fetch_all_by_external_name() in GeneAdaptor.pm, TranscriptAdaptor.pm,
and in TranslationAdaptor.pm.

This will limit the returned genes, transcripts, or translations, to
those that originates from the named external database.

This also means adding the optional argument to
list_gene_ids_by_extids(), list_transcript_ids_by_extids(),
list_translation_ids_by_extids(), and to _type_by_external_id() in
DBEntryAdaptor.pm.

The matching is done on 'db_name' in the 'external_db' table with LIKE
and an appended '%' to the given $external_db_name.

new code to work out the cmp seq regions that have multuple assemblys. This is now stored in the seq_region_attrib table

    Allow aliases to also be specified in the 'default' section, but
    require that there must also be a 'species' setting in the same
    section.  Aliases will never be installed if there is no 'speices'
    setting in the current section.  The 'species' section may be
    inherited from the 'default' section.

    Example:

    [default]
    host    = db.example.com
    user    = anonymous
    species = Human
    alias   = H_Sapiens
    alias   = Homo_Sapiens
    alias   = Me

    [my_core_db]
    group   = core
    dbname  = homo_sapiens_core_45_36g

    [my_estgene_db]
    group   = estgene
    dbname  = homo_sapiens_estgene_29_35b

The Registry will try to load the specified file (specified by argument
to the method load_all(), in the envrionment variable ENSEMBL_REGISTRY,
or the file '~/.ensembl_init', as before) as an ini-file.  If this fails
due to the file having the wrong format, the code reverts to the old
behaviour of using require() on the same file.

This is an example of a simple ini-file.  There is no check to see
that e.g. 'dbname' is actually present or to avoid calling the adaptor
initialization routines with unnecessary parameters (we trust the user
to get this right).  The correct adaptor is chosen by peeking at the
'group' setting.


; Simple Ensembl config file example

; The 'default' section is special and its settings will be implicit for
; all other sections in this file but may be overridden by individual           ; sections.  There should be no 'alias' settings in the 'default'
; section.

[default]
user    = anonymous
host    = ensembldb.ensembl.org
group   = Core

[the_human_db]
species = Human
dbname  = homo_sapiens_core_45_36g
alias   = mary
alias   = john

[the_mouse_db]
species = Mouse
dbname  = mus_musculus_core_45_36f
alias   = minnie
alias   = mickey

; Configuration ends here

The quotes are no longer necessary since the bind_param statement now explicitly specifies the type, removing the nasty MySQL casting issue that caused them to be added in the first place.

    cdna_start()
    cdna_end()

    cdna_coding_start()
    cdna_coding_end()

    coding_region_start()
    coding_region_end()

Changed POD from saying "in genomic coordinates" to say "in
slice-relative coordinates":

    coding_region_start()
    coding_region_end()

added NOTE about fetching by primary accession not being bothered about what the external database is

  implementation
- added fetch_associated_archived()
- removed deprecated methods

Only clear old registry configuration if we have a configuration file
name and if that file exists.

- rewrote fetch_by_stable_id*() methods to look up current versions of stable
  IDs in *_stable_id tables
- new method lookup_current() to assist the former (also replaces
  _lookup_version())
- new method add_all_current_to_history() to implement this behaviour for
  StableIdHistoryTree

- new methods to get properties of current db:
  - get_current_release()
  - get_current_assembly()

- fixed _resolve_type() to copy with non Ensembl-style stable IDs (e.g.
  Anopheles)

- removed deprecated methods

  - get_latest_StableIdEvent(): find latest Event in tree for a given stable ID
  - getters/setters for current db properties:
    - current_release()
    - current_assembly()
    - current_dbname()
  - is_incomplete(): flag to indicate incomplete trees

  - current_version(): retrieves current version of stable ID from db
  - is_current(): test if this is the current version
  - get_latest_incarnation(): get latest incarnation of stable ID from db
  - is_latest(): test if this is the latest version

- version() no longer lazy-loaded (now always set by adaptor)

- removed deprecated methods

as filled paragraphs.  Long lines shortened.  Spelling.