When multiple repositories exists and two of them contain scalability_test package be wiser and really check if neeeded Test Suite class exists before trying to load it.

When running tests the path to git command is just git, because we have
set $PATH

updateConfiguration is a method to set the list of partitions, not to
add partitions to an existing computer. The previous (slapos.core
<1.5.0) implementation of slap proxy was incorrect and was adding more
partitions each time it was called, but that was never the indented
behavior.

Prepare the list of partitions and call updateConfiguration only once to
set all partitions.

This should fix SLAPOS-SR-TEST that are still using erp5.util

Some tests where patching SlapOSControler to replace methods by a no-op,
but this was never reset. I believe we should rewrite this using mock
which takes care of un-doing the patches for us and makes this testing
easier, but for now I'm just reseting the missing methods like all
others were already reseted.

configparser read options as strings, node_quantity is not supposed to
be an int.

pkg_resources.resource_string reads the file as binary, if we want to
substitute in this string we need to decode it.

Value of text_content is an unicode string whose (inside XML) encoding is utf-8. This inconsistency raises in xml library thus it is safer to convert back to utf-8 which is the intended encoding anyway.

See slapos.core@42b55b5f

all repository list which will cause a mismatch between test suite class
location and repository.

/reviewed-on !945

Editing instance parameters will automatically trigger instance update at node and there's no need to stop / start instance.

evaluated before timeout and thus causing unexpected AttributeErrors
later.

password pair generated (in some cases this might happen after frontend
generation leading to 403 HTTP error codes in bootstrap code.)

Cleanup conditional code.

code will fail with :
"local variable 'user' referenced before assignment".
As callee is handling the possibility of delayed frontend and / or
instance creation this is perfectly fine.

   title is undefined, use self.name instead.

31804f68 was merged too soon.

/reviewed-on !913

  The API changed on the slap, this commit follow up the changes.

/reviewed-on !830

The way of truncating the index file was an error in git 2.11.0, but it
was not an error with git 2.17.0.

/reviewed-on !863

This allows to know the node working on test, thus we can redraft
lines when a runTestSuite command fails

Right now we have this scenario:
- test result line is started
- sometimes, runTestSuite fails (like timeout), failure is reported
  but test result line remains started (we don't know yet which
  line is associated with testnode)
- when a test result line is "started" since more than 4 hours,
  test result line is redrafted
- test can be reexecuted
Speed up by removing the need of waiting alarm, by knowing which
test result line is executed by which test node, and by redrafting
immediately the test result line on test node failure

We randomly have unkilled firefox process remaining. Make sure
to kill them.

As leaving close_fds=False can result in subtle bugs (see
http://zderadicka.eu/subtle-evil-of-close_fds-parameter-in-subprocess-popen/ )
close descriptors before launching anything with ProcessManager, especially
useful for starting things like supervisord.

/reviewed-on !841

supervirord when stopping might need several seconds to kill
properly sub processes. So give him some time before killing
more brutaly with killall.

This should solve issue with functional tests are not working
due to apache not starting. Apache when killed to brutally
let unclosed semaphore, and after some time, we have too many
semaphore to start new apache, making functional tests failing.

Kill any process having in command line the path reserved for unit
test. This would really allows to kill any remaining process if any.

Do not look if processes are child of testnode, because processes
like mariadb and others are not running as child

Kill any process having in command line the path reserved for unit
test. This would really allows to kill any remaining process if any.

Do not look if processes are child of testnode, because processes
like mariadb and others are not running as child

* Set MAKEFLAGS with -l to limit number of jobs by load and number of
processes
* set NPY_NUM_BUILD_JOBS & BUNDLE_JOBS to `node_quantity` instance
parameter

/reviewed-on !809