v2.4.0

Version 2.4.0 (2020-05-16)

This release drops special Web browser requirements to access dasht-server(1),
improves error messages in dasht-docset-install(1), ensures clean up of stale
files that contain whitespace in their names, and resolves a few other issues.

Minor:

• dasht-server-http(1): serve local file:// URLs through http:// URLs.

  Users no longer have to use special browsers or special configurations
  that allow the loading of file:// links from http://127.0.0.1 URLs.
  All browsers can now natively access the dasht-server(1) search engine.

  See #45

Patch:

• dasht-docsets-update(1): remove stale files with whitespace in names.

  Thanks to @frodeaa for reporting this issue and contributing a patch:

  Replace xargs with while read and avoid globbing. The same could be
  achived by using xargs -0 (GNU/BSD) if newline is replaced with \0.

  tr '\n' '\0' | xargs -0 sh -e -u -c '
  

  See #35

• dasht-server-http(1): escape "C++" docset name as ^C\+\+$ regex.

  When the "C++" docset was selected from the "in docsets" dropdown menu in
  dasht-server(1), the plus signs weren't properly escaped per regex syntax.

  See #49

• dasht-docsets-install(1): validate gzip(1) integrity before extraction.

  This avoids cryptic errors from tarball extraction:

  gzip: stdin: unexpected end of file
  tar: Child returned status 1
  tar: Error is not recoverable: exiting now
  

  By pointing out the cause of the error to the user:

  gzip: Apple_Guides_and_Sample_Code.tgz: unexpected end of file
  

• dasht-docsets-install(1): restore support for older wget(1) versions.

  Older wget versions don't support the --show-progress option:

  wget: unrecognized option '--show-progress'
  Usage: wget [OPTION]... [URL]...
  
  Try ‘wget --help’ for more options.
  

• dasht-query-line(1): update URI fragments for HTML and CSS docsets.