Upcoming CFV heads-up

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
34 messages Options
12
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Upcoming CFV heads-up

Donal K. Fellows-2
Hi everyone!

I'm planning to call some votes on TIPs soon. The ones I'm thinking
about doing the call on are some of mine where I think we've got an
implementation ready (or where I can make one quickly):

    * TIP #463: Command-Driven Substitutions for regsub
    * TIP #470: Reliable Access to OO Definition Context Object
    * TIP #473: Allow a Defined Target Namespace in oo::copy

I guess there might be other TIPs that could go at the same time. Let me
know if you think there's a problem with any of the above TIPs or if
there's ones that should be added to this tranche.

Donal.


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core

donal_k_fellows.vcf (241 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Donal K. Fellows-2
On 10/06/2017 15:35, Donal K. Fellows wrote:
> I guess there might be other TIPs that could go at the same time. Let me
> know if you think there's a problem with any of the above TIPs or if
> there's ones that should be added to this tranche.

I'm told that there's another about ready too:

    * TIP 472: Add Support for 0d Radix Prefix to Integer Literals

Donal.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core

donal_k_fellows.vcf (241 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Donal K. Fellows-2
In reply to this post by Donal K. Fellows-2
On 10/06/2017 15:35, Donal K. Fellows wrote:
> I'm planning to call some votes on TIPs soon.

OK, here we go!

This is a CFV for the following TIPs. A number of these propose to
target 8.6; as 8.6 is a release, I'll not where this is the case and ask
for people voting for the TIPs to say if they're OK with the feature
going into 8.6 or whether it has to just go to trunk.

Note that all these TIPs also have tests and docs in their
implementations; approvals can be finalized extremely rapidly.

  • TIP #463: Command-Driven Substitutions for regsub
    Link: http://www.tcl.tk/cgi-bin/tct/tip/463.html
    Implementation branch: tip-463
    Target: 8.7

  • TIP #470: Reliable Access to OO Definition Context Object
    Link: http://www.tcl.tk/cgi-bin/tct/tip/470.html
    Implementation branch: tip-470
    Target: 8.7

  • TIP #472: Add Support for 0d Radix Prefix to Integer Literals
    Link: http://www.tcl.tk/cgi-bin/tct/tip/473.html
    Implementation branch: bsg-0d-radix-prefix
    Target: 8.6

  • TIP #473: Allow a Defined Target Namespace in oo::copy
    Link: http://www.tcl.tk/cgi-bin/tct/tip/473.html
    Implementation branch: oo-copy-ns
    Target: 8.6

I'm relatively OK with #472 going into 8.6, but could also live with it
being an 8.7 feature. Provided we don't boil the ocean in the bikeshed
over it. :-)

I'd like #473 to go into 8.6 as it's really a bugfix for an oversight
during the creation of TclOO; [oo::copy] was supposed to have this
capability all along (well, since I allowed specific namespaces to be
targeted at all in the first place) but I forgot to fix it. :-)

Votes are to be done in the usual format to this list with a deadline of
[clock format 1498129200] (i.e., by midday my timezone on Thursday next
week). My votes are as follows:

   TIP #463: YES
     (I don't know how many years I've wanted this!)
   TIP #470: YES
     (We can backport easily enough, but there's no pressing need for
     doing so from my perspective)
   TIP #472: YES
     (I worry about %#d, but it's *REALLY* obscure! I would be
     marginally happier with going to 8.7 with this, but not so as to
     change my vote.)
   TIP #473: YES
     (Glorified bugfix, this one.)

Donal.


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core

donal_k_fellows.vcf (241 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Alexandre Ferrieux
After deep pondering (approx. 100ms),

 TIP #463: YES
 TIP #470: YES
 TIP #472: YES
 TIP #473: YES

-Alex

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Harald Oehlmann
In reply to this post by Donal K. Fellows-2
Am 15.06.2017 um 01:05 schrieb Donal K. Fellows:
>   TIP #463: YES
>     (I don't know how many years I've wanted this!)

+1 !

My first reaction was: please in 8.6.... but better late than never.

Thank you,
Harald

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Jan Nijtmans-2
In reply to this post by Donal K. Fellows-2
2017-06-15 1:05 GMT+02:00 Donal K. Fellows:
> This is a CFV for the following TIPs.

Here are my votes:

  TIP #463: YES
  TIP #470: YES for main part, ABSTAIN for %d part (see remark below)
  TIP #472: YES
  TIP #473: YES

> I'm relatively OK with #472 going into 8.6, but could also live with it
> being an 8.7 feature. Provided we don't boil the ocean in the bikeshed over
> it. :-)

First, I rebased the branch to core-8-6-branch, so this is exactly
what's proposed in the TIP: <http://core.tcl.tk/tcl/info/ae4471b37ced1b49>

>     (I worry about %#d, but it's *REALLY* obscure! I would be
>     marginally happier with going to 8.7 with this, but not so as to
>     change my vote.)

I agree with that: I also would prefer to leave out those lines
from the code for now:
    <http://core.tcl.tk/tcl/artifact/ce447bbf6963ab07?ln=2039-2043>
I already formulated a counter-proposal in another mail, giving
'#' a clear meaning, but I can imagine that people want to think
about this. It would be a pity to put the most important part
of the TIP on hold for this. I'll leave the final decision on
this to Brian. I won't block it.

Regards,
       Jan Nijtmans

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Steve Landers
In reply to this post by Donal K. Fellows-2

On 15 Jun 2017, 7:05 AM +0800, Donal K. Fellows <[hidden email]>, wrote:
On 10/06/2017 15:35, Donal K. Fellows wrote:
I'm planning to call some votes on TIPs soon.

OK, here we go!

This is a CFV for the following TIPs. A number of these propose to
target 8.6; as 8.6 is a release, I'll not where this is the case and ask
for people voting for the TIPs to say if they're OK with the feature
going into 8.6 or whether it has to just go to trunk. 

TIP #463: YES

TIP #470: YES for main part, PASS for %d part

TIP #472: YES

TIP #473: YES


My concern with the %d pass is that anyone likely to be affected probably won’t read the release notes and so it would violate the law of least surprises.


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Brian Griffin-4
In reply to this post by Jan Nijtmans-2

On Jun 15, 2017, at 1:07 AM, Jan Nijtmans <[hidden email]> wrote:

2017-06-15 1:05 GMT+02:00 Donal K. Fellows:
This is a CFV for the following TIPs.

Here are my votes:

 TIP #463: YES
 TIP #470: YES for main part, ABSTAIN for %d part (see remark below)

The above comment is meant for 472, no?

 TIP #472: YES
 TIP #473: YES

I'm relatively OK with #472 going into 8.6, but could also live with it
being an 8.7 feature. Provided we don't boil the ocean in the bikeshed over
it. :-)

First, I rebased the branch to core-8-6-branch, so this is exactly
what's proposed in the TIP: <http://core.tcl.tk/tcl/info/ae4471b37ced1b49>

Thanks!


   (I worry about %#d, but it's *REALLY* obscure! I would be
   marginally happier with going to 8.7 with this, but not so as to
   change my vote.)

I agree with that: I also would prefer to leave out those lines
from the code for now:
   <http://core.tcl.tk/tcl/artifact/ce447bbf6963ab07?ln=2039-2043>
I already formulated a counter-proposal in another mail, giving
'#' a clear meaning, but I can imagine that people want to think
about this. It would be a pity to put the most important part
of the TIP on hold for this. I'll leave the final decision on
this to Brian. I won't block it.

The '#' option is present to be compatible with C standard printf.  Extending support to %d seems logical and pragmatic.  I don't understand the concern, is it specifically with %#d, or the '#' option in general?

IMO, questioning, redefining or extending the semantics of '#' in general should be the topic of a new TIP.

Either way, I defer to Venkat, the true author of the idea (I'm just the coder), and abide by the consensus  of the TCT.

-Brian


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Brian Griffin-4
In reply to this post by Donal K. Fellows-2

On Jun 14, 2017, at 4:05 PM, Donal K. Fellows <[hidden email]> wrote:

On 10/06/2017 15:35, Donal K. Fellows wrote:
I'm planning to call some votes on TIPs soon.

OK, here we go!

This is a CFV for the following TIPs. A number of these propose to target 8.6; as 8.6 is a release, I'll not where this is the case and ask for people voting for the TIPs to say if they're OK with the feature going into 8.6 or whether it has to just go to trunk.


TIP #463: (8.7)  YES 

TIP #470: (8.7)  YES

TIP #472: (8.6)  YES

TIP #473: (8.6)  YES

-Brian



------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Kevin Kenny-6
In reply to this post by Donal K. Fellows-2
TIP 463 : YES
TIP 470 : YES
TIP 472 : YES on recognizing 0d, PRESENT on %#d [1]
TIP 473 : YES

[1] I realize that I am voting PRESENT on my own amendment. If it does not pass without my vote, I'm happy to consider it to have been proposed improvidently.


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Donald G Porter-2
In reply to this post by Donal K. Fellows-2
On 06/14/2017 07:05 PM, Donal K. Fellows wrote:
> This is a CFV for the following TIPs.

Note that these TIPs don't have their headers changed to

   Vote: In progress

which has the effect of locking out changes.

We've approved a new fossil TIP system; does it have the same
capability?  Does anyone but me care?

--
| Don Porter            Applied and Computational Mathematics Division |
| [hidden email]             Information Technology Laboratory |
| http://math.nist.gov/~DPorter/                                  NIST |
|______________________________________________________________________|

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

TIP 472 Votes

Donald G Porter-2
In reply to this post by Donal K. Fellows-2
On 06/14/2017 07:05 PM, Donal K. Fellows wrote:
>   • TIP #472: Add Support for 0d Radix Prefix to Integer Literals
>     Link: http://www.tcl.tk/cgi-bin/tct/tip/473.html
>     Implementation branch: bsg-0d-radix-prefix
>     Target: 8.6

TIP 472 for 8.6.7: NO

New features belong in alpha branches absent compelling need,
or total lack of an alpha branch where it can go. Neither applies.

Besides the general rule, there's evident uncertainty about
whether %#d behavior is the best choice to make, and release
in 8.6.7 locks one choice in place. Better to release in alpha
and gather evidence from at least alpha testers whether there's
trouble or better options, with ability to change if needed.

The recognition of 0d123 as integer is also not as incidental as
people seem to believe.  It will change the dataflow of code
like:

if {[string is integer $val]} {...} else {...}

I think most uses will see the change as harmless, but I also
know I'm not as creative as the entire universe of Tcl coders
out there. Again an alpha period to discover trouble is what
we should do.

TIP 472 for 8.7 (bsg-0d-radix-prefix): PRESENT

Nothing so bad as to oppose, if this is exactly what some
audience wants, but not something I really love either.

TIP 472 for 8.7 (jn-0d-radix-prefix): YES

What shifts my opinion is that Jan's refinement at long
last gives a principled purpose to the %# formats, one
that can be clearly documented as something other than
"something close to what C's printf does, usually".  It
even creates a reason why "0" gets treated differently. In
this functional reform, the need for a "0d" grows out of
providing the %#0d functionality, and the disappearance of it
is baked into the disappearance of the need, assuming
a Tcl 9 with TIP 114 in it.  A really nice piece of
migration support.

--
| Don Porter            Applied and Computational Mathematics Division |
| [hidden email]             Information Technology Laboratory |
| http://math.nist.gov/~DPorter/                                  NIST |
|______________________________________________________________________|

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Donal K. Fellows-2
In reply to this post by Donald G Porter-2
On 16/06/2017 14:38, Donald G Porter wrote:
> On 06/14/2017 07:05 PM, Donal K. Fellows wrote:
>> This is a CFV for the following TIPs.
>
> Note that these TIPs don't have their headers changed to
>
>   Vote: In progress
>
> which has the effect of locking out changes.

Yeah, I got lazy. Fixed now.

> We've approved a new fossil TIP system; does it have the same
> capability?  Does anyone but me care?

I don't know. I also don't know how to actually use the new system. :-)

Donal.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core

donal_k_fellows.vcf (241 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Mark Janssen-6


>> We've approved a new fossil TIP system; does it have the same
>> capability? Does anyone but me care?

> I don't know. I also don't know how to actually use the new system. :-)


My proposal is that the new system would track the CFV in a fossil event. So it's clear what version of the TIP the vote is against it would even be possible to link to the actial artifact id in the CFV.
I am currently still in holidays, will follow up on the tip move to fossil when I am back.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Upcoming CFV heads-up

Jos Decoster-2
In reply to this post by Donal K. Fellows-2
TIP #463: YES
TIP #470: YES
TIP #472: YES, my preference is to put it in 8.7 and to consider on Jan's proposal for #d handling.
TIP #473: YES




  • TIP #470: Reliable Access to OO Definition Context Object
    Link: http://www.tcl.tk/cgi-bin/tct/tip/470.html
    Implementation branch: tip-470
    Target: 8.7

  • TIP #472: Add Support for 0d Radix Prefix to Integer Literals
    Link: http://www.tcl.tk/cgi-bin/tct/tip/473.html
    Implementation branch: bsg-0d-radix-prefix
    Target: 8.6

  • TIP #473: Allow a Defined Target Namespace in oo::copy
    Link: http://www.tcl.tk/cgi-bin/tct/tip/473.html
    Implementation branch: oo-copy-ns
    Target: 8.6

I'm relatively OK with #472 going into 8.6, but could also live with it
being an 8.7 feature. Provided we don't boil the ocean in the bikeshed
over it. :-)

I'd like #473 to go into 8.6 as it's really a bugfix for an oversight
during the creation of TclOO; [oo::copy] was supposed to have this
capability all along (well, since I allowed specific namespaces to be
targeted at all in the first place) but I forgot to fix it. :-)

Votes are to be done in the usual format to this list with a deadline of
[clock format 1498129200] (i.e., by midday my timezone on Thursday next
week). My votes are as follows:

   TIP #463: YES
     (I don't know how many years I've wanted this!)
   TIP #470: YES
     (We can backport easily enough, but there's no pressing need for
     doing so from my perspective)
   TIP #472: YES
     (I worry about %#d, but it's *REALLY* obscure! I would be
     marginally happier with going to 8.7 with this, but not so as to
     change my vote.)
   TIP #473: YES
     (Glorified bugfix, this one.)

Donal.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: TIP 472 Votes

Donal K. Fellows-2
In reply to this post by Donald G Porter-2
On 16/06/2017 15:04, Donald G Porter wrote:
> TIP 472 for 8.6.7: NO

Having read your argument, I concur. I change my vote on this TIP to:

TIP 472 for 8.6.7: NO
TIP 472 for 8.7: YES

All stuff to do with formats really gets a bit of a PRESENT from me at
the moment. It might be good, but I've not had inclination to properly
review something subtle like that and what you say makes me think that
such consideration is warranted.

Donal.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core

donal_k_fellows.vcf (241 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: TIP 472 Votes

Jan Nijtmans-2
2017-06-20 9:15 GMT+02:00 Donal K. Fellows:
> On 16/06/2017 15:04, Donald G Porter wrote:
>> TIP 472 for 8.6.7: NO
> Having read your argument, I concur. I change my vote on this TIP to:
>
> TIP 472 for 8.6.7: NO
> TIP 472 for 8.7: YES

Well, counting the votes so far, I see 6 YES and 2 NO, so that
means the main part of TIP #472 would be accepted despite of
the two NO's.
Don wrote:
> New features belong in alpha branches absent compelling need,
> or total lack of an alpha branch where it can go. Neither applies.
I would rephrase that slightly:
    "New features should be tested well before ending in a release"
Adding 0d to the integer parsing is well tested IMHO: Several
people had a look at the code changes (including myself), test-cases
are added to the test-suite. There are no multiple implementation
possibilities. So, I really question what an alpha phase would add,
a RC phase would be fine with me.

Regarding the "compelling need", I think there is: Currently there
is no way in Tcl to force interpretation as a decimal. TIP #472
allows to prefix "0d", without the need for applications to
convert between 'internal' and 'external' integer format. I really
would hate to force applications to do that, better is letting
Tcl handle it transparently. So, I don't see a reason to change
my vote.

Regarding "%#d", there is no compelling need, since "format %#d"
can be implemented as:
       if {$d >= 0} {
           format 0d%d $d
       } else {
           format -0d%d [expr {-$d}]
       }
(in most cases $d will be positive, which makes this even simpler)
So, I don't see compelling need for implementing format %#d.

Since the %#d implementation has currently 3 YES votes and
2 NO votes (for inclusion in 8.6.7), I don't see a compelling
need to change my vote to NO. The message is clear already,
it's better to wait a little, using the time to sort things out.

Thanks for all feedback!
       Jan Nijtmans

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: TIP 472 Votes

Andreas Leitgeb
Jan Nijtmans <[hidden email]> wrote:
> Regarding the "compelling need", I think there is: Currently there
> is no way in Tcl to force interpretation as a decimal. TIP #472
> allows to prefix "0d", without the need for applications to
> convert between 'internal' and 'external' integer format.

Unless the tip also supports "0d-42", I'm not all that sure about the
practical value of "prefixing 0d" after all. If you still need to
check for a sign first, then using scan %d isn't quite as verbose
in comparison, anymore.

Otoh., if  0d-42  *was* legal, and if furthermore expr accepted a
"0d" bareword prefixing an "optional-sign--rest-all-digits" operand,
to the semantics of applying a decimal interpretation to the
digits, *then* it might be a convenience to use.

0d, 0o, 0x, 0b barewords could be considered for expr&co, anyway:
   if {0d$hour > 23} ...
   if {0x$hex == 0xcafebabe} ...
Maybe a separate TIP?  That would not merely save two "-quotes, but
also (eventually) still work for negatives...

But then again, I might just overvalue negative integers, ...


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: TIP 472 Votes

Kevin Kenny-6
On Wed, Jun 21, 2017 at 2:10 PM, Andreas Leitgeb <[hidden email]> wrote:

> Jan Nijtmans <[hidden email]> wrote:
>> Regarding the "compelling need", I think there is: Currently there
>> is no way in Tcl to force interpretation as a decimal. TIP #472
>> allows to prefix "0d", without the need for applications to
>> convert between 'internal' and 'external' integer format.
>
> Unless the tip also supports "0d-42", I'm not all that sure about the
> practical value of "prefixing 0d" after all. If you still need to
> check for a sign first, then using scan %d isn't quite as verbose
> in comparison, anymore.

I would have presumed that the correct notation would be -0d42,
and that [format %#d] would do The Right Thing. That's the advantage
to %#d over 0d%d.

Nevertheless, my PRESENT vote stands - I'm still willing to
consider the idea improvidently proposed if it cannot pass
without my vote.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: TIP 472 Votes

Andreas Leitgeb
On Wed, Jun 21, 2017 at 02:32:22PM -0400, Kevin Kenny wrote:

> On Wed, Jun 21, 2017 at 2:10 PM, Andreas Leitgeb <[hidden email]> wrote:
>> Jan Nijtmans <[hidden email]> wrote:
>>> Regarding the "compelling need", I think there is: Currently there
>>> is no way in Tcl to force interpretation as a decimal. TIP #472
>>> allows to prefix "0d", without the need for applications to
>>> convert between 'internal' and 'external' integer format.
>> Unless the tip also supports "0d-42", I'm not all that sure about the
>> practical value of "prefixing 0d" after all. If you still need to
>> check for a sign first, then using scan %d isn't quite as verbose
>> in comparison, anymore.
> I would have presumed that the correct notation would be -0d42,
> and that [format %#d] would do The Right Thing. That's the advantage
> to %#d over 0d%d.

I don't quite get the usecase for *producing* 0d-prefixed numbers,
anyway.  If I were to present numbers to humans, "0d" would rather
confuse them, and if I were to produce numbers to be postprocessed
by other tools, they'd likely have a hard time with the "d", too.

The usecase I did believe to understand is the one where I have a
possibly-zero-padded decimal string and want to use it with least
possible hassle. - That's the context for my remarks in my previous
post.

I think, the TIP author's intention also was merely to be able to tack
some prefix to an "incoming" number to make sure it gets parsed as
decimal, even if it originally started with a 0.  Unfortunately I
missed, if they also specified how the tag should work with negatives.

Any momentum towards [format] was only triggered by perceived
needs of orthogonality, which in this case I don't see.
Java, for example doesn't define "#" for %d in String.format(), but
does for both %o and %x/%X.
The orthogonality is broken, anyway, since octal and hexadecimal
values in IT are typically unsigned, so they just don't have that
problem.


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/tcl-core
12
Loading...