EZP-26277: Content type copies generated identifiers are too long

[alongosz]

Status: Ready for a review
Fixes: EZP-26277
Backported to: 6.4

This PR shortens generated identifiers of content type copies.
Instead of:
copy_of_<identifier>_<remoteId>
now we have:
cpy_<identifier>_<remoteId>

Also, <remoteId> is no longer a md5 hash but a sha1 hash shortened to 7 chars. We can't get rid of <remoteId> in copied content type identifier because it would be impossible to copy content type several times.

Another possible approaches:

• Make database identifier field longer than 50 chars - as suggested in JIRA issue description. However AFAIK performance of varchar varies between DBMS-es and should be tested on large databases (about 2 mln of Content Objects).
• Force API Consumer to specify copied content identifier. This is an API BC break and should not be considered for 6.x (IMHO).

TODO

• Make copied content type identifier shorter.
• Align tests and doc with this change.

[alongosz]

Review ping @bdunogier @andrerom @pspanja

[bdunogier]

I was about to say that I'd change cpyto cp, as it is a known shorthand for copy, but obviously too late.