From Newsgroup: comp.lang.tcl

Is using `lmap` the best way to copy a list? For example:

package require struct::list 1
set a {a bc def ghij klmno}
# is the following the best way to copy a list?
set b [lmap x $a {expr {$x}}]
puts "a={$a} b={$b} a==b=[struct::list equal $a $b]"

Output:

a={a bc def ghij klmno} b={a bc def ghij klmno} a==b=1
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

Am 18.06.2025 um 09:19 schrieb Mark Summerfield:

Is using `lmap` the best way to copy a list? For example:

package require struct::list 1
set a {a bc def ghij klmno}
# is the following the best way to copy a list?
set b [lmap x $a {expr {$x}}]
puts "a={$a} b={$b} a==b=[struct::list equal $a $b]"

Output:

a={a bc def ghij klmno} b={a bc def ghij klmno} a==b=1

Stupid question:

set b $a

to copy list b to a ?


--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

Not sure I understand the question. Why not just

set b $a

?

Also, your code will not always give identical results using
expr ([expr {$x}] is not always $x). For example,

% set a {a 0x10 b}
a 0x10 b
% set b [lmap x $a {expr {$x}}]
a 16 b

You could use one of the following forms instead

% set b [lmap x $a {lindex $x}]
a 0x10 b
% set b [lmap x $a {string cat $x}]
a 0x10 b
% set b [lmap x $a {return -level 0 $x}]
a 0x10 b

but as I said, why not just set a $b ?

/Ashok

set a {a 0x10 b}
On 6/18/2025 12:49 PM, Mark Summerfield wrote:

Is using `lmap` the best way to copy a list? For example:

package require struct::list 1
set a {a bc def ghij klmno}
# is the following the best way to copy a list?
set b [lmap x $a {expr {$x}}]
puts "a={$a} b={$b} a==b=[struct::list equal $a $b]"

Output:

a={a bc def ghij klmno} b={a bc def ghij klmno} a==b=1

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

On Wed, 18 Jun 2025 17:03:36 +0530, Ashok wrote:

Not sure I understand the question. Why not just

set b $a

?

Also, your code will not always give identical results using expr ([expr {$x}] is not always $x). For example,

% set a {a 0x10 b}
a 0x10 b % set b [lmap x $a {expr {$x}}]
a 16 b

You could use one of the following forms instead

% set b [lmap x $a {lindex $x}]
a 0x10 b % set b [lmap x $a {string cat $x}]
a 0x10 b % set b [lmap x $a {return -level 0 $x}]
a 0x10 b

but as I said, why not just set a $b ?

/Ashok

set a {a 0x10 b}
On 6/18/2025 12:49 PM, Mark Summerfield wrote:

Is using `lmap` the best way to copy a list? For example:

package require struct::list 1 set a {a bc def ghij klmno}
# is the following the best way to copy a list?
set b [lmap x $a {expr {$x}}]
puts "a={$a} b={$b} a==b=[struct::list equal $a $b]"

Output:

a={a bc def ghij klmno} b={a bc def ghij klmno} a==b=1

I'm coming to Tcl from Python & Go. In Python saying `a = [1, 2, 3]` and
then `b = a` makes `b` a _reference_ to `a` rather than an independent variable. Now I realise that Tcl doesn't do it that way and that `set a
$b` works fine.
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

Am 18.06.2025 um 13:59 schrieb Mark Summerfield:

I'm coming to Tcl from Python & Go. In Python saying `a = [1, 2, 3]` and
then `b = a` makes `b` a _reference_ to `a` rather than an independent variable. Now I realise that Tcl doesn't do it that way and that `set a
$b` works fine.

Well, internally you also get only a reference, in the sense, that the reference count of the data is incremented instead of copied.
But if you modify b, you will not modify a, as then, a copy will be done.

% set a {a bc def ghij klmno} ; ::tcl::unsupported::representation $a
value is a pure string with a refcount of 4, object pointer at
0x239caf217d0, string representation "a bc def ghij..."
% set b $a ; ::tcl::unsupported::representation $b
value is a pure string with a refcount of 5, object pointer at
0x239caf217d0, string representation "a bc def ghij..."
% lset b 0 b;::tcl::unsupported::representation $b
value is a list with a refcount of 2, object pointer at 0x239cc358760, internal representation 0x239ccac9410:0x0, no string representation

You see, that:
set b $a: only recount incremented, a and b have same data address
lset b 0 b: now, a copy is made, as b is different to a.

Sorry, could not resist to copy satupid internals...
Harald
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

Mark Summerfield <m.n.summerfield@gmail.com> posted:

I'm coming to Tcl from Python & Go. In Python saying `a = [1, 2, 3]` and then `b = a` makes `b` a _reference_ to `a` rather than an independent variable. Now I realise that Tcl doesn't do it that way and that `set a
$b` works fine.

When I had to do some Python for work, after being familiar with Tcl,
I was quite shocked to find that Python shares values in that way! 😲
--
Colin Macleod ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ https://cmacleod.me.uk

Warning: Gumption level low, top-up when possible!
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

Am 18.06.25 um 13:59 schrieb Mark Summerfield:

I'm coming to Tcl from Python & Go. In Python saying `a = [1, 2, 3]` and
then `b = a` makes `b` a _reference_ to `a` rather than an independent variable. Now I realise that Tcl doesn't do it that way and that `set a
$b` works fine.

I suspected something like this, when reading your question. In Python
terms, Tcl has only immutable values. Whereas in Python, the outcome of
a[2] = 'x' depends on what exact type this "a" thing is, in Tcl it is
either impossible or it creates a copy on write (lset).

This is part of the bigger "EIAS" concept ("everything is a string")
which also means that values can be type-casted automatically - however,
in good practice Tcl, you try to stick with a type nonetheless.

Christian
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

On 6/18/2025 5:19 AM, Colin Macleod wrote:

Mark Summerfield <m.n.summerfield@gmail.com> posted:

I'm coming to Tcl from Python & Go. In Python saying `a = [1, 2, 3]` and
then `b = a` makes `b` a _reference_ to `a` rather than an independent
variable. Now I realise that Tcl doesn't do it that way and that `set a
$b` works fine.

When I had to do some Python for work, after being familiar with Tcl,
I was quite shocked to find that Python shares values in that way! 😲

Me too!

This python feature goes against my favorite rule of language design:

"The principle of least astonishment"

I'd have to say that python gets a low score on that one!

especially how with a=b, you don't even know the type (and thus immutability) of b without looking elsewhere, which can lead to some nasty runtime surprises.


--- Synchronet 3.21a-Linux NewsLink 1.2