Support for JSDoc in services

[ghost]

Fixes #2452 and #14397
Does not fix #15855, #15853, or #13371

Notes:

• We now always bind jsdoc, even in typescript files. This doesn't necessarily mean that we'll get a good Symbol, but at least parent pointers will be set. I'll have to test how this branch affects performance.

• forEachChild is now typed as providing a NodeArray<Node> to cbNodeArray instead of a Node[]. We already had casts to this effect.

• .getChildren() on a JSDocComment node used to provide a garbage array of "tokens" as if the content was TypeScript code. Now avoid collecting tokens at all and just have it be the array of nodes from forEachChild.

[mhegazy]

We now always bind jsdoc, even in typescript files. This doesn't necessarily mean that we'll get a good Symbol, but at least parent pointers will be set. I'll have to test how this branch affects performance.

Why do we need to do this all the time? can we just do the resolution on demand when an operation is needed in a jsdoc?

can you please share the perf numbers.

[mhegazy]

should not this one be marked as well?

[ghost]

D'oh, shouldn't have left this in. We don't support goto-implementation for parameters anyway.

[ghost]

The current baselines show no significant change.

Monaco - node (v7.10.0, x64) |

Project	Baseline	Current	Delta	Best	Worst
Memory used	359,036k (± 0.00%)	359,062k (± 0.01%)	+25k (+ 0.01%)	359,038k	359,087k
Parse Time	2.00s (± 0.67%)	2.00s (± 0.95%)	+0.00s (+ 0.08%)	1.98s	2.02s
Bind Time	0.80s (± 2.27%)	0.80s (± 2.13%)	+0.00s (+ 0.31%)	0.78s	0.82s
Check Time	3.99s (± 1.83%)	3.97s (± 0.88%)	-0.02s (- 0.44%)	3.93s	3.99s
Emit Time	2.71s (± 1.41%)	2.65s (± 0.57%)	-0.07s (- 2.43%)	2.63s	2.66s
Total Time	9.49s (± 1.22%)	9.41s (± 0.59%)	-0.08s (- 0.84%)	9.37s	9.47s

TFS - node (v7.10.0, x64)

Project	Baseline	Current	Delta	Best	Worst
Memory used	312,738k (± 0.01%)	312,728k (± 0.03%)	-9k (- 0.00%)	312,632k	312,795k
Parse Time	1.45s (± 0.34%)	1.46s (± 0.96%)	+0.01s (+ 0.62%)	1.44s	1.47s
Bind Time	0.74s (± 1.12%)	0.74s (± 1.76%)	+0.00s (+ 0.54%)	0.73s	0.76s
Check Time	3.40s (± 0.43%)	3.44s (± 0.99%)	+0.03s (+ 0.98%)	3.39s	3.47s
Emit Time	2.41s (± 0.61%)	2.39s (± 1.11%)	-0.02s (- 0.73%)	2.37s	2.43s
Total Time	8.01s (± 0.25%)	8.03s (± 0.76%)	+0.03s (+ 0.31%)	7.95s	8.08s

This may just be because our test cases don't use JSDoc very much, although that is probably representative of real code.
If we want to do #13371 we will probably want to bind the JSDoc anyway. The problem is that trying to set parent pointers on demand would lead to a lot of extra code every time we traverse down a tree, and we'd also likely miss some and get lots of NPEs in services, where we generally always assume a parent pointer to be set.

[mhegazy]

if so, then you want to make sure that all the special handling of jsdoc tags in the binder is limited to js files, e.g. @typedef declaring a type, or @property and @type generating symbols.

[mhegazy]

Please run the perf tests one more time. the binder has been historically sensitive to reorganizing the bind function

[ghost]

OK, looks like performance is still fine.

Monaco - node (v7.10.0, x64)

Project	Baseline	Current	Delta	Best	Worst
Memory used	359,897k (± 0.00%)	359,932k (± 0.00%)	+35k (+ 0.01%)	359,910k	359,966k
Parse Time	2.00s (± 0.37%)	2.01s (± 0.77%)	+0.01s (+ 0.65%)	1.98s	2.05s
Bind Time	0.79s (± 0.52%)	0.80s (± 0.75%)	+0.01s (+ 0.80%)	0.78s	0.81s
Check Time	3.96s (± 0.35%)	3.98s (± 0.29%)	+0.02s (+ 0.61%)	3.96s	4.00s
Emit Time	2.70s (± 0.54%)	2.70s (± 0.41%)	+0.00s (+ 0.12%)	2.68s	2.72s
Total Time	9.44s (± 0.18%)	9.49s (± 0.30%)	+0.05s (+ 0.51%)	9.41s	9.54s

TFS - node (v7.10.0, x64)

Project	Baseline	Current	Delta	Best	Worst
Memory used	313,022k (± 0.01%)	313,056k (± 0.01%)	+34k (+ 0.01%)	312,935k	313,091k
Parse Time	1.48s (± 2.95%)	1.44s (± 1.00%)	-0.04s (- 2.52%)	1.39s	1.46s
Bind Time	0.77s (± 3.43%)	0.77s (± 2.97%)	-0.00s (- 0.13%)	0.73s	0.84s
Check Time	3.50s (± 1.63%)	3.44s (± 0.42%)	-0.06s (- 1.82%)	3.42s	3.48s
Emit Time	2.41s (± 1.96%)	2.35s (± 0.69%)	-0.06s (- 2.50%)	2.32s	2.39s
Total Time	8.17s (± 2.01%)	8.00s (± 0.25%)	-0.17s (- 2.02%)	7.96s	8.04s