The Trunk: Files-dtl.61.mcz

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

The Trunk: Files-dtl.61.mcz

commits-2
David T. Lewis uploaded a new version of Files to project The Trunk:
http://source.squeak.org/trunk/Files-dtl.61.mcz

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

Name: Files-dtl.61
Author: dtl
Time: 25 December 2009, 11:24:59 am
UUID: 395f5d5e-f046-4747-8f2a-9cdb898fddcf
Ancestors: Files-dtl.59

Revert Files-dtl.60.mcz, which caused problems with Monticello saves.

=============== Diff against Files-dtl.59 ===============


Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Files-dtl.61.mcz

David T. Lewis
I ran into problems saving the large changes file patches, and reverted
the update.  Sorry for the noise.

On a positive note, I have a changes file that has reached 426 MB without
problems, so the updated sourcePointer address mapping works fine.

>From my test image:

  (CompiledMethod allInstances collect: [:e | e sourcePointer]) max hex ==> '36A39BF3'

  (SourceFiles at: 2) size ==> 446930379

Dave

On Sat, Dec 26, 2009 at 04:29:49AM +0000, [hidden email] wrote:

> David T. Lewis uploaded a new version of Files to project The Trunk:
> http://source.squeak.org/trunk/Files-dtl.61.mcz
>
> ==================== Summary ====================
>
> Name: Files-dtl.61
> Author: dtl
> Time: 25 December 2009, 11:24:59 am
> UUID: 395f5d5e-f046-4747-8f2a-9cdb898fddcf
> Ancestors: Files-dtl.59
>
> Revert Files-dtl.60.mcz, which caused problems with Monticello saves.
>
> =============== Diff against Files-dtl.59 ===============
>

Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Files-dtl.61.mcz

Levente Uzonyi-2
On Fri, 25 Dec 2009, David T. Lewis wrote:

> I ran into problems saving the large changes file patches, and reverted
> the update.  Sorry for the noise.
>

I guess the problem is related to #fileIndexFromSourcePointer: which
returns 0 if sent to a StandardSourceFileArray when the argument is less
than 16777216 (16r1000000), while ExpandedSourceFileArray returns 2. I
guess that 0 is a valid result and some code expects this value.

Another difference between the two implementations is that
ExpandedSourceFileArray new filePositionFromSourcePointer: 0 ===> -16777216
while
StandardSourceFileArray new filePositionFromSourcePointer: 0 ===> 0

> On a positive note, I have a changes file that has reached 426 MB without
> problems, so the updated sourcePointer address mapping works fine.

Great. :)


Levente

>
>> From my test image:
>
>  (CompiledMethod allInstances collect: [:e | e sourcePointer]) max hex ==> '36A39BF3'
>
>  (SourceFiles at: 2) size ==> 446930379
>
> Dave
>
> On Sat, Dec 26, 2009 at 04:29:49AM +0000, [hidden email] wrote:
>> David T. Lewis uploaded a new version of Files to project The Trunk:
>> http://source.squeak.org/trunk/Files-dtl.61.mcz
>>
>> ==================== Summary ====================
>>
>> Name: Files-dtl.61
>> Author: dtl
>> Time: 25 December 2009, 11:24:59 am
>> UUID: 395f5d5e-f046-4747-8f2a-9cdb898fddcf
>> Ancestors: Files-dtl.59
>>
>> Revert Files-dtl.60.mcz, which caused problems with Monticello saves.
>>
>> =============== Diff against Files-dtl.59 ===============
>>
>
>

Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Files-dtl.61.mcz

David T. Lewis
On Sat, Dec 26, 2009 at 05:49:48AM +0100, Levente Uzonyi wrote:

> On Fri, 25 Dec 2009, David T. Lewis wrote:
>
> >I ran into problems saving the large changes file patches, and reverted
> >the update.  Sorry for the noise.
> >
>
> I guess the problem is related to #fileIndexFromSourcePointer: which
> returns 0 if sent to a StandardSourceFileArray when the argument is less
> than 16777216 (16r1000000), while ExpandedSourceFileArray returns 2. I
> guess that 0 is a valid result and some code expects this value.
>
> Another difference between the two implementations is that
> ExpandedSourceFileArray new filePositionFromSourcePointer: 0 ===> -16777216
> while
> StandardSourceFileArray new filePositionFromSourcePointer: 0 ===> 0

Thank you, you're right. I'll fix this and spend some more time testing.

Dave

>
> >On a positive note, I have a changes file that has reached 426 MB without
> >problems, so the updated sourcePointer address mapping works fine.
>
> Great. :)
>

Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Files-dtl.61.mcz

David T. Lewis
On Sat, Dec 26, 2009 at 02:37:48PM -0500, David T. Lewis wrote:

> On Sat, Dec 26, 2009 at 05:49:48AM +0100, Levente Uzonyi wrote:
> > On Fri, 25 Dec 2009, David T. Lewis wrote:
> >
> > >I ran into problems saving the large changes file patches, and reverted
> > >the update.  Sorry for the noise.
> > >
> >
> > I guess the problem is related to #fileIndexFromSourcePointer: which
> > returns 0 if sent to a StandardSourceFileArray when the argument is less
> > than 16777216 (16r1000000), while ExpandedSourceFileArray returns 2. I
> > guess that 0 is a valid result and some code expects this value.
> >
> > Another difference between the two implementations is that
> > ExpandedSourceFileArray new filePositionFromSourcePointer: 0 ===> -16777216
> > while
> > StandardSourceFileArray new filePositionFromSourcePointer: 0 ===> 0
>
> Thank you, you're right. I'll fix this and spend some more time testing.
>
For reference, I'm attaching an update that is backward compatible
over the sourcePointer range 0 through 16rFFFFFF.

I think the actual problem that I had in loading this to truck was
something to do with Montecello load ordering, which I obviously
don't understand well enough at this point. I'll spend some more
time figuring this out before I risk messing up the trunk again.
Meantime these change sets can be loaded directly with no apparent
problems.

Dave
 



ExpandedSourceFileArray-part1-dtl.2.cs (21K) Download Attachment
ExpandedSourceFileArray-part2-dtl.2.cs (739 bytes) Download Attachment