Hi Ed.
Interpreters, ExecStates, Binding
-
- ConvertValueToNPVariant() and convertValueToObjcValue(), which are
used in the plugin interface binding code, both contain special logic
that basically says if wrapping up a window object, store the
RootObject
Dan,
Do you really need to customize these things through the editing
delegate, or do you just need WebKit to fix them in its own editing
code? Most of your comments suggest that the latter is the case.
Geoff
On Dec 21, 2006, at 11:14 AM, Dan Wood wrote:
Hi folks,
I've been working
The buildbot is red.
3 JavaScriptCore tests are failing, presumably due to the DST change.
Kevin McCullough is working on this.
7 layout tests began failing with r20130 or r20131.
50-60 new leaks showed up with r20129, r20130, or r20131.
These are the three check-ins in question:
Changed
+1
On May 7, 2007, at 11:28 AM, Darin Adler wrote:
I was thinking we should change the tool's name from DumpRenderTree
to RunWebKitTest.
What do you all think?
-- Darin
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
I believe we should add this one to the list: rdar://problem/
5198885 Crash or ASSERT below FrameLoader::commitProvisionalLoad
when navigating away from bookmarks view
Geoff
On May 12, 2007, at 1:14 AM, Maciej Stachowiak wrote:
My change to make Frames start with an empty document instead
Does anyone know what class/function I should use in WebKit to execute
a JavaScript function from C++?
KJS::JSObject::call invokes a function object as a function. You will
need to retrieve the appendMessage function object from the global
object in order to call it.
If you're asking
?
Thanks, Oleg.
Geoffrey Garen wrote:
I'm investigating possibility to use WebKit as a embedded html
renderer, but I've failed to find any documentation on API I could
use for this :(
http://developer.apple.com/documentation/Cocoa/Conceptual/DisplayWebContent/index.html#/
/apple_ref/doc/uid
I believe that testcase tests for a security problem that can result
from NULL characters. And it looks like your anti-virus software
detected the same problem.
Geoff
On Jul 17, 2007, at 10:53 AM, Yuehfu Shih wrote:
Hi,
I am new to Webkit. I am downloading WebKit source code using
setContentRect: is only called when opening a new window through
JavaScript, and only with Safari 2.
In general, these methods are not notifications -- they are requests.
So, for example, a user resize of the window may change the WebView's
contentRect but, because the user -- not WebKit
WebKit is not thread-safe. If you want to perform external processing
on a secondary thread, that's OK, but any calls into the DOM will have
to happen on the main thread.
Geoff
On Aug 10, 2007, at 3:00 AM, Julien Chaffraix wrote:
Hi all,
I am trying to make a wrapper in C++ for the Ivy
Hi Sajesh.
The debug menu is a Safari feature, not a WebKit feature. You can add
a menu to your app using Interface Builder.
Geoff
On Sep 7, 2007, at 3:29 PM, sajesh wrote:
Hi All,
I am new to web-kit, i am trying to customize the debug menu of web-
kit to add my own custom menu for
Having this document as a reference will be great!
History
Many objects in WebKit are reference counted. The pattern used is
that classes have member functionsref and deref that increment and
decrement the reference count. Each call to ref has to be matched by
a call to deref. When the
Mainly I'd like to hear if any major contributors would have a
problem with this move.
I guess I'll sound like a bit of a shill here, but:
I contributed to the WTF code and I wouldn't mind moving it to a BSD
license.
Cheers,
Geoff
___
webkit-dev
Hi Michael.
One solution would be to maintain a hash or bitmap indicating which
lazy properties you had allocated so far:
JSValueRef GetProperty(JSContextRef ctx, JSObjectRef obj, JSStringRef
prop,
JSValueRef* exception)
{
if (!bitmapContains(prop)) {
JSValueRef lazy_loaded =
, Geoffrey Garen [EMAIL PROTECTED] wrote:
Hi Michael.
One solution would be to maintain a hash or bitmap indicating which
lazy properties you had allocated so far:
JSValueRef GetProperty(JSContextRef ctx, JSObjectRef obj, JSStringRef
prop,
JSValueRef* exception)
{
if (!bitmapContains(prop
Hi Akos.
Any help speeding up JavaScriptCore is welcome.
The current plan is:
1. Convert from an AST execution engine to a custom bytecode engine
2. Perform low-cost optimizations on the custom bytecode (e.g., simple
type inference, peephole optimization, constant folding, copy
propagation,
Hi Akos.
Hi Geoff,
thanks for the info. It seems to me that your answers cancelled my
options. :) Optimizing the AST is not the best way to go since
execution moves to bytecode. And designing a new bytecode is not an
option anymore since you are already doing it. :)
Sorry to spoil
In seeing the last project listed, C Language Binding I was
reminded of the JavaScript Core API that was released in Leopard.
Is the list of projects at that site out-of-date relative to the
latest releases of WebKit?
Yes, it's a little out of date.
Geoff
Which binding should I use that would fit the architecture best?
Must I understand JavaScriptCore/API in order to add the object?
What is the life cycle of the Javascript objects?
Yes, I think using the API in JavaScriptCore/API will give you the
best results. It's also the easiest API to
Is there a way to set a timeout using this API or kill / interrupt a
running wild script?
No.
Geoff
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo/webkit-dev
I want to know if it's possible to create a click event on a tag
within
a page without using javascript. I know the tag name.
I assume you mean without using JavaScript or the mouse.
Depending on the type of application you're writing, you could use the
Objective-C API to the DOM.
Is there a way to get this Document object from a WebView object or
a WebFrame object?
No. The C++ components of WebCore are not API.
Geoff
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
On Jun 12, 2008, at 11:25 PM, Maciej Stachowiak wrote:
Now that SquirrelFish has landed, I'd like to take a look at
reorganizing JavaScriptCore's files and directories a bit.
JavaScriptCore has the following subdirectories:
API/
VM/
docs/
icu/
kjs/
os-win32/
pcre/
profiler/
tests/
Do these also assume that we'd change the directory layout to match
the SVN project folder layout? (That seems like a reasonably good
starting point.)
Yeah.
Geoff
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
These look good to me.
Cut down on confusing uses of Object and Imp.
Should we add the JS prefix to these, too?
Geoff
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
I like the idea. But I'm not crazy about the three WebKit namespaces
being WTF, JSC, and WebCore. One of these things is not like the
others!
How about JSCore? We already use JS as an abbreviation for
JavaScript all over the place, so JSCore seems like a natural
shortening of
I'm not sure where to put these:
ExecState.cpp (runtime or vm?)
ExecState.h
I like vm. Mostly, an ExecState represents the state of execution in
the VM. The runtime relies on the ExecState, but I wouldn't say that
the ExecState was a component of the runtime.
completion.h
test.js
I think you attached the wrong file. All I see is comments.
Geoff
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
I'm looking for a way to measure the stack usage of webkit
especially then how the distribution of different size of allocated
blocks. Is there any developed method yet ?
Have looked at export MallocStackLogging=YES. Could I get a log to a
text file where allocation and deallocation are
With MallocStackLogging enabled, you can use the malloc_history tool.
You can also use the ObjectAlloc Instrument in Instruments.app.
Geoff
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
You might also want to read http://trac.webkit.org/wiki/Memory%20Use,
although that's mostly about the Mac and Windows ports.
Yes, MallocStackLogging and Instruments are Mac OS X only. umdh.exe is
Windows only.
Geoff
___
webkit-dev mailing list
After rendering about 200 pages on my platform port of WebKit, I
seem to be seeing some memory leaks involving the code in KJS,
totalling 30-70MB. I have included a subgraph of the call graph for
the leaking functions. I'm not familiar with the KJS code. Is it
possible that circular
Or there is option 3:
3) Restore the clamp for setTimeout and setInterval to 10ms for
compatibility, and add a new setHighResTimer API that does not have
any lower bound.
I'd like to tweak this suggestion a bit:
Let's make this new timer API object-oriented, so it can be both less
In the Chrome tree, every object inheriting from RefCounted incurs
an extra pointer in size, but this is clearly more than necessary
since many RefCounted objects do not have bindings.
If we believe that JS wrappers are relatively uncommon, we can store
them in a Node's rare data
If we believe that JS wrappers are relatively uncommon, we can
store them in a Node's rare data structure, and bloat only those
uncommon nodes that have JS wrappers.
Depending on exactly how common they are, this could be more net
memory use, if it causes Nodes to have a NodeRareData
I would argue for enum Color { BrownColor };
I believe (based on recent patch reviews) that we have undocumented
consensus that this style of first-letter-capitalized CamelCase is
preferred for enums. I tend to agree with Peter that the Color
suffix is unnecessary, but I'm not sure if
One thing I'd like to highlight: It is a requirement for Chromium
to use consistent URL parsing throughout the entire application.
Can you explain this requirement more?
I think that's an important data point, since other WebKit ports,
including the Mac port, use KURL in WebKit and a
KURL depends on WebCore types, and was
found to have bugs, so Brett did a study of other URL parsers and
wrote the Google URL library as a template library (and wrote GURL as
an example class using it with std::string).
This part I don't get. Brett found some bugs. Why didn't he fix them?
Again, I'm wondering how many legitimate uses are there for short
timeouts in background tabs/windows.
In a background window:
animation
video
audio
work queues for database or other background processing
something interesting the web hasn't invented yet
To give you some context, Safari used
Hi Chris.
I really like the idea of a Timer object. It would allow you to
separate creation from starting, allows you to pause and add other
API's to the interface. Can the constructor be used to simplify the
creation:
var t = new Timer(0, false, function() { ...});
which would start
I seen also this in cache :
~Cache(); // Not implemented to make sure nobody accidentally calls
delete --
WebCore does not delete singletons.
it's also to make shutdown faster ? this can create a big leak, no ?
Destruction of items in the cache is not tied to the lifetime of the
cache.
Thank you, Dimitri. I found http://build.chromium.org/merge/ very
informative.
Cheers,
Geoff___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
Now should I file a new bug on https://bugs.webkit.org/ for
improving markup, then write the patch and send to you (plus
someone else I should involve) for review? Is that a right process?
Yes, that's right. But you shouldn't target the patches at a
specific person for review unless it's a
Hi François.
What are Protected Objects?
Protected objects are objects that have no direct references from the
stack or the JavaScript heap, but must not be destroyed by the garbage
collector. Typically, they are objects that are referenced by non-GC
heap-allocated storage, like the
Hi Dave.
We tend to write unit tests in HTML or JavaScript. If it's possible to
trigger all of the relevant test conditions in HTML or JavaScript, I'd
recommend that.
The next best thing to do is to write a C++ test function, add it to
WebKitTools/DumpRenderTree, and invoke it during
I was about to upstream some patches that add #ifdefs to the WebCore
IDL files. In many cases, the only delta is the removal of the
Custom attribute. Eric Seidel pointed out that it might be better
to rename those usages of Custom to JSCCustom to avoid #ifdefs.
This change would avoid
Hi Laurent.
The API shouldn't have any subtle dependency on
kJSClassAttributeNoAutomaticPrototype. Please file a bug report @
bugs.webkit.or, and attach a source file that we can compile in order
to reproduce this bug.
Thanks,
Geoff
On Feb 4, 2009, at 1:55 PM, Laurent Calburtin wrote:
Hi.
You can use 'make' or any of the platform-specific project files in
the JavaScriptCore subdirectory.
Geoff
On Feb 19, 2009, at 11:52 PM, thouraya andolsi wrote:
Hi all,
How can i compile JSC standalone ? ( I would like to have just the
jsc executable)
I looked at
I would like to know how to parse the js code through the function of
webkit.
Could you tell me the main functions and the files?
JavaScriptCore/parser/Lexer.h
JavaScriptCore/parser/Lexer.cpp
JavaScriptCore/parser/Parser.h
JavaScriptCore/parser/Parser.cpp
Cheers,
Geoff
Hi Zhe.
I'm developing an application which uses webkit's JavaScriptCore
and a customized global object (by providing a special class when
creating the context). My customized global object provides some
built-in properties that can be accessed by javascript code. Now, I
want to
I'm writing a C++ to JavaScript binding based on JavaScriptCore,
and would like to know current source URL and line number
information in the callback functions. However I didn't find any
related API. It there any way to do it?
No.
Sounds like a reasonable feature request.
Geoff
Hi Ross.
While I can grab the JSGlobalContext reference from a WebView's
mainFrame,
and subsequently use the public JS APIs to extend the underlying JS
global
object, I believe that subsequent WebView allocations will cause the
underlying JSGlobalObject to be reinitialized forcing me to
This code was tested against webkit svn. Anyone can help me explain
this behavior? Is it a bug of JavaScriptCore?
Most likely, the conservative nature of the garbage collector prevents
the function from being reclaimed in this case.
You can step through JSC::Heap::collect() to see exactly
I was wondering whether it is possible to remove the constant pool
area in
jit.
It is.
RegisterFile contents right now:
... [ CallFrame ] [ Cleared const pool ] [ temporary SF
registers ] ...
(SF means SquirellFish)
The constant pool is only used by interpreter, the jit simply clears
-EventListener* Node::getInlineEventListener(const AtomicString
eventType) const
+EventListener* Node::getAttributeEventListener(const AtomicString
eventType) const
I think that per our coding style guidelines, this function should
be named attributeEventListener():
6. Precede setters
Right now, the branch is about even with trunk on SunSpider.
Geoff
On May 5, 2009, at 6:11 AM, Zoltan Herczeg wrote:
Hi,
we have noticed that there is a new SquirellFish branch, called
Nitro-Extreme. You mentioned before (perhaps a year ago), that you
plan to
change the JSNumber format to
1) Does it make sense to enable ENABLE_YARR by default and use it even
when we have no JIT support.
No.
2) Does the YARR JIT depend on the JSCore JIT, or is it a separate
JIT.
Not entirely.
If so, does it work on the same platforms at the JSCore JIT?
What problems are those?
Geoff
If so, does it work on the same platforms at the JSCore JIT?
What problems are those?
Oops!
platforms
Yes, it works on the same platforms as the JSCore JIT.
Geoff
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
Hi Akos.
Today, we realized that there is a new ARM JIT port for WebKit. (http://trac.webkit.org/changeset/44514
) Congratulations on getting this working!, great job.
Thanks.
I cannot conceal how disappointed I am, as is the whole team at
Szeged.
I'm sorry to hear that. I understand
This expands out to 95 inline instructions on the MIPS for just the
slow case alone, of which 3 are functions calls to other functions.
So this probably requires thousands of clock cycles to execute.
IMHO it doesn't make sense to inline op_call because:
You've made some interesting
It could be worth trying a stub function that triggers the
compilation of the function should it not be present, but i'm not
sure what that would really save as we still need the arity checks
inline
A design that I like is a stub function that triggers compilation (so
the caller can
Hi Sebastian.
JSEvaluateScript can only evaluate JavaScript; it can't parse HTML.
Cheers,
Geoff
On Jun 10, 2009, at 10:05 PM, Sebastian Linke wrote:
Hi,
based on my experiences when I ran some javascript code on a webkit
context,
I was trying to do the same with normal html code.
It may
Can the arity check be performed at compile time as in C++?
C++ can perform arity checks at compile time because C++ uses early
binding. JavaScript uses late binding.
Geoff___
webkit-dev mailing list
webkit-dev@lists.webkit.org
it looks we are in the way of the train. You have plans, we don't know
about them, you have commit rights, we don't, so the tides are
against us.
If you're interested in review or commit rights, they're granted based
on a track record of good work, good judgement, and good
collaboration.
The Popup blocking feature, is that part of Webkit code space?
It's mostly in WebCore (a component of WebKit).
See DOMWindow::allowPopUp().
Geoff
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
Is there a straigthforwarded way to collect all the relationships
between the
site's object and to get a js-context of it?
What I want in the end is to get access on the `document`-element,
e.g. to
perform calls like `document.forms`.
If you have a JSContextRef that corresponds to a
Function objects have a property named name, and also a property
named displayName.
Geoff
On Jun 16, 2009, at 3:55 AM, Zoltan Herczeg wrote:
Hi,
how can I extract the name of a JS function in JIT.cpp?
Zoltan
___
webkit-dev mailing list
If someone could be kind enough to explain how it works on other
platforms, it would be very helpful.
I'm not really interested in helping with the dubiously legal task of
taking JavaScriptCore, which is an OpenSource library, and turning it
into a closed, proprietary library.
Ditto. The last time I wondered about this, I grepped through the
code and found [1] to be the most prevalent. So, in code reviews I
have been recommending people do [1].
I think that's the nicest style (just like a variable, no special
prefix). Let's go with that and update the coding
1. Is it correct for the ExecState to carry the thisValue?
As Adam realized later (I think), ExecState carries the value for
this inside the calling function. It does not carry the object whose
property is being accessed.
Geoff
___
webkit-dev
Our current thinking is to add a parameter to toJS() to receive the
JSGlobalObject in which to create the wrapper.
Seems logical.
Once we do that, the question is how to find the proper
JSGlobalObject at each call site.
In most cases, you have another JSObject sitting around that is from
Is it definitely right for document.body to make a wrapper using
prototypes
from the document's host window, rather than from the accessing
function's
window? What do other browsers do?
That's correct. Other browser's get this case right.
Is there a particular security or other benefit
I'm first fixing JSCell::new subclasses to make sure they're always
allocating in the correct heap. If we're to map from objects to the
associated globalobject via the heap, we need to fix allocation
first.
I'm not sure what you guys have been meaning by the phrase correct
heap. Barring
1. Pass a current global object through to all toJS calls (lots of
callsites changed)
2. Store a current global object off of the ExecState (set by the
JS engine before leaving into custom native code for property lookup
or function execution).
I discussed this a bit with Darin and Geoff, and
I discussed this a bit with Darin and Geoff, and we came to the
conclusion that the correct fix is to have each JS DOMObject store
a JSGlobalObject pointer and augment the toJS methods to pass a
global object instead of an ExecState (close to you #1).
You probably mean in addition to
I discussed this a bit with Darin and Geoff, and we came to the
conclusion
that the correct fix is to have each JS DOMObject store a
JSGlobalObject
pointer and augment the toJS methods to pass a global object
instead of an
ExecState (close to you #1).
There are classes in JavaScriptCore
Is there a particular security or other benefit here, or do we just
want to
make this change to match other browsers?
Our current behavior is buggy, unpredictable, and out of spec. This
has led to security bugs in the past and will lead to security bugs in
the future.
I don't disagree with
I discussed this a bit with Darin and Geoff, and we came to the
conclusion that the correct fix is to have each JS DOMObject store
a JSGlobalObject pointer and augment the toJS methods to pass a
global object instead of an ExecState (close to you #1). I would
not advocate storing more
Which spec did you have in mind? I'd like to read it.
Essentially, the ECMAScript spec requires this. In spec-land, these
objects are all created at the beginning of time. The fact that we
create them lazily is what leads to this bug. Depending on who
touches them first, they end up with
Yep... My guess is that (a) is probably better, but you two are more
expert on why we use lazy construction.
Heh -- lazy construction in JSC is even older than I am! :)
The goal is to reduce memory use and startup cost in cases where
certain functions aren't used. I'm not sure how much of a
Consider this case, which does not involve a DOM object:
frames[0].Array.prototype.push.__proto__ ==
Array.prototype.push.__proto__
Built-in classes work somewhat differently. I believe they use the
calling function's global object (lexical global object) rather
than having some notion
That's correct. Other browser's get this case right. Here are a
couple test cases you might find interesting:
http://webblaze.org/abarth/tests/protoconfused/test1.html
http://webblaze.org/abarth/tests/protoconfused/test2.html
I tried these tests, with mixed results:
IE8: Exception thrown
Also, once we've established the model, we'll need to propose it to
some
standards body -- probably HTML5.
I believe the correct spec to describe this behavior is WebIDL, which
controls how the abstract DOM interfaces are realized in ECMAScript.
Sounds good.
Geoff
Hi Brian.
I don't understand the distinction you're drawing between the
property level and the object level. Can you explain what those
mean and give an example of each?
Thanks,
Geoff
On Jul 20, 2009, at 9:14 AM, Brian Barnes wrote:
I was getting ready to try the first move from
(green)
if (name == 'blue') return(blue)
}
And property level, we'd have 3 getters:
void getMyObjectRed(...)
{
return(red);
}
void getMyObjectGreen(...)
{
return(green);
}
void getMyObjectBlue(...)
{
return(blue);
}
[] Brian
Geoffrey Garen wrote:
Hi Brian.
I don't understand the distinction
I'm not sure you get exactly what I'm saying as you put class in
places where I'd expect property. For instance:
I looked briefly, and it seems like it would be relatively easy to
add an API for adding C getters and setters to a class individually.
So, that seems like a reasonable
How about separate call backs at the class level? That would solve
my problem with minimal code movement. Something like:
JSClassSetPropertyGetterSetter
(ctx,class,red,myRedGetter,myRedSetter);
Would that be more within the design?
Yes. That's what I had in mind when I mentioned an API
Sure!
Geoff
On Jul 20, 2009, at 3:51 PM, Brian Barnes wrote:
What should be my next step, here? Submit a bug for this?
[] Brian
On Jul 20, 2009, at 6:05 PM, Geoffrey Garen wrote:
How about separate call backs at the class level? That would
solve my problem with minimal code movement
Am I missing something? Is there some real benefit to using types
like int32_t instead of int in some places in our code and not
others? If so, what are these critical places and what makes them
different from the rest of the code?
In JavaScriptCore, some structures have integer members
In JavaScriptCore, some structures have integer members that must be
32bits in size, regardless of processor type. In those places, int32_t
and uint32_t are useful.
Less clear to me is whether clients of such structures should also use
int32_t / uint32_t. For example:
struct {
int32_t i;
Hi.
r45939 broke my workflow. Here's the related bugzilla bug: https://bugs.webkit.org/show_bug.cgi?id=26999
.
Old Roll out a patch workflow:
cd JavaScriptCore
svn-create-patch patch.txt
svn-unapply patch.txt
Old Roll in a patch workflow:
cd JavaScriptCore
svn-apply patch.txt
These old
Detailed descriptions, bug links, test instructions, and a link back
to the entire original review history are all part of Chromium
commits, yet we don't use ChangeLogs. I think discipline about what
to include + tooling to support it are orthogonal to a project's use
of a ChangeLog as
Hi Simon.
What about functionality where the C API would slow down the C++ API
but the
internal JSC API is stable enough/good enough?
If there are such areas, and they turn out to be substantial, we
should look at speeding them up, for the benefit of all clients of the
C API.
I can
Hi Zoltan.
JSWrapperObject::markChildren is responsible for marking the
internalValue of a DateInstance. Is that function not being called?
Geoff
On Sep 1, 2009, at 6:16 AM, Zoltan Herczeg wrote:
Hi Oliver,
it seems on ARM using WTF_USE_JSVALUE32, the internal value of a date
object is
Comments should look like sentences by beginning with a capital and
ending with a period (punctation).
I think this is a generally good recommendation, but sometimes a
sentence fragment makes for a better comment, e.g.:
if (x == y) // false for NaN
Don't add explicit line breaks in the
The better option, as Adam noted on the bug, is one in which the JS
engine itself measures how much work has been performed. A count of
instructions or function calls is a truer measure than the proposed
count of time checks, but imposes costs on the JS engine that the V8
team have argued
* I notice that even when I set the WebCore::Cache capacity to zero,
I can't necessarily dump _everything_ out of it. Is there some
other set of calls I should make to drop more references somewhere?
Cache::setCapacities() calls Cache::prune(), which should remove all
dead items.
Live
Live items cannot be removed from the cache.
Yeah, I think that is what I'm seeing.
Assume the page is not being painted (e.g. the window is
minimized). Is there a way to turn the live items into dead ones
so they can be flushed? Obviously the instant we repaint we will
have to obtain
I'm convinced now, too. Let's disable it.
Geoff
On Oct 5, 2009, at 12:53 PM, Maciej Stachowiak wrote:
On Oct 5, 2009, at 12:23 PM, Eric Seidel wrote:
It seems that the requestee field is a source of confusion for new
contributers. Especially so when the new contributor comes from
Hi Zoltan.
Sounds like a bug.
(I tend to dislike #ifdefs like JIT_OPTIMIZE_CALL because of their
tendency to fall prey to bit rot like this.)
Loading the callee into regT1, regT0 is best. I'd recommend changing !
JIT_OPTIMIZE_CALL to respect that convention.
Geoff
On Oct 7, 2009, at
1 - 100 of 380 matches
Mail list logo