On Tuesday, April 10, 2012 18:44:40 H. S. Teoh wrote:
> TDPL, p.117, last para:
> 
> 	... For a user-defined type to be used as a key in an
> 	associative array, it must define two special methods, opHash
> 	and opCmp.
> 
> Really? I thought the convention was toHash (TDPL, p.205). So, which is
> it? Which *should* it be?
> 
> To me, it seems utterly arbitrary that classes should use toHash whereas non-class user-defined types should use opHash. Shouldn't we make it consistent across the board?

I expect that opHash was a mistake and that there should be an errata for that line on page 117: http://erdani.com/tdpl/errata/

- Jonathan M Davis

On Tue, Apr 10, 2012 at 06:49:07PM -0700, Jonathan M Davis wrote:
> On Tuesday, April 10, 2012 18:44:40 H. S. Teoh wrote:
> > TDPL, p.117, last para:
> > 
> > 	... For a user-defined type to be used as a key in an
> > 	associative array, it must define two special methods, opHash
> > 	and opCmp.
> > 
> > Really? I thought the convention was toHash (TDPL, p.205). So, which is
> > it? Which *should* it be?
> > 
> > To me, it seems utterly arbitrary that classes should use toHash whereas non-class user-defined types should use opHash. Shouldn't we make it consistent across the board?
> 
> I expect that opHash was a mistake and that there should be an errata for that line on page 117: http://erdani.com/tdpl/errata/
[...]

Actually, I looked, but it wasn't listed.

Andrei? Is this an error?


T

-- 
Do not reason with the unreasonable; you lose by definition.

On 4/10/12 11:10 PM, H. S. Teoh wrote:
> On Tue, Apr 10, 2012 at 06:49:07PM -0700, Jonathan M Davis wrote:
>> On Tuesday, April 10, 2012 18:44:40 H. S. Teoh wrote:
>>> TDPL, p.117, last para:
>>>
>>> 	... For a user-defined type to be used as a key in an
>>> 	associative array, it must define two special methods, opHash
>>> 	and opCmp.
>>>
>>> Really? I thought the convention was toHash (TDPL, p.205). So, which is
>>> it? Which *should* it be?
>>>
>>> To me, it seems utterly arbitrary that classes should use toHash whereas
>>> non-class user-defined types should use opHash. Shouldn't we make it
>>> consistent across the board?
>>
>> I expect that opHash was a mistake and that there should be an errata
>> for that line on page 117: http://erdani.com/tdpl/errata/
> [...]
>
> Actually, I looked, but it wasn't listed.
>
> Andrei? Is this an error?

Most likely!

Andrei

On Wed, Apr 11, 2012 at 09:01:41AM -0500, Andrei Alexandrescu wrote:
> On 4/10/12 11:10 PM, H. S. Teoh wrote:
> >On Tue, Apr 10, 2012 at 06:49:07PM -0700, Jonathan M Davis wrote:
> >>On Tuesday, April 10, 2012 18:44:40 H. S. Teoh wrote:
> >>>TDPL, p.117, last para:
> >>>
> >>>	... For a user-defined type to be used as a key in an
> >>>	associative array, it must define two special methods, opHash
> >>>	and opCmp.
> >>>
> >>>Really? I thought the convention was toHash (TDPL, p.205). So, which is
> >>>it? Which *should* it be?
> >>>
> >>>To me, it seems utterly arbitrary that classes should use toHash whereas non-class user-defined types should use opHash. Shouldn't we make it consistent across the board?
> >>
> >>I expect that opHash was a mistake and that there should be an errata for that line on page 117: http://erdani.com/tdpl/errata/
> >[...]
> >
> >Actually, I looked, but it wasn't listed.
> >
> >Andrei? Is this an error?
> 
> Most likely!
[...]

So this should be added to the errata, then?


T

-- 
In theory, there is no difference between theory and practice.

On 4/13/12 12:15 PM, H. S. Teoh wrote:
> So this should be added to the errata, then?

I'm on it.

Andrei

On Fri, Apr 13, 2012 at 01:07:39PM -0500, Andrei Alexandrescu wrote:
> On 4/13/12 12:15 PM, H. S. Teoh wrote:
> >So this should be added to the errata, then?
> 
> I'm on it.
[...]

Speaking of which, are you planning at some point to publish a second edition to TDPL? Maybe after the language settles down a bit more? Just curious.


T

-- 
Recently, our IT department hired a bug-fix engineer. He used to work for Volkswagen.

On 4/13/12 1:16 PM, H. S. Teoh wrote:
> On Fri, Apr 13, 2012 at 01:07:39PM -0500, Andrei Alexandrescu wrote:
>> On 4/13/12 12:15 PM, H. S. Teoh wrote:
>>> So this should be added to the errata, then?
>>
>> I'm on it.
> [...]

Added errata with credit. http://erdani.com/tdpl/errata

> Speaking of which, are you planning at some point to publish a second
> edition to TDPL? Maybe after the language settles down a bit more? Just
> curious.

New edition == major changes

New printing == small changes, errata fixes

I think we'll go with a new printing soon.


Andrei

On Friday, April 13, 2012 13:23:00 Andrei Alexandrescu wrote:
> I think we'll go with a new printing soon.

But then I'll have to buy another copy! ;)

Though trying to get your hands on the new printing once it comes out will probably be a bit hard for a while, because the odds of getting a copy of the first printing instead are likely to be fairly high at first.

- Jonathan M Davis

On 4/13/12, Andrei Alexandrescu <SeeWebsiteForEmail@erdani.org> wrote:
> New printing == small changes, errata fixes

Be sure to sneak in the lambda syntax as an errata fix. ;-)