Import improvement from robvdl fork

[claudyus]

This thread is to track improvement that could be imported from the robvdl fork, otherwise we cannot simple cherry-pick commits due to his deep app rework.

[claudyus]

https://github.com/robvdl/LXC-Web-Panel/commit/318ad33642afbc8424bc1a8dcf28ac1f531a2b04

[claudyus]

https://github.com/robvdl/LXC-Web-Panel/commit/e1a379d8ff0589c5b3f5790a5333849821cdbd46

[claudyus]

https://github.com/robvdl/LXC-Web-Panel/commit/0d4a9fcfc16328d80768b18de7ccc2cf661b6892

[claudyus]

https://github.com/robvdl/LXC-Web-Panel/commit/e10facbc35f675ff4a3c12faf5820ba3c16df551 investigate uppercase in check_ubuntu

[claudyus]

https://github.com/robvdl/LXC-Web-Panel/commit/09d161d8ad37e70352d0fd4214e882fca979818c

[claudyus]

Migrate to wtform as per https://github.com/robvdl/LXC-Web-Panel/issues/31, investigate for ipv4/mad validation

TODO after 0.7 release

[robvdl]

Claudyus: The edit form is almost done, I've also added in a few fields that I noticed you added recently, these are:

id_map, hook_pre_start, hook_pre_mount, hook_mount, hook_start, hook_post_stop and hook_clone

I am pretty sure these new fields are working (saving) correctly, I ended up added them to the form as I noticed they were defined in code, but not yet in the form.

There is still one experimental thing I am trying right now in another branch of my repo, and that is to unify the backend validation regex, and pattern on the from end... if it works (I am not sure yet)

Then, I am hoping I can automatically generate the front-end pattern from the backend regex, so you only need to the define the regex once per field. In most cases, the JavaScript regex style is compatible with Python's anyway.

When this is all done, I am happy to help bring all this WTForm work into your repo, and do this in a branch on your repo so you can look at it.

I suggest the bit where I am using my SQLAlchemy models to update the bucket token, to just use your existing SQL queries as I don't really want to bring in both the forms and the model stuff at once.

[robvdl]

One of the things that I am having to deal with, is that sometimes complexity of the front end regex (for example utsname) is greater (so generally better and more accurate), but sometimes the backend regex is better.

So as I am unifying the regexes I need to basically decide which one is best, if the backend and front end regexes are different.