[Glass] createStone test 3.6.0 failed
Dale Henrichs
dale.henrichs at gemtalksystems.com
Wed Dec 2 09:22:26 PST 2020
Marten,
I need to see stacks, and the expression used to create them, please.
Dale
On 12/1/20 12:13 PM, Marten Feldtmann wrote:
> Well, a simple DateAndTime now asString will also fail. In 3.5.x the
> printing stuff ends somewhere in SmallDouble and uses asStringLocaleC,
> now the method is ending up in ScaledDecimal and there is US oriented
> parsing on a string created via asString (and not asStringLocaleC).
>
> Marten
>> Marten Feldtmann via Glass <glass at lists.gemtalksystems.com> hat am
>> 01.12.2020 20:49 geschrieben:
>>
>>
>> Hello Dale,
>>
>> I did a "git pull", I did a " updateGsDevKit -g -t -i" , but the
>> createStone script is still crashing at the end due to the same
>> error. "git log -2" shows only changes up to 24.11.2020 ...
>>
>> The reason are still some *squeak extension methods, so that
>>
>> SmallDateAndTime now inspect
>>
>> throws an unexpected Format error.
>>
>>
>> Marten
>>
>>
>>> Dale Henrichs <dale.henrichs at gemtalksystems.com> hat am 01.12.2020
>>> 18:22 geschrieben:
>>>
>>>
>>> Marten,
>>>
>>> fix for #131 (glassdb/glass) has been pushed to github on the master
>>> branch, so do a `git pull`, try another 3.6.0 createStone and let me
>>> know how it works ...
>>>
>>> Dale
>>>
>>> On 11/30/20 6:46 PM, Dale Henrichs wrote:
>>>
>>>> I've got a fix[1] queued up for
>>>> ScaledDecimal>>_printAsDateAndTimeANSISecondsOn:[2]. It looks like
>>>> I should be able to push the fix tomorrow (travis runs take a
>>>> loooooooong time these days).
>>>>
>>>> Dale
>>>>
>>>> [1] https://github.com/GsDevKit/GsDevKit/issues/131
>>>> [2] https://travis-ci.org/github/glassdb/glass/builds/746889098
>>>>
>>>> On 11/30/20 6:30 PM, Dale Henrichs wrote:
>>>>>
>>>>> That was my guess ... so `Locale decimalPoint` returns a $, ... as
>>>>> it should?
>>>>>
>>>>> Dale
>>>>>
>>>>> On 11/30/20 11:39 AM, Marten Feldtmann wrote:
>>>>>> My decimal point here in Germany is $, and therefore NOT $.
>>>>>>
>>>>>>
>>>>>> Marten
>>>>>>> Dale Henrichs via Glass <glass at lists.gemtalksystems.com> hat am
>>>>>>> 30.11.2020 20:08 geschrieben:
>>>>>>>
>>>>>>>
>>>>>>> Marten,
>>>>>>>
>>>>>>> Could you confirm that your `Locale decimalPoint` is not $. ...
>>>>>>> I'm assuming that that's the case, but I would like a
>>>>>>> confirmation ...
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>> Dale
>>>>>>>
>>>>>>> On 11/30/20 10:13 AM, Dale Henrichs wrote:
>>>>>>>>
>>>>>>>> Hold the phone ... it turns out that your the code in 3.6.0 is
>>>>>>>> expecting there to be a $. in the printString for a
>>>>>>>> ScaledDecimal, and I'm going to assume that you've got a comma
>>>>>>>> ... I should be able to patch that this afternoon ...
>>>>>>>>
>>>>>>>> Sorry for the confusion, but I _assumed_ that the unexpected
>>>>>>>> format error was related to an instance variable reference in a
>>>>>>>> method executed against a Small* class, since we had that
>>>>>>>> problem along the way during the development of 3.6.0 ...
>>>>>>>>
>>>>>>>> Also not that the old implementation of
>>>>>>>> ScaledDecimal>>_printAsDateAndTimeANSISecondsOn: has the same
>>>>>>>> test (for $.), so I guess it _is_ surprising that you haven't
>>>>>>>> hit this problem before ... hmm, maybe we fixed a bug with
>>>>>>>> ScaledDecimal so that it's printString uses the local for
>>>>>>>> printing ...
>>>>>>>>
>>>>>>>> Dale
>>>>>>>>
>>>>>>>> On 11/30/20 10:02 AM, Dale Henrichs wrote:
>>>>>>>>>
>>>>>>>>> could you send me the implementation of
>>>>>>>>> SmallScaledDecimal(ScaledDecimal)>>_printAsDateAndTimeANSISecondsOn:
>>>>>>>>> in your image?
>>>>>>>>>
>>>>>>>>> Use the following topaz command in a topaz session, logged in
>>>>>>>>> as DataCurator:
>>>>>>>>>
>>>>>>>>> lookup ScaledDecimal>>_printAsDateAndTimeANSISecondsOn:
>>>>>>>>>
>>>>>>>>> From the error you are getting it seems that you must be
>>>>>>>>> touching an instance variable in ScaledDecimal in the method
>>>>>>>>> that causes the message to fail with the "unexpected format"
>>>>>>>>> message..
>>>>>>>>>
>>>>>>>>> For the record here is the old implementation (from 3.5.4):
>>>>>>>>>
>>>>>>>>> topaz 1> lookup ScaledDecimal>>_printAsDateAndTimeANSISecondsOn:
>>>>>>>>> category: '*squeak'
>>>>>>>>> method: ScaledDecimal
>>>>>>>>> _printAsDateAndTimeANSISecondsOn: aStream
>>>>>>>>>
>>>>>>>>> | string i |
>>>>>>>>> "Adding 100 causes a very tiny loss of precission but makes parsing much easier"
>>>>>>>>> string := (self + 100) asString.
>>>>>>>>> (string at: 4) = $. ifFalse: [self error: 'unexpected format'].
>>>>>>>>> i := string size.
>>>>>>>>> [
>>>>>>>>> (string at: i) = $0 or: [ (string at: i) = $. ].
>>>>>>>>> ] whileTrue: [
>>>>>>>>> i := i - 1.
>>>>>>>>> ].
>>>>>>>>> aStream nextPutAll: (string copyFrom: 2 to: i).
>>>>>>>>> %
>>>>>>>>>
>>>>>>>>> and the new implementation (for 3.6.0):
>>>>>>>>>
>>>>>>>>> topaz 1> lookup ScaledDecimal>>_printAsDateAndTimeANSISecondsOn:
>>>>>>>>> category: '*squeak'
>>>>>>>>> method: ScaledDecimal
>>>>>>>>> _printAsDateAndTimeANSISecondsOn: aStream
>>>>>>>>> | string i noDecimalPoint |
>>>>>>>>> "Adding 100 causes a very tiny loss of precission but makes parsing much easier"
>>>>>>>>> self = 0
>>>>>>>>> ifTrue: [ aStream nextPutAll: '00' ]
>>>>>>>>> ifFalse: [
>>>>>>>>> string := (self + 100) asString.
>>>>>>>>> (string at: 4) = $.
>>>>>>>>> ifFalse: [ self error: 'unexpected format' ].
>>>>>>>>> i := string size.
>>>>>>>>> noDecimalPoint := true.
>>>>>>>>> [
>>>>>>>>> | char |
>>>>>>>>> "strip trailing 0's"
>>>>>>>>> char := string at: i.
>>>>>>>>> noDecimalPoint
>>>>>>>>> ifTrue: [ noDecimalPoint := char ~= $. ].
>>>>>>>>> noDecimalPoint and: [ char = $0 ] ]
>>>>>>>>> whileTrue: [ i := i - 1 ].
>>>>>>>>> noDecimalPoint
>>>>>>>>> ifFalse: [
>>>>>>>>> "don't include the decimal point in the printString, if $. is last character"
>>>>>>>>> i := i - 1 ].
>>>>>>>>> aStream nextPutAll: (string copyFrom: 2 to: i) ]
>>>>>>>>> %
>>>>>>>>>
>>>>>>>>> I pretty much compared all of the code loaded into the image,
>>>>>>>>> and other than an older version of AST-Core (AST-Core-dkh.105
>>>>>>>>> instead of AST-Core-dkh.106) ... so the RB git repo is
>>>>>>>>> apparently not on the latest version ...
>>>>>>>>>
>>>>>>>>> The code in question is in Squeak.v36-dkh.351 and you have
>>>>>>>>> loaded that version as well ...
>>>>>>>>>
>>>>>>>>> Dale
>>>>>>>>>
>>>>>>>>> On 11/27/20 12:06 AM, Marten Feldtmann via Glass wrote:
>>>>>>>>>> Hey,
>>>>>>>>>>
>>>>>>>>>> actually something must be strange on my system. The original
>>>>>>>>>> error is still there, but I was able to create a stone using
>>>>>>>>>> "createStone -g test 3.6.0", but I can't do anything with it.
>>>>>>>>>>
>>>>>>>>>> I tried Jade 2.4.6 with this running stone, but Jade
>>>>>>>>>> complains (while connecting and starting) a compile error ...
>>>>>>>>>>
>>>>>>>>>> _keysForDictionary: aDictionary
>>>>>>>>>>
>>>>>>>>>> and as I understand it complains an undefined symbol (which
>>>>>>>>>> seems to be the parameter aDictionary):
>>>>>>>>>>
>>>>>>>>>> | keys |
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> (aDictionary isKindOf: RubyHash) ifFalse:[ ^super _keysForDictionary: aDictionary ]
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> ^1
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> I know it does not make much sense ...
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Marten
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Glass mailing list
>>>>>>>>>> Glass at lists.gemtalksystems.com
>>>>>>>>>> https://lists.gemtalksystems.com/mailman/listinfo/glass
>>>>>>> _______________________________________________ Glass mailing
>>>>>>> list Glass at lists.gemtalksystems.com
>>>>>>> https://lists.gemtalksystems.com/mailman/listinfo/glass
>> _______________________________________________ Glass mailing list
>> Glass at lists.gemtalksystems.com
>> https://lists.gemtalksystems.com/mailman/listinfo/glass
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.gemtalksystems.com/mailman/private/glass/attachments/20201202/f9337d1d/attachment.htm>
More information about the Glass
mailing list