Re: Minor backend fr.po mistake

Hi,

#, c-format
msgid "This may be because of a non-immutable index expression."
msgstr "Ceci peut être dû à une expression d'index immutable."

Should be "NON immutable".

(Or non immuable, or is that just muable... Just kidding but I am a
bit confused about why "immutable" is preserved in English. In this
case it's referring to observed behaviour of an expression that must
have been considered immutable by static analysis but seems not to be
at execution time. In another case we use the French word "immuable"
(no T) when talking about immutability by static analysis (that is,
whether all functions are declared IMMUTABLE). Other messages and
languages made various choices about when to translate, quote or
uppercase the words immutable and volatile. Wow, this is hard.)

Hi,

Le sam. 30 nov. 2024 à 05:15, Thomas Munro <thomas(dot)munro(at)gmail(dot)com> a
écrit :

> Hi,
>
> #, c-format
> msgid "This may be because of a non-immutable index expression."
> msgstr "Ceci peut être dû à une expression d'index immutable."
>
> Should be "NON immutable".
>
>
You're right, my mistake. I've fixed it on all maintained releases.

> (Or non immuable, or is that just muable... Just kidding but I am a
> bit confused about why "immutable" is preserved in English. In this
> case it's referring to observed behaviour of an expression that must
> have been considered immutable by static analysis but seems not to be
> at execution time. In another case we use the French word "immuable"
> (no T) when talking about immutability by static analysis (that is,
> whether all functions are declared IMMUTABLE). Other messages and
> languages made various choices about when to translate, quote or
> uppercase the words immutable and volatile. Wow, this is hard.)
>
>
Immutable and volatile exist in French (though the right word should be
"immuable", but usually French DBAs know what immutable is).

Thanks for your report.

--
Guillaume.

On 2024-Nov-30, Thomas Munro wrote:

> #, c-format
> msgid "This may be because of a non-immutable index expression."
> msgstr "Ceci peut être dû à une expression d'index immutable."
>
> Should be "NON immutable".

Agreed.

> (Or non immuable, or is that just muable... Just kidding but I am a
> bit confused about why "immutable" is preserved in English. In this
> case it's referring to observed behaviour of an expression that must
> have been considered immutable by static analysis but seems not to be
> at execution time. In another case we use the French word "immuable"
> (no T) when talking about immutability by static analysis (that is,
> whether all functions are declared IMMUTABLE). Other messages and
> languages made various choices about when to translate, quote or
> uppercase the words immutable and volatile. Wow, this is hard.)

Yeah, it is hard. Sometimes we keep the English term because for the
user it's clearer exactly which property is meant. For instance, in the
Spanish translation I used to translate "locks" as "candados" and
"triggers" as "disparadores", but per user feedback I ended reverting
that into using the English words in quotes. Examples:

msgid "permission denied for event trigger %s"
msgstr "permiso denegado al “trigger” por eventos %s"

msgid "could not obtain lock on relation \"%s.%s\""
msgstr "no se pudo bloquear un “lock” en la relación «%s.%s»"

(This is quite new and has not yet propagated to 17, but it will by the
February release). I think users understand this better.

Anyway I see that the French translation tends to use "immutable"
everywhere rather than "immuable". It might be just a typo, or somebody
was not consistent.

--
Álvaro Herrera 48°01'N 7°57'E — https://www.EnterpriseDB.com/

Le sam. 30 nov. 2024 à 15:00, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> a
écrit :

> On 2024-Nov-30, Thomas Munro wrote:
>
> > #, c-format
> > msgid "This may be because of a non-immutable index expression."
> > msgstr "Ceci peut être dû à une expression d'index immutable."
> >
> > Should be "NON immutable".
>
> Agreed.
>
> > (Or non immuable, or is that just muable... Just kidding but I am a
> > bit confused about why "immutable" is preserved in English. In this
> > case it's referring to observed behaviour of an expression that must
> > have been considered immutable by static analysis but seems not to be
> > at execution time. In another case we use the French word "immuable"
> > (no T) when talking about immutability by static analysis (that is,
> > whether all functions are declared IMMUTABLE). Other messages and
> > languages made various choices about when to translate, quote or
> > uppercase the words immutable and volatile. Wow, this is hard.)
>
> Yeah, it is hard. Sometimes we keep the English term because for the
> user it's clearer exactly which property is meant. For instance, in the
> Spanish translation I used to translate "locks" as "candados" and
> "triggers" as "disparadores", but per user feedback I ended reverting
> that into using the English words in quotes. Examples:
>
> msgid "permission denied for event trigger %s"
> msgstr "permiso denegado al “trigger” por eventos %s"
>
> msgid "could not obtain lock on relation \"%s.%s\""
> msgstr "no se pudo bloquear un “lock” en la relación «%s.%s»"
>
> (This is quite new and has not yet propagated to 17, but it will by the
> February release). I think users understand this better.
>
>
While I still translate lock in french ("verrou"), I don't translate
trigger anymore. Noone understood what I meant with "déclencheur". And
every DBA, even French ones :) , knows what a trigger is.

> Anyway I see that the French translation tends to use "immutable"
> everywhere rather than "immuable". It might be just a typo, or somebody
> was not consistent.
>
>
I tend to stick with "immutable", as French people understand this term.
Maybe there's still some "immuable", I'll try to fix them.

--
Guillaume.

On Sun, Dec 1, 2024 at 3:09 AM Guillaume Lelarge <guillaume(at)lelarge(dot)info> wrote:
> Noone understood what I meant with "déclencheur". And every DBA, even French ones :) , knows what a trigger is.

Right, they had to type CREATE TRIGGER to get there. For the names of
database objects and other things in the syntax of the language that
you have to use to talk to the computer, it makes sense that some
translators decide that it should talk back to you in the same way.
It's murkier for the more descriptive stuff like mutability. (I
wonder what would happen if the message required you decide if the
foreign word used as an adjective is variable, but by coincidence you
can't tell in any of the current messages, the only ones that come
close used a different formulation "marked/marquées/marcadas
IMMUTABLE" :-))

Anyway, thanks for the discussion!

On 01.12.24 02:24, Thomas Munro wrote:
> On Sun, Dec 1, 2024 at 3:09 AM Guillaume Lelarge <guillaume(at)lelarge(dot)info> wrote:
>> Noone understood what I meant with "déclencheur". And every DBA, even French ones :) , knows what a trigger is.
>
> Right, they had to type CREATE TRIGGER to get there. For the names of
> database objects and other things in the syntax of the language that
> you have to use to talk to the computer, it makes sense that some
> translators decide that it should talk back to you in the same way.
> It's murkier for the more descriptive stuff like mutability. (I
> wonder what would happen if the message required you decide if the
> foreign word used as an adjective is variable, but by coincidence you
> can't tell in any of the current messages, the only ones that come
> close used a different formulation "marked/marquées/marcadas
> IMMUTABLE" :-))

I find the Microsoft Terminology Search a good resource for translations
of technical terms:
https://learn.microsoft.com/en-us/globalization/reference/microsoft-terminology