Commented by: Sean Leyne (seanleyne)

NBackup already has this function, via the lock/unlock option.

See https://firebirdsql.org/manual/nbackup-lock-unlock.html

Commented by: Arioch (arioch)

No it does not have "this function"

It is three commands instead of one and it is still less intuitive.

Commented by: @hvlad

You don't have to "restore" backup of level 0, if you need a database copy instead of backup.

Just "fix" it: https://firebirdsql.org/manual/nbackup-lock-unlock.html#nbackup-restore-and-fixup

Commented by: Arioch (arioch)

The "fix" is described in both your and Sean's docs as following -L operation not -B 0 one.

And it is still less convinient and less simple

Commented by: @hvlad

> The "fix" is described in both your and Sean's docs as following -L operation not -B 0 one.
If you read the docs you should understand that "backup" of level zero is the copy of locked database.
So, you could just "fix" it instead of "restore".

Commented by: @dyemanov

Maybe it's worth allowing both -b 0 and -f together? Or some other modifier for -b 0 to fixup the resulting copy automatically.

Commented by: Arioch (arioch)

> If you read the docs you should understand

There are two problems with this claim.

1. Is user really expected to be deducing from implications, rather than reading explicitly and clearly stated things? If so, why need documentation at all? Sources are the ultimate documentations for themselves. User might read sources and should understand, etc.

2. That seems to be "implementation details" as opposed to "code by contract".
Is it clearly and explicitly stated in documentation that "-b 0" is the same as "-l" + copy + "-u", and, importantly, that it WON'T CHANGE in foreseeble future?

If not, then you as developer can use implementation details in production, but users can not.
You would be among first to blame users for any screw up caused by their cutting corners and using historical features that were not explicitly promised in documentation. You would be among first to claim that users' habit of using undocumented things which "always worked so for years" is holding you back and prohibits you from enhancing Firebird, which would require changing habitual side effects of legacy implementation.

Introducign such a switch would serve both makling it intuitive and obvious AND making it the contract users can rely upon and developers knowing not to break without BIG RED WARNING in all the docs around.

Commented by: @hvlad

Dmitry,

I don't think that adding modifier for -b 0 is good idea. It will not make things simpler. It will not work with -b 1, for example.
We could add special new switch to do {-b 0, -f} at one step but I see no value in it.
It adds nothing, except maybe additional confusion for end user.

Commented by: @hvlad

Arioch,

you never listen to others arguments. There is only one correct opinion - yours.
I see no sense to continue this useless discussion.

Commented by: Arioch (arioch)

It is half the problem if i can't hear your arguments.
It is much worse when you don't hear your own arguments.

Just today you said : "Ибо не надо закладываться на чужое внутреннее поведение, которое может и будет меняться." - "Because one should never found his actions upon others implementation derails which can and eventually would get changed".
https://www.sql.ru/forum/actualutils.aspx?action=gotomsg&tid=1321312&msg=22061422

That is exactly the case with "-b 0 is the same as -l".
Unless that is explicitly stated in documentation as a contract that Firebird Team pledges not to violate - this can not be relied upon by users because it is exactly implementatio ndetail that must remain volatile and arbitrary..

Commented by: @hvlad

1. The cite you show above is abosilutely different case.
2. Nobody never said that "-b 0" is the same as "-l". Both have some common parts, but thay are not the same. You should be more attentive.
3. Ask to update documentation, if you consider it incomplete.

This ticket is closed, accept it and keep calm.