C++: Fix join order in 'getConversionType4'

[MathiasVP]

Before:

Tuple counts for Printf::FormatLiteral::getConversionType_dispred#fff/3@10f6e8dt after 90ms:
  ...
  3914   ~2%     {3} r61 = SELECT project#Printf::FormatLiteral::parseConvSpec_dispred#fffffffff#4 ON In.2 != "l"
  3914   ~2%     {3} r62 = SELECT r61 ON In.2 != "w"
  3914   ~2%     {3} r63 = SELECT r62 ON In.2 != "h"
  3914   ~2%     {2} r64 = SCAN r63 OUTPUT In.0 'this', In.1 'n'
  3914   ~2%     {2} r65 = STREAM DEDUP r64
  3914   ~1%     {3} r66 = JOIN r65 WITH Printf::FormatLiteral::getDefaultCharType_dispred#ff ON FIRST 1 OUTPUT Rhs.1 'result', Lhs.0 'this', Lhs.1 'n'
  900220 ~0%     {3} r67 = JOIN r66 WITH Type::DerivedType::getBaseType_dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Rhs.1 'result', Lhs.1 'this', Lhs.2 'n'
  3914   ~1%     {3} r68 = JOIN r67 WITH Type::PointerType#f ON FIRST 1 OUTPUT Lhs.1 'this', Lhs.2 'n', Lhs.0 'result'
  ...

After:

Tuple counts for Printf::FormatLiteral::getConversionType_dispred#fff/3@06dbbet0 after 9ms:
  ...
  3914  ~2%     {3} r44 = SELECT project#Printf::FormatLiteral::parseConvSpec_dispred#fffffffff#4 ON In.2 != "l"
  3914  ~2%     {3} r45 = SELECT r44 ON In.2 != "w"
  3914  ~2%     {3} r46 = SELECT r45 ON In.2 != "h"
  3914  ~2%     {2} r47 = SCAN r46 OUTPUT In.0 'this', In.1 'n'
  3914  ~2%     {2} r48 = STREAM DEDUP r47
  3914  ~1%     {3} r49 = JOIN r48 WITH Printf::FormatLiteral::getDefaultCharType_dispred#ff ON FIRST 1 OUTPUT Rhs.1 'result', Lhs.0 'this', Lhs.1 'n'
  3914  ~1%     {3} r50 = JOIN r49 WITH Printf::isPointerTypeWithBase#ff ON FIRST 1 OUTPUT Lhs.1 'this', Lhs.2 'n', Rhs.1 'result'
  ...

Notice that the final tuple count (from this disjunction) doesn't change.

[MathiasVP]

Thanks for approving it so efficiently, @aschackmull! I found a copy/paste error just after you approved it, though :( I force-pushed it away, but I had written:

isPointerTypeWithBase(this.getNonDefaultCharType(), result.(PointerType).getBaseType())

where I meant to write

isPointerTypeWithBase(this.getNonDefaultCharType(), result)

[redsun82]

LGTM! Let's see if DCA agrees 🙂

[redsun82]

Still LGTM! 🙂