The Trunk: System-ul.937.mcz

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

The Trunk: System-ul.937.mcz

commits-2
Levente Uzonyi uploaded a new version of System to project The Trunk:
http://source.squeak.org/trunk/System-ul.937.mcz

==================== Summary ====================

Name: System-ul.937
Author: ul
Time: 20 March 2017, 12:35:03.813931 am
UUID: 01691bb8-181f-4795-83ce-20130031a81d
Ancestors: System-ul.936

Fixed SmalltalkImage >> #specialSelectors. SystemTracer doesn't use it any more, but it's being used by tests, which expect it to return only the selectors without their argument count.

=============== Diff against System-ul.936 ===============

Item was changed:
  ----- Method: SmalltalkImage>>specialSelectors (in category 'special objects') -----
  specialSelectors
+
+ | arrayOfPairs |
+ arrayOfPairs := self specialObjectsArray at: 24.
+ ^Array new: arrayOfPairs size // 2 streamContents: [ :stream |
+ arrayOfPairs pairsDo: [ :selector :numArgs |
+ stream nextPut: selector ] ]
+ !
- "Used by SystemTracer only."
-
- ^SpecialSelectors!


Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: The Trunk: System-ul.937.mcz

Eliot Miranda-2
Hi Levente,

    please revert the change that eliminates the argument counts from the specialSelectors array ASAP.  The VM uses these!!  This is causing untold damage in the VM as we speak (for example, it is related to Hernan's crash).  The VM *must* be able to determine the argument count for a special selector it does not inline.

On Sun, Mar 19, 2017 at 4:35 PM, <[hidden email]> wrote:
Levente Uzonyi uploaded a new version of System to project The Trunk:
http://source.squeak.org/trunk/System-ul.937.mcz

==================== Summary ====================

Name: System-ul.937
Author: ul
Time: 20 March 2017, 12:35:03.813931 am
UUID: 01691bb8-181f-4795-83ce-20130031a81d
Ancestors: System-ul.936

Fixed SmalltalkImage >> #specialSelectors. SystemTracer doesn't use it any more, but it's being used by tests, which expect it to return only the selectors without their argument count.

=============== Diff against System-ul.936 ===============

Item was changed:
  ----- Method: SmalltalkImage>>specialSelectors (in category 'special objects') -----
  specialSelectors
+
+       | arrayOfPairs |
+       arrayOfPairs := self specialObjectsArray at: 24.
+       ^Array new: arrayOfPairs size // 2 streamContents: [ :stream |
+               arrayOfPairs pairsDo: [ :selector :numArgs |
+                       stream nextPut: selector ] ]
+ !
-       "Used by SystemTracer only."
-
-       ^SpecialSelectors!





--
_,,,^..^,,,_
best, Eliot


Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: The Trunk: System-ul.937.mcz

Eliot Miranda-2
Hi Levente,

On Tue, Mar 21, 2017 at 9:46 AM, Eliot Miranda <[hidden email]> wrote:
Hi Levente,

    please revert the change that eliminates the argument counts from the specialSelectors array ASAP.  The VM uses these!!  This is causing untold damage in the VM as we speak (for example, it is related to Hernan's crash).  The VM *must* be able to determine the argument count for a special selector it does not inline.

Ah, OK.  It's not creating damage in the VM, but it is creating damage in the simulator.  Forgive me :-).  I overreacted.
 

On Sun, Mar 19, 2017 at 4:35 PM, <[hidden email]> wrote:
Levente Uzonyi uploaded a new version of System to project The Trunk:
http://source.squeak.org/trunk/System-ul.937.mcz

==================== Summary ====================

Name: System-ul.937
Author: ul
Time: 20 March 2017, 12:35:03.813931 am
UUID: 01691bb8-181f-4795-83ce-20130031a81d
Ancestors: System-ul.936

Fixed SmalltalkImage >> #specialSelectors. SystemTracer doesn't use it any more, but it's being used by tests, which expect it to return only the selectors without their argument count.

=============== Diff against System-ul.936 ===============

Item was changed:
  ----- Method: SmalltalkImage>>specialSelectors (in category 'special objects') -----
  specialSelectors
+
+       | arrayOfPairs |
+       arrayOfPairs := self specialObjectsArray at: 24.
+       ^Array new: arrayOfPairs size // 2 streamContents: [ :stream |
+               arrayOfPairs pairsDo: [ :selector :numArgs |
+                       stream nextPut: selector ] ]
+ !
-       "Used by SystemTracer only."
-
-       ^SpecialSelectors!





--
_,,,^..^,,,_
best, Eliot



--
_,,,^..^,,,_
best, Eliot


Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: The Trunk: System-ul.937.mcz

Levente Uzonyi
Hi Eliot,

Anyway, I reverted the method to return the raw array, and introduced
another one to return just the names.
The old implementation could have caused some head-scrathing, because it
just returned a class variable which was not referenced from any other
place, so it was never updated.

Levente

On Tue, 21 Mar 2017, Eliot Miranda wrote:

> Hi Levente,
> On Tue, Mar 21, 2017 at 9:46 AM, Eliot Miranda <[hidden email]> wrote:
>       Hi Levente,
>     please revert the change that eliminates the argument counts from the specialSelectors array ASAP.  The VM uses these!!  This is causing untold damage in the VM as we speak (for example, it is
> related to Hernan's crash).  The VM *must* be able to determine the argument count for a special selector it does not inline.
>
>
> Ah, OK.  It's not creating damage in the VM, but it is creating damage in the simulator.  Forgive me :-).  I overreacted.
>  
>
>       On Sun, Mar 19, 2017 at 4:35 PM, <[hidden email]> wrote:
>             Levente Uzonyi uploaded a new version of System to project The Trunk:
>             http://source.squeak.org/trunk/System-ul.937.mcz
>
>             ==================== Summary ====================
>
>             Name: System-ul.937
>             Author: ul
>             Time: 20 March 2017, 12:35:03.813931 am
>             UUID: 01691bb8-181f-4795-83ce-20130031a81d
>             Ancestors: System-ul.936
>
>             Fixed SmalltalkImage >> #specialSelectors. SystemTracer doesn't use it any more, but it's being used by tests, which expect it to return only the selectors without their argument
>             count.
>
>             =============== Diff against System-ul.936 ===============
>
>             Item was changed:
>               ----- Method: SmalltalkImage>>specialSelectors (in category 'special objects') -----
>               specialSelectors
>             +
>             +       | arrayOfPairs |
>             +       arrayOfPairs := self specialObjectsArray at: 24.
>             +       ^Array new: arrayOfPairs size // 2 streamContents: [ :stream |
>             +               arrayOfPairs pairsDo: [ :selector :numArgs |
>             +                       stream nextPut: selector ] ]
>             + !
>             -       "Used by SystemTracer only."
>             -
>             -       ^SpecialSelectors!
>
>
>
>
>
> --
> _,,,^..^,,,_
> best, Eliot
>
>
>
>
> --
> _,,,^..^,,,_
> best, Eliot
>
>

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: The Trunk: System-ul.937.mcz

Eliot Miranda-2


On Tue, Mar 21, 2017 at 11:03 AM, Levente Uzonyi <[hidden email]> wrote:
Hi Eliot,

Anyway, I reverted the method to return the raw array, and introduced another one to return just the names.
The old implementation could have caused some head-scrathing, because it just returned a class variable which was not referenced from any other place, so it was never updated.

Hmmm, now I have to decide if I should revert ;-)
 


Levente


On Tue, 21 Mar 2017, Eliot Miranda wrote:

Hi Levente,
On Tue, Mar 21, 2017 at 9:46 AM, Eliot Miranda <[hidden email]> wrote:
      Hi Levente,
    please revert the change that eliminates the argument counts from the specialSelectors array ASAP.  The VM uses these!!  This is causing untold damage in the VM as we speak (for example, it is
related to Hernan's crash).  The VM *must* be able to determine the argument count for a special selector it does not inline.


Ah, OK.  It's not creating damage in the VM, but it is creating damage in the simulator.  Forgive me :-).  I overreacted.
 

      On Sun, Mar 19, 2017 at 4:35 PM, <[hidden email]> wrote:
            Levente Uzonyi uploaded a new version of System to project The Trunk:
            http://source.squeak.org/trunk/System-ul.937.mcz

            ==================== Summary ====================

            Name: System-ul.937
            Author: ul
            Time: 20 March 2017, 12:35:03.813931 am
            UUID: 01691bb8-181f-4795-83ce-20130031a81d
            Ancestors: System-ul.936

            Fixed SmalltalkImage >> #specialSelectors. SystemTracer doesn't use it any more, but it's being used by tests, which expect it to return only the selectors without their argument
            count.

            =============== Diff against System-ul.936 ===============

            Item was changed:
              ----- Method: SmalltalkImage>>specialSelectors (in category 'special objects') -----
              specialSelectors
            +
            +       | arrayOfPairs |
            +       arrayOfPairs := self specialObjectsArray at: 24.
            +       ^Array new: arrayOfPairs size // 2 streamContents: [ :stream |
            +               arrayOfPairs pairsDo: [ :selector :numArgs |
            +                       stream nextPut: selector ] ]
            + !
            -       "Used by SystemTracer only."
            -
            -       ^SpecialSelectors!





--
_,,,^..^,,,_
best, Eliot




--
_,,,^..^,,,_
best, Eliot




--
_,,,^..^,,,_
best, Eliot


Loading...