise NoSuchProcess(self.pid, self._name)
    NoSuchProcess: psutil.NoSuchProcess process no longer exists (pid=19382)

Update to catch and ignore such exceptions, and minor refactorings such as:
 - Not calculate the path of openoffice binary in the loop, this is not
   supposed to change.
 - Remove reference to lsof in logged message, this does not use lsof.
 - Catch all exceptions.

In python2, mimetypes module was not deterministic and this snippet:

    import mimetypes; print(mimetypes.guess_extension("text/html"))

use to be ".html" on python 2.7.14, but is ".htm" on python 2.7.18

Similarly:

    import mimetypes; print(mimetypes.guess_extension("application/msword"))

was ".doc" on 2.7.14 and ".dot" on 2.7.18 (this was in my observations, this
does not look deterministic as it is iterating on a dict, maybe this behaviour
is not always same)

For html conversion engine, ERP5 expect that the extension for text/html is
.html, not .htm. Some tests also expect that the conversion for word is .doc so
to keep compatibility with extensions used in ERP5 compatibility mode, define
explicitly these two extensions instead of depending on python standard library.

See merge request !25

See merge request !24

Tests passing at
https://nexedi.erp5.net/test_result_module/20200102-6DAB14B9

/reviewed-on !23

/reviewed-on !22

Due to a5157949 cloudooo will retry 10 times before raising thus take this into account
Raise rather than swallow.

Test do pass at
https://nexedi.erp5.net/test_result_module/20191227-20A35E0F

/reviewed-on nexedi/cloudooo!21

/reviewed-on !18

using python csv module

latin1 is still supported as a fallback when input csv does not decode
as utf8

Split into several test classes for easier maintainance

no need for test_suite everywhere

The cloudooo test on erp5-master is not able to finish after last erp5.util release: https://nexedijs.erp5.net/#/test_result_module/20180719-86977F42

After fix, there is a cloned test suite and the test is completed: https://nexedijs.erp5.net/#/test_result_module/20180719-781978C2 , but some test fails.

I don't know these particular tests neither if they are failing due to the egg release, but the change in this MR solves the problem with test execution and it is ready to merge.

/reviewed-on nexedi/cloudooo!17

(cherry picked from commit e006854)

(cherry picked from commit 50329148)

/reviewed-on nexedi/cloudooo!14

`cloudooo.handler.yformat.handler` was mainly contributed by
Boris Kocherov. Commit 0bb5fbdc
should have been credited to him.

fix backward compatibily + update test

- source format is html
- destination format is pdf
- conversion_kw is not empty