GM: stop recording all *_pdf.png results as "INVALID" on non-Mac platforms

gm/tests/outputs/add-config-pdf/output-expected/stdout

Index: gm/tests/outputs/add-config-pdf/output-expected/stdout
===================================================================
--- gm/tests/outputs/add-config-pdf/output-expected/stdout	(revision 0)
+++ gm/tests/outputs/add-config-pdf/output-expected/stdout	(revision 0)
@@ -0,0 +1,17 @@
+GM: These configs will be run: 8888 565 pdf
+GM: reading expectations from JSON summary file gm/tests/inputs/json/identical-bytes.json
+GM: writing to gm/tests/outputs/add-config-pdf/output-actual/writePath
+GM: writing mismatches to gm/tests/outputs/add-config-pdf/output-actual/mismatchPath
+GM: drawing... selftest1 [300 200]
+GM: Ran 1 GMs
+GM: ... over  3 configs ["8888", "565", "pdf"]
+GM: ...  and  7 modes   ["pipe", "pipe cross-process", "pipe cross-process, shared address", "replay", "rtree", "serialize", "tilegrid"]
+GM: ... so there should be a total of 10 tests.
+GM: Ran 10 tests: NoGpuContext=0 IntentionallySkipped=1 RenderModeMismatch=0 ExpectationsMismatch=0 MissingExpectations=0 WritingReferenceImage=0
+GM: [*] 0 NoGpuContext:
+GM: [ ] 1 IntentionallySkipped: pdf/selftest1

[epoger, 2013/06/18 05:20:10]

Even if we HAD image expectations for the PDF version, we wouldn't have been
able to compare them (because this platform doesn't know how to convert PDFs to
SkBitmaps).  So IntentionallySkipped makes more sense than MissingExpectations.

+GM: [*] 0 RenderModeMismatch:
+GM: [*] 0 ExpectationsMismatch:
+GM: [ ] 0 MissingExpectations:
+GM: [*] 0 WritingReferenceImage:
+GM: (results marked with [*] will cause nonzero return value)