rspamd: Update to 2.5

from GitHub user @sjorge + extra patch from me

Closes NetBSD/pkgsrc#60

2.5: 01 Apr 2020
  * [Conf] Mark Rspamd emailbl as ignore whitelist
  * [Conf] RBL: Add missing emails = true option
  * [Feature] Add support for scripts in fuzzy storage
  * [Feature] Arc: Add whitelisted_signers_map option
  * [Feature] Implement hosts file processing
  * [Feature] Neural: Introduce classes bias that allows non-equal classes learning
  * [Feature] Update libev to 4.33
  * [Fix] Another brain damage html standard adoptions
  * [Fix] Another fix for brain damaged obs-fws state
  * [Fix] Fix flags that caused force_actions failure
  * [Fix] Fix logging issue
  * [Fix] Fix lua symbols scores registration when config does not define scores
  * [Fix] Fix opaque maps logic
  * [Fix] Fix parsing of the html tags with no spaces after attributes
  * [Fix] Fix some corner cases in urls parsing, add limits
  * [Fix] Fix tlds extraction if custom composition rules are used
  * [Fix] Fix variables replacement in mempool
  * [Fix] Improve base64 detection
  * [Fix] Normalize dynamic scores in ANN correctly
  * [Fix] Plug memory leak introduced by #3153
  * [Fix] Stat_redis_backend: Fix memory leak and simplify learn path
  * [Fix] Try hard to deal with ghost workers
  * [Fix] metadata_exporter default formatter
  * [Rework] Change the way to extract URLs when dealing with alternative parts
  * [Rework] Fix various url extraction issues
  * [Rework] Re cache: Load compiled hyperscan in the main process as well
  * [Rework] Re cache: Load hyperscan early
  * [Rework] Rework URL structure: adjust tld part
  * [Rework] Rework URL structure: host field
  * [Rework] Rework URL structure: more structure optimisations
  * [Rework] Rework URL structure: user field
  * [Rework] URL: Another update for urls extraction logic
  * [Rework] Urls: Improve query urls handling
  * [Rework] Urls: adopt html related stuff
  * [Rework] Urls: more rework of the urls sets
  * [Rework] Urls: process query urls in HTML urls correctly
  * [Rework] Urls: rework urls hash structure
  * [Rework] Urls: update lua libraries
  * [Rework] Use multiple search tries for different url extraction types

2.4: 26 Feb 2020
  * [CritFix] Fix parsing of the content type attributes
  * [Feature] Clickhouse: Add extra columns support
  * [Feature] Rbl: Add url_compose_map option for RBL rules
  * [Fix] 'R' flag is for all headers regexp
  * [Fix] Allow to reset settings id from Lua (e.g. because of the priority)
  * [Fix] Avoid collisions in mempool variables by changing fuzzy caching logic
  * [Fix] Avoid strdup usage for symbols options
  * [Fix] Do not trust stat(2) it lies
  * [Fix] Filter all options for symbols to have sane characters
  * [Fix] Fix all headers iteration
  * [Fix] Fix allowed_settings for neural
  * [Fix] Fix listen socket parsing
  * [Fix] Fix maps expressions evaluation
  * [Fix] Fix sentinel connections leak by using async connections
  * [Fix] Fix smtp message on passthrough result
  * [Fix] Fix tld compositon rules
  * [Fix] Fuzzy_storage: Do not check for shingles if a direct hash has been found
  * [Fix] Lua_mime: Do not perform QP encoding for 7bit parts
  * [Fix] Neural: Distinguish missing symbols from symbols with low scores
  * [Fix] Support listening on systemd sockets by name
  * [Project] Add lua_urls_compose library
  * [Project] Allow to set a custom log function to the logger
  * [Project] CDB maps: Start making cdb a first class citizen
  * [Project] Clickhouse: Add extra columns concept
  * [Project] Fix urls composition rules, add unit tests
  * [Project] Unify cdb maps
  * [Rework] Logger infrastructure rework
  * [Rework] Refactor libraries structure
  * [Rework] Rework SSL caching
  * [Rework] Update snowball stemmer to 2.0 and remove all crap aside of UTF8

(nia)