Cleaning up template.distillery files (for 1.4 release)

[RogerTarani]

1/ In Makefile.os there are 10 times -ppx -ppx.
Is it normal?
As a test, replacing them by -ppx had no effect on make byte (but had effect on make opt that failed, as it seems to be useful to eliomopt. Anyway it's not blocking)

2/ Replacing deprecated Eliom_content.Html.F.pcdata with Eliom_content.Html.F.txt saved about 700 warning lines when doing make byte (out of about 1200 lines).

List of deprecated modules or functions:
Eliom_content.Html.F.pcdata -> JEliom_content.Html.F.txt
Js -> Js_of_ocaml.Js
Dom_html -> Js_of_ocaml.Dom_html
Dom -> Js_of_ocaml.Dom
Firebug -> Js_of_ocaml.Firebug

Can you update the ocsigen-start *.eliom files in template.distillery/ for the next 1.3.1 release?
It will make the build much less verbose.

You may do it with a sed command much more quickly than handling a PR.
I used the following command for example for Js -> Js_of_ocaml.Js :
$ sed -i 's/Js./Js_of_ocaml.Js./g' $(grep -l "Js." *.eliom)

3/ In demo_carousel1.eliom L33, the function To_dom.of_element belongs to the module Eliom_content.Html.
As I had a "Warning 33: unused open Eliom_content.Html." I deleted this open Eliom_content.Html directive, but then the compiler said Unbound module To_dom.
Tweaking theses directives, exchanging their order had no effect. At last, I just opened the module as a prefix to To_dom:

L34  let%lwt () = Ot_nodeready.nodeready (Eliom_content.Html.To_dom.of_element ~%carousel) in

Can you explain why this "unused open Eliom_content.Html" happened?

4/ In file "oc_1_3_0_test2_i18n.eliom" there are several Warning 27: unused variable f (4 lines).
(cause : line 147 and 852)

| En -> (fun ?(capitalize=false) ?(f=false) () -> List.flatten [[txt (if capitalize then "V" else "v")];[txt "alidated"]])
Inserting a null test worked:

L147 | En -> (fun ?(capitalize=false) ?(f=false) () -> List.flatten [[txt (if capitalize then "V" else "v")];[txt "alidated"];[txt (if f then "" else "")]])

L852 | En -> (fun ?(capitalize=false) ?(f=false) () -> String.concat "" [(if capitalize then "V" else "v");"alidated";(if f then "" else "")])

Does this null test still makes sense according to you?

Now I have a zero warning build.