Smalltalk › Squeak › Squeak - Dev

How to create a SqueakSource image?

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

32 messages Options

Philippe Marschall

Re: How to create a SqueakSource image?

> 3.7 vs. 3.8 issue?

www.squeaksource.com is 3.8 too. I'm really confused.

Frank Shearar

Re: How to create a SqueakSource image?

In reply to this post by Andreas.Raab

This sounds like http://bugs.impara.de/view.php?id=2229 which looks like it
was added to 3.9a-6707.

IntegerTest>>testPrintPaddedWith does exercise the code, although it's fair
to say that this is hardly obvious. Integer>>printOn:base:showRadix:'s sole
sender is Integer>>printPaddedWith:to:base:, but perhaps I should write a
test suite for #printOn:base:showRadix.

I completely agree that (in 6706) there's horrible special-casing for base
10: I wrote some tests detailing what I _thought_ should happen, and made
only the smallest changes I could to make the tests pass. In particular, I
made it work; I didn't make it right.

frank
----- Original Message -----
From: "Andreas Raab" <[hidden email]>
To: "The general-purpose Squeak developers list"
<[hidden email]>
Sent: Tuesday, February 07, 2006 11:52 PM
Subject: Re: How to create a SqueakSource image?

> Actually #printOn:base:showRadix: will fail for any showRadix: false
> unless base is 10. Sounds weird? Yeah. Here are examples:
>
> 100 printOn: Transcript base: 8 showRadix: false "->fails"
> 100 printOn: Transcript base: 10 showRadix: false "->succeeds"
> 100 printOn: Transcript base: 16 showRadix: false "->fails"
>
> Looks like #printStringRadix: is supposed to return a radix whenever
> invoked but doesn't. Note how it is actually special-casing radix 10 -
> it looks like it's expecting #printStringBase: to add the "16r" prefix
> which apparently #printStringBase: doesn't (or rather "no longer") do.
>
> BTW, this is a classic screw-up that tests would have found in no time.
> Where are they when you need 'em?
>
> Cheers,
> - Andreas
>
> Wilkes Joiner wrote:
> > image 3.8 6665
> > it fails on the default color on red 54.
> >
> > On 2/7/06, Philippe Marschall <[hidden email]> wrote:
> >> 2006/2/7, Wilkes Joiner <[hidden email]>:
> >>> That seems to be part of the problem. The error is being generated by
> >>> Color>>hex from
> >>> Integer>>printOn: base: showRadix:
> >>>
> >>> "radix indicator not found."
> >> Can you go up the stack and see for what number this fails?
> >>
> >> What an Image (version) do you have?
> >>
> >>
> >
> >
>
>
>

Bert Freudenberg-3

Re: How to create a SqueakSource image?

In reply to this post by Philippe Marschall

Am 08.02.2006 um 12:10 schrieb Philippe Marschall:

>> 3.7 vs. 3.8 issue?
>
> www.squeaksource.com is 3.8 too. I'm really confused.

source.impara.de is 3.7.

- Bert -

stéphane ducasse-2

Re: How to create a SqueakSource image?

In reply to this post by Andreas.Raab

In 3.9

100 printOn: Transcript base: 8 showRadix: false 100
100 printOn: Transcript base: 10 showRadix: false 100
100 printOn: Transcript base: 16 showRadix: false 100

> Actually #printOn:base:showRadix: will fail for any showRadix:
> false unless base is 10. Sounds weird? Yeah. Here are examples:
>
> 100 printOn: Transcript base: 8 showRadix: false "->fails"
> 100 printOn: Transcript base: 10 showRadix: false "->succeeds"
> 100 printOn: Transcript base: 16 showRadix: false "->fails"
>
> Looks like #printStringRadix: is supposed to return a radix
> whenever invoked but doesn't. Note how it is actually special-
> casing radix 10 - it looks like it's expecting #printStringBase: to
> add the "16r" prefix which apparently #printStringBase: doesn't (or
> rather "no longer") do.
>
> BTW, this is a classic screw-up that tests would have found in no
> time. Where are they when you need 'em?

Indeed.
Send them I will add them to 3.9 (else I will do it).

stéphane ducasse-2

Re: How to create a SqueakSource image?

In reply to this post by Frank Shearar

Yes I just checked that.

testPrintPaddedWith

self assert: (123 printPaddedWith: $0 to: 10 base: 2) = '0001111011'.
self assert: (123 printPaddedWith: $0 to: 10 base: 8) = '0000000173'.
self assert: (123 printPaddedWith: $0 to: 10 base: 10) = '0000000123'.
self assert: (123 printPaddedWith: $0 to: 10 base: 16) = '000000007B'.

So creating a couple of tests with the correct results would be nice.

Stef

On 8 févr. 06, at 12:48, Frank Shearar wrote:

> This sounds like http://bugs.impara.de/view.php?id=2229 which looks
> like it
> was added to 3.9a-6707.
>
> IntegerTest>>testPrintPaddedWith does exercise the code, although
> it's fair
> to say that this is hardly obvious.
> Integer>>printOn:base:showRadix:'s sole
> sender is Integer>>printPaddedWith:to:base:, but perhaps I should
> write a
> test suite for #printOn:base:showRadix.
>
> I completely agree that (in 6706) there's horrible special-casing
> for base
> 10: I wrote some tests detailing what I _thought_ should happen,
> and made
> only the smallest changes I could to make the tests pass. In
> particular, I
> made it work; I didn't make it right.
>
> frank
> ----- Original Message -----
> From: "Andreas Raab" <[hidden email]>
> To: "The general-purpose Squeak developers list"
> <[hidden email]>
> Sent: Tuesday, February 07, 2006 11:52 PM
> Subject: Re: How to create a SqueakSource image?
>
>
>> Actually #printOn:base:showRadix: will fail for any showRadix: false
>> unless base is 10. Sounds weird? Yeah. Here are examples:
>>
>> 100 printOn: Transcript base: 8 showRadix: false "->fails"
>> 100 printOn: Transcript base: 10 showRadix: false "->succeeds"
>> 100 printOn: Transcript base: 16 showRadix: false "->fails"
>>
>> Looks like #printStringRadix: is supposed to return a radix whenever
>> invoked but doesn't. Note how it is actually special-casing radix
>> 10 -
>> it looks like it's expecting #printStringBase: to add the "16r"
>> prefix
>> which apparently #printStringBase: doesn't (or rather "no longer")
>> do.
>>
>> BTW, this is a classic screw-up that tests would have found in no
>> time.
>> Where are they when you need 'em?
>>
>> Cheers,
>> - Andreas
>>
>> Wilkes Joiner wrote:
>>> image 3.8 6665
>>> it fails on the default color on red 54.
>>>
>>> On 2/7/06, Philippe Marschall <[hidden email]> wrote:
>>>> 2006/2/7, Wilkes Joiner <[hidden email]>:
>>>>> That seems to be part of the problem. The error is being
>>>>> generated by
>>>>> Color>>hex from
>>>>> Integer>>printOn: base: showRadix:
>>>>>
>>>>> "radix indicator not found."
>>>> Can you go up the stack and see for what number this fails?
>>>>
>>>> What an Image (version) do you have?
>>>>
>>>>
>>>
>>>
>>
>>
>>
>
>

stéphane ducasse-2

Re: How to create a SqueakSource image?

In reply to this post by wilkesj

hi all

it seems to me that the next endeavour for our nice community
will be to have a way to distribute coherent package set.

Stef

wilkesj

Re: How to create a SqueakSource image?

I think that once MCConfigurations is part of the main image then we
will have a pretty good one. I've become totally dependent on it. I
find it helpful to update a clean image with an mcm as a sanity check
for a project every so often.

On 2/8/06, stéphane ducasse <[hidden email]> wrote:
> hi all
>
> it seems to me that the next endeavour for our nice community
> will be to have a way to distribute coherent package set.
>
> Stef
>
>

Frank Shearar

Re: How to create a SqueakSource image?

In reply to this post by stéphane ducasse-2

Hi Stéphane,

By "tests with the correct results" what do you mean?
IntegerTest>>testPrintOnBaseShowRadix?

I presume that these are the correct results? If so I'll write something up
and put it on Mantis today.

123 printOn: foo base: 10 showRadix: true => 10r123
123 printOn: foo base: 10 showRadix: false => 123
123 printOn: foo base: 8 showRadix: true => 8r173
123 printOn: foo base: 8 showRadix: false => 173

frank

----- Original Message -----
From: "stéphane ducasse" <[hidden email]>
To: "The general-purpose Squeak developers list"
<[hidden email]>
Sent: Wednesday, February 08, 2006 8:18 PM
Subject: Re: How to create a SqueakSource image?

Yes I just checked that.

testPrintPaddedWith

self assert: (123 printPaddedWith: $0 to: 10 base: 2) = '0001111011'.
self assert: (123 printPaddedWith: $0 to: 10 base: 8) = '0000000173'.
self assert: (123 printPaddedWith: $0 to: 10 base: 10) = '0000000123'.
self assert: (123 printPaddedWith: $0 to: 10 base: 16) = '000000007B'.

So creating a couple of tests with the correct results would be nice.

Stef

On 8 févr. 06, at 12:48, Frank Shearar wrote:

Frank Shearar

Re: How to create a SqueakSource image?

http://bugs.impara.de/view.php?id=2698 demonstrates a bug in
Integer>>printOn:base:showRadix: (123 printOn: foo base: 8 showRadix: false
puts '8r173' in s, not '173'), and supplies a fix.

frank
----- Original Message -----
From: "Frank Shearar" <[hidden email]>
To: "The general-purpose Squeak developers list"
<[hidden email]>
Sent: Thursday, February 09, 2006 8:02 AM
Subject: Re: How to create a SqueakSource image?

> Hi Stéphane,
>
> By "tests with the correct results" what do you mean?
> IntegerTest>>testPrintOnBaseShowRadix?
>
> I presume that these are the correct results? If so I'll write something
up

> and put it on Mantis today.
>
> 123 printOn: foo base: 10 showRadix: true => 10r123
> 123 printOn: foo base: 10 showRadix: false => 123
> 123 printOn: foo base: 8 showRadix: true => 8r173
> 123 printOn: foo base: 8 showRadix: false => 173
>
> frank
>
> ----- Original Message -----
> From: "stéphane ducasse" <[hidden email]>
> To: "The general-purpose Squeak developers list"
> <[hidden email]>
> Sent: Wednesday, February 08, 2006 8:18 PM
> Subject: Re: How to create a SqueakSource image?
>
>
> Yes I just checked that.
>
>
> testPrintPaddedWith
>
> self assert: (123 printPaddedWith: $0 to: 10 base: 2) = '0001111011'.
> self assert: (123 printPaddedWith: $0 to: 10 base: 8) = '0000000173'.
> self assert: (123 printPaddedWith: $0 to: 10 base: 10) = '0000000123'.
> self assert: (123 printPaddedWith: $0 to: 10 base: 16) = '000000007B'.
>
> So creating a couple of tests with the correct results would be nice.
>
> Stef
>
>
> On 8 févr. 06, at 12:48, Frank Shearar wrote:
>
> > This sounds like http://bugs.impara.de/view.php?id=2229 which looks
> > like it
> > was added to 3.9a-6707.
> >
> > IntegerTest>>testPrintPaddedWith does exercise the code, although
> > it's fair
> > to say that this is hardly obvious.
> > Integer>>printOn:base:showRadix:'s sole
> > sender is Integer>>printPaddedWith:to:base:, but perhaps I should
> > write a
> > test suite for #printOn:base:showRadix.
> >
> > I completely agree that (in 6706) there's horrible special-casing
> > for base
> > 10: I wrote some tests detailing what I _thought_ should happen,
> > and made
> > only the smallest changes I could to make the tests pass. In
> > particular, I
> > made it work; I didn't make it right.
> >
> > frank
> > ----- Original Message -----
> > From: "Andreas Raab" <[hidden email]>
> > To: "The general-purpose Squeak developers list"
> > <[hidden email]>
> > Sent: Tuesday, February 07, 2006 11:52 PM
> > Subject: Re: How to create a SqueakSource image?
> >
> >
> >> Actually #printOn:base:showRadix: will fail for any showRadix: false
> >> unless base is 10. Sounds weird? Yeah. Here are examples:
> >>
> >> 100 printOn: Transcript base: 8 showRadix: false "->fails"
> >> 100 printOn: Transcript base: 10 showRadix: false "->succeeds"
> >> 100 printOn: Transcript base: 16 showRadix: false "->fails"
> >>
> >> Looks like #printStringRadix: is supposed to return a radix whenever
> >> invoked but doesn't. Note how it is actually special-casing radix
> >> 10 -
> >> it looks like it's expecting #printStringBase: to add the "16r"
> >> prefix
> >> which apparently #printStringBase: doesn't (or rather "no longer")
> >> do.
> >>
> >> BTW, this is a classic screw-up that tests would have found in no
> >> time.
> >> Where are they when you need 'em?
> >>
> >> Cheers,
> >> - Andreas
> >>
> >> Wilkes Joiner wrote:
> >>> image 3.8 6665
> >>> it fails on the default color on red 54.
> >>>
> >>> On 2/7/06, Philippe Marschall <[hidden email]> wrote:
> >>>> 2006/2/7, Wilkes Joiner <[hidden email]>:
> >>>>> That seems to be part of the problem. The error is being
> >>>>> generated by
> >>>>> Color>>hex from
> >>>>> Integer>>printOn: base: showRadix:
> >>>>>
> >>>>> "radix indicator not found."
> >>>> Can you go up the stack and see for what number this fails?
> >>>>
> >>>> What an Image (version) do you have?
> >>>>
> >>>>
> >>>
> >>>
> >>
> >>
> >>
> >
> >
>
>
>
>
>

stéphane ducasse-2

Re: How to create a SqueakSource image?

In reply to this post by Frank Shearar

> Hi Stéphane,
>
> By "tests with the correct results" what do you mean?
> IntegerTest>>testPrintOnBaseShowRadix?

Yes sorry.

> I presume that these are the correct results? If so I'll write
> something up
> and put it on Mantis today.

If you publish directly on the inbox (and ping us) then we will
directly integrate them.

>
> 123 printOn: foo base: 10 showRadix: true => 10r123
> 123 printOn: foo base: 10 showRadix: false => 123
> 123 printOn: foo base: 8 showRadix: true => 8r173
> 123 printOn: foo base: 8 showRadix: false => 173

When I tried yesterday I got really strange results.

>
> frank
>
> ----- Original Message -----
> From: "stéphane ducasse" <[hidden email]>
> To: "The general-purpose Squeak developers list"
> <[hidden email]>
> Sent: Wednesday, February 08, 2006 8:18 PM
> Subject: Re: How to create a SqueakSource image?
>
>
> Yes I just checked that.
>
>
> testPrintPaddedWith
>
> self assert: (123 printPaddedWith: $0 to: 10 base: 2) = '0001111011'.
> self assert: (123 printPaddedWith: $0 to: 10 base: 8) = '0000000173'.
> self assert: (123 printPaddedWith: $0 to: 10 base: 10) = '0000000123'.
> self assert: (123 printPaddedWith: $0 to: 10 base: 16) = '000000007B'.
>
> So creating a couple of tests with the correct results would be nice.
>
> Stef
>
>
> On 8 févr. 06, at 12:48, Frank Shearar wrote:
>
>> This sounds like http://bugs.impara.de/view.php?id=2229 which looks
>> like it
>> was added to 3.9a-6707.
>>
>> IntegerTest>>testPrintPaddedWith does exercise the code, although
>> it's fair
>> to say that this is hardly obvious.
>> Integer>>printOn:base:showRadix:'s sole
>> sender is Integer>>printPaddedWith:to:base:, but perhaps I should
>> write a
>> test suite for #printOn:base:showRadix.
>>
>> I completely agree that (in 6706) there's horrible special-casing
>> for base
>> 10: I wrote some tests detailing what I _thought_ should happen,
>> and made
>> only the smallest changes I could to make the tests pass. In
>> particular, I
>> made it work; I didn't make it right.
>>
>> frank
>> ----- Original Message -----
>> From: "Andreas Raab" <[hidden email]>
>> To: "The general-purpose Squeak developers list"
>> <[hidden email]>
>> Sent: Tuesday, February 07, 2006 11:52 PM
>> Subject: Re: How to create a SqueakSource image?
>>
>>
>>> Actually #printOn:base:showRadix: will fail for any showRadix: false
>>> unless base is 10. Sounds weird? Yeah. Here are examples:
>>>
>>> 100 printOn: Transcript base: 8 showRadix: false "->fails"
>>> 100 printOn: Transcript base: 10 showRadix: false "->succeeds"
>>> 100 printOn: Transcript base: 16 showRadix: false "->fails"
>>>
>>> Looks like #printStringRadix: is supposed to return a radix whenever
>>> invoked but doesn't. Note how it is actually special-casing radix
>>> 10 -
>>> it looks like it's expecting #printStringBase: to add the "16r"
>>> prefix
>>> which apparently #printStringBase: doesn't (or rather "no longer")
>>> do.
>>>
>>> BTW, this is a classic screw-up that tests would have found in no
>>> time.
>>> Where are they when you need 'em?
>>>
>>> Cheers,
>>> - Andreas
>>>
>>> Wilkes Joiner wrote:
>>>> image 3.8 6665
>>>> it fails on the default color on red 54.
>>>>
>>>> On 2/7/06, Philippe Marschall <[hidden email]> wrote:
>>>>> 2006/2/7, Wilkes Joiner <[hidden email]>:
>>>>>> That seems to be part of the problem. The error is being
>>>>>> generated by
>>>>>> Color>>hex from
>>>>>> Integer>>printOn: base: showRadix:
>>>>>>
>>>>>> "radix indicator not found."
>>>>> Can you go up the stack and see for what number this fails?
>>>>>
>>>>> What an Image (version) do you have?
>>>>>
>>>>>
>>>>
>>>>
>>>
>>>
>>>
>>
>>
>
>
>
>

Frank Shearar

Re: How to create a SqueakSource image?

I already submitted a changeset to Mantis
(http://bugs.impara.de/view.php?id=2698) but just so I know, which inbox do
you mean? This would be the Harvester inbox?

I assume that, in general, we still want to post fixes on Mantis?

frank
----- Original Message -----
From: "stéphane ducasse" <[hidden email]>
To: "The general-purpose Squeak developers list"
<[hidden email]>
Sent: Thursday, February 09, 2006 10:43 AM
Subject: Re: How to create a SqueakSource image?

> Hi Stéphane,
>
> By "tests with the correct results" what do you mean?
> IntegerTest>>testPrintOnBaseShowRadix?

Yes sorry.

> I presume that these are the correct results? If so I'll write
> something up
> and put it on Mantis today.

If you publish directly on the inbox (and ping us) then we will
directly integrate them.

>
> 123 printOn: foo base: 10 showRadix: true => 10r123
> 123 printOn: foo base: 10 showRadix: false => 123
> 123 printOn: foo base: 8 showRadix: true => 8r173
> 123 printOn: foo base: 8 showRadix: false => 173

When I tried yesterday I got really strange results.

stéphane ducasse-2

Re: How to create a SqueakSource image?

Reply

> I already submitted a changeset to Mantis
> (http://bugs.impara.de/view.php?id=2698) but just so I know, which
> inbox do
> you mean? This would be the Harvester inbox?

source.squeakfoundation.com/inbox

> I assume that, in general, we still want to post fixes on Mantis?

sure.