[OpenSmalltalk/opensmalltalk-vm] 0222b9: Fix minheadless crash at startup

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

[OpenSmalltalk/opensmalltalk-vm] 0222b9: Fix minheadless crash at startup

Tobias Pape-4
 
  Branch: refs/heads/Cog
  Home:   https://github.com/OpenSmalltalk/opensmalltalk-vm
  Commit: 0222b96ffc697f25aadba4fca0672ccb78a7cb94
      https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/0222b96ffc697f25aadba4fca0672ccb78a7cb94
  Author: Nicolas Cellier <[hidden email]>
  Date:   2020-03-26 (Thu, 26 Mar 2020)

  Changed paths:
    M platforms/minheadless/unix/sqUnixSpurMemory.c
    M platforms/minheadless/windows/sqWin32Alloc.c
    M platforms/minheadless/windows/sqWin32SpurAlloc.c

  Log Message:
  -----------
  Fix minheadless crash at startup

The changes to `sqMakeMemoryExecutableFromToCodeToDataDelta` performed in
https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/051a0fe56561b586c8e36f8c2950efd391059605

Notably:

> Don't attempt to take the address of 0 in the DUAL_MAPPED_CODE_ZONE regime.

The minheadless being now mostly unmaintained, and having large duplication of platforms code base, it can only rot...


Reply | Threaded
Open this post in threaded view
|

Re: [OpenSmalltalk/opensmalltalk-vm] 0222b9: Fix minheadless crash at startup

Tobias Pape
 

> On 26.03.2020, at 00:03, Nicolas Cellier <[hidden email]> wrote:
>
>
>  Branch: refs/heads/Cog
>  Home:   https://github.com/OpenSmalltalk/opensmalltalk-vm
>  Commit: 0222b96ffc697f25aadba4fca0672ccb78a7cb94
>      https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/0222b96ffc697f25aadba4fca0672ccb78a7cb94
>  Author: Nicolas Cellier <[hidden email]>
>  Date:   2020-03-26 (Thu, 26 Mar 2020)
>
>  Changed paths:
>    M platforms/minheadless/unix/sqUnixSpurMemory.c
>    M platforms/minheadless/windows/sqWin32Alloc.c
>    M platforms/minheadless/windows/sqWin32SpurAlloc.c
>
>  Log Message:
>  -----------
>  Fix minheadless crash at startup
>
> The changes to `sqMakeMemoryExecutableFromToCodeToDataDelta` performed in
> https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/051a0fe56561b586c8e36f8c2950efd391059605
>
> Notably:
>
>> Don't attempt to take the address of 0 in the DUAL_MAPPED_CODE_ZONE regime.
>
> The minheadless being now mostly unmaintained, and having large duplication of platforms code base, it can only rot...

what is it in the first place?
-t
>


Reply | Threaded
Open this post in threaded view
|

Re: [OpenSmalltalk/opensmalltalk-vm] 0222b9: Fix minheadless crash at startup

Eliot Miranda-2
 


> On Mar 25, 2020, at 5:07 PM, Tobias Pape <[hidden email]> wrote:
>
> 
>
>> On 26.03.2020, at 00:03, Nicolas Cellier <[hidden email]> wrote:
>>
>>
>> Branch: refs/heads/Cog
>> Home:   https://github.com/OpenSmalltalk/opensmalltalk-vm
>> Commit: 0222b96ffc697f25aadba4fca0672ccb78a7cb94
>>     https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/0222b96ffc697f25aadba4fca0672ccb78a7cb94
>> Author: Nicolas Cellier <[hidden email]>
>> Date:   2020-03-26 (Thu, 26 Mar 2020)
>>
>> Changed paths:
>>   M platforms/minheadless/unix/sqUnixSpurMemory.c
>>   M platforms/minheadless/windows/sqWin32Alloc.c
>>   M platforms/minheadless/windows/sqWin32SpurAlloc.c
>>
>> Log Message:
>> -----------
>> Fix minheadless crash at startup
>>
>> The changes to `sqMakeMemoryExecutableFromToCodeToDataDelta` performed in
>> https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/051a0fe56561b586c8e36f8c2950efd391059605
>>
>> Notably:
>>
>>> Don't attempt to take the address of 0 in the DUAL_MAPPED_CODE_ZONE regime.
>>
>> The minheadless being now mostly unmaintained, and having large duplication of platforms code base, it can only rot...
>
> what is it in the first place?

Ronie Salgado’s code for a minimal headless vm suitable for a GUI interface implemented entirely with FFI calls and callbacks.  This should not be allowed to rot.  For example it is useful for a scripting vm.

It is good work and a tragedy that Pharo has forked.  Ronie himself is heading towards a non-vm implementation of Smalltalk.  I think that’s a mistake, but that’s a different discussion.  I say this because I fear that now Ronie has reduced interest in either the OpenSmalltalk or Pharo VM efforts.

> -t
>>
>
>
Reply | Threaded
Open this post in threaded view
|

Re: [OpenSmalltalk/opensmalltalk-vm] 0222b9: Fix minheadless crash at startup

Tobias Pape
 

> On 26.03.2020, at 03:17, Eliot Miranda <[hidden email]> wrote:
>
>
>
>
>> On Mar 25, 2020, at 5:07 PM, Tobias Pape <[hidden email]> wrote:
>>
>> 
>>
>>> On 26.03.2020, at 00:03, Nicolas Cellier <[hidden email]> wrote:
>>>
>>>
>>> Branch: refs/heads/Cog
>>> Home:   https://github.com/OpenSmalltalk/opensmalltalk-vm
>>> Commit: 0222b96ffc697f25aadba4fca0672ccb78a7cb94
>>>    https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/0222b96ffc697f25aadba4fca0672ccb78a7cb94
>>> Author: Nicolas Cellier <[hidden email]>
>>> Date:   2020-03-26 (Thu, 26 Mar 2020)
>>>
>>> Changed paths:
>>>  M platforms/minheadless/unix/sqUnixSpurMemory.c
>>>  M platforms/minheadless/windows/sqWin32Alloc.c
>>>  M platforms/minheadless/windows/sqWin32SpurAlloc.c
>>>
>>> Log Message:
>>> -----------
>>> Fix minheadless crash at startup
>>>
>>> The changes to `sqMakeMemoryExecutableFromToCodeToDataDelta` performed in
>>> https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/051a0fe56561b586c8e36f8c2950efd391059605
>>>
>>> Notably:
>>>
>>>> Don't attempt to take the address of 0 in the DUAL_MAPPED_CODE_ZONE regime.
>>>
>>> The minheadless being now mostly unmaintained, and having large duplication of platforms code base, it can only rot...
>>
>> what is it in the first place?
>
> Ronie Salgado’s code for a minimal headless vm suitable for a GUI interface implemented entirely with FFI calls and callbacks.  This should not be allowed to rot.  For example it is useful for a scripting vm.
>
> It is good work and a tragedy that Pharo has forked.  Ronie himself is heading towards a non-vm implementation of Smalltalk.  I think that’s a mistake, but that’s a different discussion.  I say this because I fear that now Ronie has reduced interest in either the OpenSmalltalk or Pharo VM efforts.

Thanks for the clarification.
Best regards
        -Tobias