Utf32 buffer

[jerch]

This PR applies the UTF32 buffer layout as discussed in #791. This layout rearranges the stored information to make room for the upcoming true color support.

Note: The PR is based on #1796 as it relies on changes from there. After #1796 has landed it can be rebased/merged.

Fixes #791.

[jerch]

Some speed numbers (tested with this xterm-benchmark script https://gist.github.com/jerch/07ea5e2dfc0548df63a9113d6d721aee):

• master

   Context "lib/xterm_perfcases/input.js"
      Context "Terminal: ls -lR /usr/lib"
         Case "#1" : 1 - runtime: 2511.44 ms
         Case "#1" : 1 - throughput: 19.19 MB/s
         Case "#1" : 2 - runtime: 2374.15 ms
         Case "#1" : 2 - throughput: 20.30 MB/s
         Case "#1" : 3 - runtime: 2348.06 ms
         Case "#1" : 3 - throughput: 20.53 MB/s
         Case "#1" : 3 runs - average runtime: 2411.22 ms
         Case "#1" : 3 runs - average throughput: 20.01 MB/s

• this branch

   Context "lib/xterm_perfcases/input.js"
      Context "Terminal: ls -lR /usr/lib"
         Case "#1" : 1 - runtime: 2088.04 ms
         Case "#1" : 1 - throughput: 23.08 MB/s
         Case "#1" : 2 - runtime: 1857.33 ms
         Case "#1" : 2 - throughput: 25.95 MB/s
         Case "#1" : 3 - runtime: 1835.57 ms
         Case "#1" : 3 - throughput: 26.26 MB/s
         Case "#1" : 3 runs - average runtime: 1926.98 ms
         Case "#1" : 3 runs - average throughput: 25.10 MB/s

[juancampa]

This is great! Both in terms of performance and because it unblocks the coming WebGL renderer from customizing selection color, and of course, true color support :)

@jerch Do you have an ETA for this to be ready for review/merge? I can help with reviews, if you guys need an extra (albeit not as knowledgeable in xterm internals) pair of eyes

[jerch]

@juancampa The active codebase should already work with this. The test cases are still missing, they are not yet ported to the new interface (tedious to get those moved over). I plan to get it ready for review by the weekend.

This is planned for the 3.11 release, but we will have to wait for the reflow PR #1864 and PR #1796 as they have interdependencies. After that the road is clear to work directly on true color and UTF8 support.

[jerch]

#1796 got merged, thus the change list here got way shorter. Up for a "normal" review against master.

[Tyriar]

This is only used in tests so I think we should move it to TestUtils.test.ts (and pass in the ICellData as an arg)

[jerch]

Yepp, good idea.

Fixed