Locked by Smalltalkhub and Tonel

On 10/16/20 3:21 PM, bruno buzzi brassesco via Glass wrote:
> How do you want to proceed ?

Well, we should make the tests pass?

I don't have the code loaded in an image, so I can't run tests and debug
them myself ... this is the part where I was hoping to get help and not
have to spend time in the details (until I'm out from under the critical
path at least):)

At this point I don't know if the test is failing because the
associations aren't in the same order or if one of the associations was
incorrectly copied.

I don't even know which collections are involved ... or which copy
method has changed (if indeed #copy is the reason that the test is
failing) ...

I suppose I would need to understand exactly the point where the
execution goes bad and then I should be able to suggest a course of
action ...

Dale

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

Dale,

I will help no problem :)
Maybe i was not clear enough.

The tests all green in GS versions: 3.2.17, 3.3.9, 3.4.5, 3.5.0 and 3.5.4.
In GS 3.1.0.6 almost all tests pass but #testCopy .

After debugging i found that :
|d1 d2 |
d1 := OrderPreservingIdentityStringDictionary new.
d1 at: '1' put: '1'.
d2 := d1 copy.
(d1 associations at: 1) == (d2 associations at: 1)

GS 3.1.0.6 answer true. In others versions answer false.
Any way on Monday afternoon i will research further.

regards,
bruno

On 10/16/20 3:21 PM, bruno buzzi brassesco via Glass wrote:

How do you want to proceed ?

Well, we should make the tests pass?

I don't have the code loaded in an image, so I can't run tests and debug them myself ... this is the part where I was hoping to get help and not have to spend time in the details (until I'm out from under the critical path at least):)

At this point I don't know if the test is failing because the associations aren't in the same order or if one of the associations was incorrectly copied.

I don't even know which collections are involved ... or which copy method has changed (if indeed #copy is the reason that the test is failing) ...

I suppose I would need to understand exactly the point where the execution goes bad and then I should be able to suggest a course of action ...

Dale

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

Dale,

I will help no problem :)
Maybe i was not clear enough.

The tests all green in GS versions: 3.2.17, 3.3.9, 3.4.5, 3.5.0 and 3.5.4.
In GS 3.1.0.6 almost all tests pass but #testCopy .

Ah, much clearer...

After debugging i found that :
|d1 d2 |
d1 := OrderPreservingIdentityStringDictionary new.
d1 at: '1' put: '1'.
d2 := d1 copy.
(d1 associations at: 1) == (d2 associations at: 1)

GS 3.1.0.6 answer true. In others versions answer false.
Any way on Monday afternoon i will research further.

It does appear that one of the #copy methods is doing a deeper copy in 3.1.0.6 which IS interesting ... have a good weekend!

Dale

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

Dale,

Now all Tests are green for GS 3.1.0.6, 3.2.17, 3.3.9, 3.4.5, 3.5.0, 3.5.4.
Check at:
https://travis-ci.org/github/brunobuzzi/OrderPreservingDictionary/builds/737198238

Do you want me to add another GS version ?

Tomorrow I will send the PR so now we have 2 of 5.

regards,
bruno

On 10/16/20 4:59 PM, bruno buzzi brassesco via Glass wrote:

Dale,

I will help no problem :)
Maybe i was not clear enough.

The tests all green in GS versions: 3.2.17, 3.3.9, 3.4.5, 3.5.0 and 3.5.4.
In GS 3.1.0.6 almost all tests pass but #testCopy .

Ah, much clearer...

After debugging i found that :
|d1 d2 |
d1 := OrderPreservingIdentityStringDictionary new.
d1 at: '1' put: '1'.
d2 := d1 copy.
(d1 associations at: 1) == (d2 associations at: 1)

GS 3.1.0.6 answer true. In others versions answer false.
Any way on Monday afternoon i will research further.

It does appear that one of the #copy methods is doing a deeper copy in 3.1.0.6 which IS interesting ... have a good weekend!

Dale

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

Excellent ... I don't think we need to add additional versions unless we find a regression that is version specific ...

Dale

Dale,

Now all Tests are green for GS 3.1.0.6, 3.2.17, 3.3.9, 3.4.5, 3.5.0, 3.5.4.
Check at:
https://travis-ci.org/github/brunobuzzi/OrderPreservingDictionary/builds/737198238

Do you want me to add another GS version ?

Tomorrow I will send the PR so now we have 2 of 5.

regards,
bruno

On 16/10/2020 21:07, Dale Henrichs via Glass wrote:

On 10/16/20 4:59 PM, bruno buzzi brassesco via Glass wrote:

Dale,

I will help no problem :)
Maybe i was not clear enough.

The tests all green in GS versions: 3.2.17, 3.3.9, 3.4.5, 3.5.0 and 3.5.4.
In GS 3.1.0.6 almost all tests pass but #testCopy .

Ah, much clearer...

After debugging i found that :
|d1 d2 |
d1 := OrderPreservingIdentityStringDictionary new.
d1 at: '1' put: '1'.
d2 := d1 copy.
(d1 associations at: 1) == (d2 associations at: 1)

GS 3.1.0.6 answer true. In others versions answer false.
Any way on Monday afternoon i will research further.

It does appear that one of the #copy methods is doing a deeper copy in 3.1.0.6 which IS interesting ... have a good weekend!

Dale

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

Dale,

XMLWriter now has all tests green:
https://travis-ci.org/github/brunobuzzi/XML-XMLWriter/builds/737835246

Already sent the PR.

regards,
bruno

Excellent ... I don't think we need to add additional versions unless we find a regression that is version specific ...

Dale

On 10/19/20 1:37 PM, bruno buzzi brassesco via Glass wrote:

Dale,

Now all Tests are green for GS 3.1.0.6, 3.2.17, 3.3.9, 3.4.5, 3.5.0, 3.5.4.
Check at:
https://travis-ci.org/github/brunobuzzi/OrderPreservingDictionary/builds/737198238

Do you want me to add another GS version ?

Tomorrow I will send the PR so now we have 2 of 5.

regards,
bruno

On 16/10/2020 21:07, Dale Henrichs via Glass wrote:

On 10/16/20 4:59 PM, bruno buzzi brassesco via Glass wrote:

Dale,

I will help no problem :)
Maybe i was not clear enough.

The tests all green in GS versions: 3.2.17, 3.3.9, 3.4.5, 3.5.0 and 3.5.4.
In GS 3.1.0.6 almost all tests pass but #testCopy .

Ah, much clearer...

After debugging i found that :
|d1 d2 |
d1 := OrderPreservingIdentityStringDictionary new.
d1 at: '1' put: '1'.
d2 := d1 copy.
(d1 associations at: 1) == (d2 associations at: 1)

GS 3.1.0.6 answer true. In others versions answer false.
Any way on Monday afternoon i will research further.

It does appear that one of the #copy methods is doing a deeper copy in 3.1.0.6 which IS interesting ... have a good weekend!

Dale

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

Excellent!

Dale,

XMLWriter now has all tests green:
https://travis-ci.org/github/brunobuzzi/XML-XMLWriter/builds/737835246

Already sent the PR.

regards,
bruno

On 19/10/2020 20:54, Dale Henrichs via Glass wrote:

Excellent ... I don't think we need to add additional versions unless we find a regression that is version specific ...

Dale

On 10/19/20 1:37 PM, bruno buzzi brassesco via Glass wrote:

Dale,

Now all Tests are green for GS 3.1.0.6, 3.2.17, 3.3.9, 3.4.5, 3.5.0, 3.5.4.
Check at:
https://travis-ci.org/github/brunobuzzi/OrderPreservingDictionary/builds/737198238

Do you want me to add another GS version ?

Tomorrow I will send the PR so now we have 2 of 5.

regards,
bruno

On 16/10/2020 21:07, Dale Henrichs via Glass wrote:

On 10/16/20 4:59 PM, bruno buzzi brassesco via Glass wrote:

Dale,

I will help no problem :)
Maybe i was not clear enough.

The tests all green in GS versions: 3.2.17, 3.3.9, 3.4.5, 3.5.0 and 3.5.4.
In GS 3.1.0.6 almost all tests pass but #testCopy .

Ah, much clearer...

After debugging i found that :
|d1 d2 |
d1 := OrderPreservingIdentityStringDictionary new.
d1 at: '1' put: '1'.
d2 := d1 copy.
(d1 associations at: 1) == (d2 associations at: 1)

GS 3.1.0.6 answer true. In others versions answer false.
Any way on Monday afternoon i will research further.

It does appear that one of the #copy methods is doing a deeper copy in 3.1.0.6 which IS interesting ... have a good weekend!

Dale

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

Dale,

XMLParser is fixed for GS:
https://travis-ci.org/github/brunobuzzi/XML-XMLParser/builds/738430669

Next week is XPath turn :)

regards,
bruno

Excellent ... I don't think we need to add additional versions unless we find a regression that is version specific ...

Dale

On 10/19/20 1:37 PM, bruno buzzi brassesco via Glass wrote:

Dale,

Now all Tests are green for GS 3.1.0.6, 3.2.17, 3.3.9, 3.4.5, 3.5.0, 3.5.4.
Check at:
https://travis-ci.org/github/brunobuzzi/OrderPreservingDictionary/builds/737198238

Do you want me to add another GS version ?

Tomorrow I will send the PR so now we have 2 of 5.

regards,
bruno

On 16/10/2020 21:07, Dale Henrichs via Glass wrote:

On 10/16/20 4:59 PM, bruno buzzi brassesco via Glass wrote:

Dale,

I will help no problem :)
Maybe i was not clear enough.

The tests all green in GS versions: 3.2.17, 3.3.9, 3.4.5, 3.5.0 and 3.5.4.
In GS 3.1.0.6 almost all tests pass but #testCopy .

Ah, much clearer...

After debugging i found that :
|d1 d2 |
d1 := OrderPreservingIdentityStringDictionary new.
d1 at: '1' put: '1'.
d2 := d1 copy.
(d1 associations at: 1) == (d2 associations at: 1)

GS 3.1.0.6 answer true. In others versions answer false.
Any way on Monday afternoon i will research further.

It does appear that one of the #copy methods is doing a deeper copy in 3.1.0.6 which IS interesting ... have a good weekend!

Dale

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass

_______________________________________________
Glass mailing list
[hidden email]
https://lists.gemtalksystems.com/mailman/listinfo/glass