Paste from the changelog:
There have been a lot of changes. Here are the highlights:
* Improved command line parser: terminator -- handling, UUID
recognition, DOM recognition, red-herring pairs (foo:bar), escaped
slashes in patterns (/one\/two/), substitutions (/one\/two/one-two/),
Unicode U+NNNN and \uNNNN, escaped entities (\n, \t etc) for use in
descriptions/annotations, abbreviated day and month names, ISO-8601
durations (PT4H, P1Y etc).
* New virtual tags UDA, ORPHAN, PROJECT, PRIORITY, and LATEST.
* Improved support for DOM references in filters task
'due.year = 2015 and due.week > 20' list.
* New configuration settings urgency.inherit, rule.color.merge,
urgency.user.tag.next.coefficient, color.uda.<name>.none, sugar,
report.<name>.sort:none, json.depends.array.
* Improved urgency: the urgency.inherit setting propagates urgency
along dependency chains.
* Improved searching: more powerful regular expressions.
* Improved attribute modifiers: is/isnt are now consistent exact
match operator equivalents to ==/!==.
* New command _unique, for generating unique lists of values, and
will ultimately replace several helper commands.
* New command commands, which lists commands and details about how
they affect filters, GC, context, and syntax.
* New verbosity tokens recur (feedback when a recurring task instance
is created), unwait (for when a waiting task becomes visible).
* Improved extensions: zsh completion, Fish shell completion, add-on
scripts now online.
* Improved documentation: help command, man pages, more online docs,
negation ! operator, sample hook scripts.
* Improved sync: GnuTLS now mandatory, so everyone has the sync
command.
* Improved JSON import/export support: free-format JSON, task arrays
assumed (which makes import/export work with out-of-the-box parsers),
dependencies optionally modeled as an array, UUIDs validated, tasks
added/updated without duplication, import from STDIN using -.
* Improved performance: less data is read from disk depending on the
filter.
* Improved diagnostics: duplicate dependency warnings, multi-task
edit failures, changes to tasks without IDs, certificate file sizes.
* Improved testing: migrated all Perl tests to Python, parallelized
test suite, colorized output, detection of newly passing tests,
Python 2.7 and 3 support, better debug output, Bash test library,
stress test tool, no more disabled tests - everything runs, test
coverage is now 87.3%.
* Widespread code cleanup, removal of dead code, C++11 enhancements,
improved portability, merged ISO-8601 and legacy durations, less
code, happier developers.
Built and run locally.
From the Changelog:
```
Version 0.7.78, 2015-10-02
+ MOV: AVrp CodecID support
+ Video Stored_Width/Stored_Height and Sampled_Width/Sampled_Height
added (hidden by default)
+ Speed optimization for the parsing of tiny files e.g. JPEG files
+ Option (command line / DLL) --Legacy=0 for disabling some legacy fields
+ Option (command line / DLL) --Output=MAXML, XML with MediaInfo and
MediaTrace output together, technology preview (not for production)
x MPEG-TS: Teletext and Teletext Subtitle were missing in the menu list
x Chroma subsampling "4:4:4" was sometimes indicated for RGB, which is
not useful
```
Upstream changes to the build system required adjusting many packages'
dependencies. On the Nixpkgs side, we no longer propagate the dependency
on cmake (to reduce closure size), so downstream dependencies had to be
adjusted for most packages that depend on kdelibs.
You can now pass
separateDebugInfo = true;
to mkDerivation. This causes debug info to be separated from ELF
binaries and stored in the "debug" output. The advantage is that it
enables installing lean binaries, while still having the ability to
make sense of core dumps, etc.
