Simplify map.jinja revert deep_merge solution from #193

[aboe76]

In a discussion on #193 and a pr #248 there is a attempt to make map.jinja easier to maintain and functional.

The deep_merge solution was created based on the fact that centos 6 didn't have mapping support in jinja2, but with the saltstack repos and epel repos having a newer version of jinja2 > 2.6 this is no longer necessary. Jinja2 from 2.6 and higher support mapping.

I have tested this new setup, with defaults.yaml, osfamilymap.yaml and osmap.yaml on the following systems without issues:

• centos 7
• redhat 7
• opensuse
• fedora 29
• ubuntu 18.04
• debian 9
• Centos 6 ( clean minimal install with only saltstack repo added)

!!!! Please be carefull with merging, could lead to unexpected behavior without testing first. !!!!

[myii]

@aboe76 Initial tests are getting consistent results so far. One section of the rendering is different, though:

--- master
+++ aboe76-simplify-map.jinja
@@ -1,13 +1,9 @@
 'gitfs': {
     'pygit2': {
         'git': {
-            'install_from_package': None,
+            'install_from_package': 'None',
             'require_state': False
         },
-        'install_from_source': False,
-        'libgit2': {
-            'install_from_source': False
-        },
-        'version': '0.22.1'
+        'install_from_source': False
     }
 },
     }
 },

---

Update: Forgot to mention, these are various Ubuntu minions.

[myii]

Ensures 'install_from_package': None rather than as a string: 'None'.

[myii]

Changed the title to prefix with WIP: to prevent an accidental merge, which works on GitLab. What's the equivalent in GitHub?

[aboe76]

@myii about the title, thanks for changing it to something usefull, about protecting from PR merging without aproval with github, this is only possible in settings, which is unavailable at our level in this organization....see https://help.github.com/articles/enabling-required-reviews-for-pull-requests/

[myii]

@aboe76 Got a new diff for you after 66c0699:

--- master
+++ aboe76-simplify-map.jinja
@@ -1,5 +1,10 @@
- 'gitfs': {'pygit2': {'git': {'install_from_package': None,
+ 'gitfs': {'dulwich': {'install_from_source': True},
+           'gitpython': {'install_from_source': False},
+           'pygit2': {'git': {'install_from_package': None,
                               'require_state': False},
                       'install_from_source': False,
-                      'libgit2': {'install_from_source': False},
-                      'version': '0.22.1'}},
+                      'libgit2': {'build_parent_dir': '/usr/src/',
+                                  'download_hash': '683d1164e361e2a0a8d52652840e2340',
+                                  'install_from_source': False,
+                                  'version': '0.23.0'},
+                      'version': '0.22.1'}},

As far as I can tell, this is an improvement, where the defaults are being merged properly rather than overwritten. Meaning, it appears that the current method used in the master branch has a problem and this PR will fix that.

[aboe76]

@myii nice, thanks for reviewing,

Installing a centos 6 vm to test against.

[noelmcloughlin]

I cannot see any code issues, LGTM. thanks @aboe76 and @myii

[aboe76]

@noelmcloughlin would like some more input before merging from:
@javierbertoli

and maybe somebody from:
@saltstack-formulas/core
@saltstack-formulas/moderators

[javierbertoli]

It LGTM too. Nice refactoring, @aboe76! As soon as you don't consider it more a WIP I think we're OK to merge?

[myii]

@noelmcloughlin @aboe76 Can this be merged? I've got another PR to submit for PY3 support and I've based it on these changes since I don't want to have to redo it again when this gets merged.

[noelmcloughlin]

+1 merge. @javierbertoli gave +1 while back.

[myii]

@noelmcloughlin Great, I've gone ahead and removed the references to WIP. Are we still waiting for another review from one of the groups mentioned above?

[noelmcloughlin]

No merge away!!!

[myii]

Confirmed all OK with @noelmcloughlin via. Slack.

Great work, @aboe76. Thanks to @javierbertoli and @noelmcloughlin for the reviews.

[aboe76]

@myii, @noelmcloughlin and @javierbertoli thanks for reviewing..