While experimenting with the new llint in JSCore, I noticed that the Not class (located in offlineasm/ast.rb) may have an incorrect method: class Not < Node attr_reader :child def initialize(codeOrigin, child) super(codeOrigin) @child = child end def children [@left, @right] end def mapChildren Not.new(codeOrigin, (yield @child)) end def dump "(not #{child.dump})" end end Shouldn't children of this 'Not' class be def children [@child] end instead, since it's a unary operator?
Created attachment 191731 [details] Fix for the children method I think this is also a problem. So this patch should fix it.
Comment on attachment 191731 [details] Fix for the children method Attachment 191731 [details] did not pass mac-ews (mac): Output: http://webkit-commit-queue.appspot.com/results/17067209 New failing tests: editing/selection/selection-modify-crash.html
(In reply to comment #2) > (From update of attachment 191731 [details]) > Attachment 191731 [details] did not pass mac-ews (mac): > Output: http://webkit-commit-queue.appspot.com/results/17067209 > > New failing tests: > editing/selection/selection-modify-crash.html This patch can not cause this fail, because at this time the 'Not' node is not used at all.
(In reply to comment #3) > (In reply to comment #2) > > (From update of attachment 191731 [details] [details]) > > Attachment 191731 [details] [details] did not pass mac-ews (mac): > > Output: http://webkit-commit-queue.appspot.com/results/17067209 > > > > New failing tests: > > editing/selection/selection-modify-crash.html > > This patch can not cause this fail, because at this time the 'Not' node is not used at all. Agreed.
Comment on attachment 191731 [details] Fix for the children method Clearing flags on attachment: 191731 Committed r145047: <http://trac.webkit.org/changeset/145047>
All reviewed patches have been landed. Closing bug.