======================================================================== http://mondrian.corp.google.com/file/8186649///depot/googleclient/gears/opensource/gears/base/common/js_types.cc?a=1 File //depot/googleclient/gears/opensource/gears/base/common/js_types.cc (snapshot 1) ------------------------------------ Line 2395: bool JsCallContext::GetArgumentAsBool(int i, On Mon Sep 8 11:31:51 2008 PDT, mpcomplete wrote: > We should probably do range checking on the 'i' passed to GetArgument(), since > these are public APIs. Either: > - check i < GetArgumentCount() in each of these methods, or > - check i inside GetArgument(), and return a null token if it's out of range?
Duh, you're absolutely right. I've gone with the first option that you've suggested. ======================================================================== http://mondrian.corp.google.com/file/8186649///depot/googleclient/gears/opensource/gears/base/common/js_types.h?a=1 File //depot/googleclient/gears/opensource/gears/base/common/js_types.h (snapshot 1) ------------------------------------ Line 353: // TODO(mpcomplete): Figure out if context_ is necessary on BROWSER_NPAPI. On Mon Sep 8 11:22:11 2008 PDT, mpcomplete wrote: > I think this TODO can go away (even if it isn't necessary, we have to leave it > in for the other ports). Done. ------------------------------------ Line 395: bool coerce=false) const; On Mon Sep 8 11:23:31 2008 PDT, mpcomplete wrote: > Google style doesn't like default params. But if this is just a temporary > stopgap, it's fine to leave it for now. Rewrote to make the coercion explicit in the method name. ------------------------------------ Line 544: bool GetArgumentAsBool(int i, bool *out, bool coerce=false); On Mon Sep 8 11:24:53 2008 PDT, mpcomplete wrote: > same here re: default param. This one is a stopgap. I intend to remove the coercion option. ======================================================================== -- To respond, reply to this email or visit http://mondrian.corp.google.com/8186649
