Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Alan Grunwald <nospam.nurdglaw@gmail.com>
Newsgroups: comp.lang.tcl
Subject: Re: Parsing namespace name strings
Date: Sat, 30 Nov 2024 13:12:34 +0000
Organization: A noiseless patient Spider
Lines: 71
Message-ID: <vif31r$1nu0c$1@dont-email.me>
References: <vida8m$186cp$1@dont-email.me>
 <20241129201601.dde7fad0c89a3239a0ae6b8a@example.invalid>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Sat, 30 Nov 2024 14:13:32 +0100 (CET)
Injection-Info: dont-email.me; posting-host="6af356727b6a4f4743bf56fcc8c3e6e8";
	logging-data="1832972"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1/c+mpMM+2Du/O+REf3k0KVGujoxXRl7UA="
User-Agent: Mozilla Thunderbird
Cancel-Lock: sha1:brHvty164f+b7F7usKYm6a8A7QA=
In-Reply-To: <20241129201601.dde7fad0c89a3239a0ae6b8a@example.invalid>
Content-Language: en-US
Bytes: 3313

On 29/11/2024 23:16, Emiliano wrote:
> On Fri, 29 Nov 2024 21:03:25 +0000
> Alan Grunwald <nospam.nurdglaw@gmail.com> wrote:
> 
>> I find I regularly want to know things like "what is first part of a
>> namespace name", "the first two parts" etc. I generally want to get the
>> same answer whether the name is a::b::c or ::a::b::c.
>>
>> I can (and do) do this by getting a list of parts via something like
>>
>> split [string map {"::" ":"} $name] ":"
>>
>> but this is clunky - is there something like [namespace split] that
>> would return a list of parts?
>>
>> For example, I'd like namespace split a::b::c to return {a b c}, and
>> namespace split ::d::e::f::g to return {d e f g}.
> 
> Easy combining [namespace tail] and [namespace qualifiers]:
> 
> proc ns2list {ns} {
>    set l {}
>    while {$ns ne {}} {
>      lappend l [namespace tail $ns]
>      set ns [namespace qualifiers $ns]
>    }
>    lreverse $l
> }
> 
> % ns2list foo::bar::
> foo bar {}
> % ns2list foo::bar
> foo bar
> % ns2list ::::foo::bar::
> foo bar {}
> % ns2list ::foo::bar
> foo bar
> % ns2list {}
> %
> 
>>
>> As a followup, if I write a proc namespaceSplit that does what I want,
>> is there a user-level way to modify the [namespace] command so that a
>> can execute namespaceSplit via [namespace split]?
> 
> Since the [namespace] command is an ensemble, its easy to add a subcommand
> 
> % namespace ensemble configure namespace -map [dict merge [namespace ensemble configure namespace -map] {split ::ns2list}]
> % namespace split ::foo::bar::baz
> foo bar baz
> 
>>
>> Many thanks
> 
> Regards
Thanks Emiliano.

I did some testing after I posted last night, and I found that 
manipulating the name with [string] commands ran faster than with 
[namespace tail] and [namespace qualifiers]. I must admit that when 
using the [namespace] subcommands, I was processing the list with 
[linsert l 0 ...] rather than using [lappend l] and [lreverse] which 
probably makes a difference.

Thanks also for the pointer about adding a new subcommand to the 
namespace ensemble. I've never played with [namespace ensemble] before 
and I can see why I was a little intimidated.

-- 

Alan