807518 - SpiderMonkey Regression on misc-bugs-659467-jsunzip (oct 29 - oct 30)

Status: RESOLVED INCOMPLETE

(Core :: JavaScript Engine, defect)

Description

[Nicolas B. Pierron [:nbp]]

Attachment: misc-bugs-659467-jsunzip MAC OS X 32bits image

Regression reported by AWFY, from 55ms to 79ms. The regression appears for both JM and JM+Ion.  The regression appeared between changesets 70d9ba6b84f8 and 8bc60d1ab725.

Comment 1

[Nicolas B. Pierron [:nbp]]

This range include the following JS change sets.

* a306ea0502f7d2e01d0950fddde6637bbfa00351 - Bug 806643 - Negate doubles by flipping signed bit. r=sstangl (26 hours ago)
* e7d9b7fadd242e92f3c0a31a9ece2c83c61fee29 - Bug 803707 - Don't assume pointers are 64 bits. r=Waldo. Use reinterpret_cast. (26 hours ago)                        
* e436ff984d566471ee901575f7e8ce469ecd7ee9 - Bug 804064 - TableSwitchV, unbox integers values before branching. r=h4writer (27 hours ago)                             
* b2620ac72e946b964250d40a002e8c6deca9b4a5 - Bug 805080 - Remove unused CESU8 support from SpiderMonkey; r=luke (2 days ago)
* df94ebb4b929618642e05e7d5d7bf7c2016670a6 - Backed out 2 changesets (bug 805883, bug 787703) because of Windows PGO bustage on non-profiling builds (27 hours ago)
* cb997ba8433bf03e4350a294ef7d52f756268d9a - Bug 806644 - Back out 782072 (r=terrence) (2 days ago)

Comment 2

[Nicolas B. Pierron [:nbp]]

Ignoring Ion specific patches, backouts (as I don't see any inverted jumps in the history of AWFY), and small modifications.  The only remaining patch seems to be Bug 805080.

Comment 3

[Terrence Cole [:terrence]]

The attached image appears as a blank white square for me.

Comment 4

[Terrence Cole [:terrence]]

Sorry, I should say rectangle.  The pixel dimensions of the image appear to be correct, it's just blank.