Fix a crash in DeadArgumentTransform

[meg-gupta]

When there are no non-type dead args, DeadArgumentTransform should not
specialize the function. While marking dead type args as not dead, make
sure we go over all the args.
If not, the following incorrect things happen:

• the function will be specialized with just dead type args
• a "dead" type arg with debug uses will be attempted to be deleted in FunctionSignatureTransform::DeadArgumentFinalizeOptimizedFunction.

[meg-gupta]

@swift-ci test

[meg-gupta]

Running benchmarks to check any code size changes

[meg-gupta]

@swift-ci benchmark

[swift-ci]

Performance: -O

Improvement	OLD	NEW	DELTA	RATIO
FloatingPointPrinting_Float_description_small	5616	5184	-7.7%	1.08x (?)

Code size: -O

Performance: -Osize

Improvement	OLD	NEW	DELTA	RATIO
FlattenListLoop	2514	1629	-35.2%	1.54x (?)
FlattenListFlatMap	6852	5385	-21.4%	1.27x (?)
String.data.Medium	112	97	-13.4%	1.15x (?)
NSStringConversion.MutableCopy.Rebridge.Medium	644	600	-6.8%	1.07x (?)

Code size: -Osize

Performance: -Onone

Regression	OLD	NEW	DELTA	RATIO
UTF8Decode_InitFromBytes_ascii_as_ascii	458	519	+13.3%	0.88x (?)
ErrorHandling	4340	4780	+10.1%	0.91x (?)
Dictionary4	989	1065	+7.7%	0.93x (?)
Improvement	OLD	NEW	DELTA	RATIO
CharIndexing_ascii_unicodeScalars_Backwards	378960	333200	-12.1%	1.14x (?)
CharIndexing_japanese_unicodeScalars_Backwards	465440	409520	-12.0%	1.14x (?)
CharIndexing_tweet_unicodeScalars_Backwards	764520	674600	-11.8%	1.13x (?)
CharIndexing_punctuatedJapanese_unicodeScalars_Backwards	66640	58840	-11.7%	1.13x (?)
CharIndexing_chinese_unicodeScalars_Backwards	288520	255680	-11.4%	1.13x (?)
CharIndexing_punctuated_unicodeScalars_Backwards	85040	75720	-11.0%	1.12x (?)
CharIndexing_korean_unicodeScalars_Backwards	374520	333560	-10.9%	1.12x (?)
CharIndexing_russian_unicodeScalars_Backwards	319240	286640	-10.2%	1.11x (?)
CharIndexing_utf16_unicodeScalars_Backwards	373720	343760	-8.0%	1.09x (?)
StringBuilderWithLongSubstring	5510	5070	-8.0%	1.09x (?)
DictionaryOfAnyHashableStrings_lookup	5232	4848	-7.3%	1.08x (?)

Code size: -swiftlibs

How to read the data

The tables contain differences in performance which are larger than 8% and differences in code size which are larger than 1%.

If you see any unexpected regressions, you should consider fixing the
regressions before you merge the PR.

Noise: Sometimes the performance results (not code size!) contain false
alarms. Unexpected regressions which are marked with '(?)' are probably noise.
If you see regressions which you cannot explain you can try to run the
benchmarks again. If regressions still show up, please consult with the
performance team (@eeckstein).

Hardware Overview

  Model Name: Mac Pro
  Model Identifier: MacPro6,1
  Processor Name: 12-Core Intel Xeon E5
  Processor Speed: 2.7 GHz
  Number of Processors: 1
  Total Number of Cores: 12
  L2 Cache (per Core): 256 KB
  L3 Cache: 30 MB
  Memory: 64 GB

[meg-gupta]

@swift-ci test Windows Platform

[atrick]

I don't really understand the surrounding code, but this fix looks great.