Fix regression in up

[aanand]

When running docker-compose up A B, logs are shown for those services and all their dependencies, which is not what the user expects. This regressed in #1645.

Not wild about the variable names here - would appreciate a second pair of eyes.

Closes#1856.

[dnephin]

I'm not sure about this. I think project.up() should still return all the containers it created.

I would think the filtering should happen in the cli.main.up() (or it can call a function to do the filtering) since that's the layer that deals with user expectations.

I think it's conceivable that something else might want to know all the containers created by up (I say this because I've actually written code that has that expectation).

[aanand]

Fair enough - only problem with doing it in cli.main.up is that it can't be tested.

[aanand]

(easily)

[dnephin]

Ya, that's true. At some point I think it might make sense to split up some of the larger cli functions into some smaller functions (to make it easier to test), but that doesn't need to happen now.

In this case I think everything that happens after project.up() could be in a separate function:

ifdetached:
returnprint_logs_for_up(containers, service_names, monochrome, timeout)

[dnephin]

Closed by #1931

[aanand]

When running docker-compose up A B, logs are shown for those services and all their dependencies, which is not what the user expects. This regressed in #1645.

Not wild about the variable names here - would appreciate a second pair of eyes.

Closes#1856.

[dnephin]

I'm not sure about this. I think project.up() should still return all the containers it created.

I would think the filtering should happen in the cli.main.up() (or it can call a function to do the filtering) since that's the layer that deals with user expectations.

I think it's conceivable that something else might want to know all the containers created by up (I say this because I've actually written code that has that expectation).

[aanand]

Fair enough - only problem with doing it in cli.main.up is that it can't be tested.

[aanand]

(easily)

[dnephin]

Ya, that's true. At some point I think it might make sense to split up some of the larger cli functions into some smaller functions (to make it easier to test), but that doesn't need to happen now.

In this case I think everything that happens after project.up() could be in a separate function:

ifdetached:
returnprint_logs_for_up(containers, service_names, monochrome, timeout)

[dnephin]

Closed by #1931

[aanand]

When running docker-compose up A B, logs are shown for those services and all their dependencies, which is not what the user expects. This regressed in #1645.

Not wild about the variable names here - would appreciate a second pair of eyes.

Closes#1856.

[dnephin]

I'm not sure about this. I think project.up() should still return all the containers it created.

I would think the filtering should happen in the cli.main.up() (or it can call a function to do the filtering) since that's the layer that deals with user expectations.

I think it's conceivable that something else might want to know all the containers created by up (I say this because I've actually written code that has that expectation).

[aanand]

Fair enough - only problem with doing it in cli.main.up is that it can't be tested.

[aanand]

(easily)

[dnephin]

Ya, that's true. At some point I think it might make sense to split up some of the larger cli functions into some smaller functions (to make it easier to test), but that doesn't need to happen now.

In this case I think everything that happens after project.up() could be in a separate function:

ifdetached:
returnprint_logs_for_up(containers, service_names, monochrome, timeout)

[dnephin]

Closed by #1931

[aanand]

When running docker-compose up A B, logs are shown for those services and all their dependencies, which is not what the user expects. This regressed in #1645.

Not wild about the variable names here - would appreciate a second pair of eyes.

Closes#1856.

[dnephin]

I'm not sure about this. I think project.up() should still return all the containers it created.

I would think the filtering should happen in the cli.main.up() (or it can call a function to do the filtering) since that's the layer that deals with user expectations.

I think it's conceivable that something else might want to know all the containers created by up (I say this because I've actually written code that has that expectation).

[aanand]

Fair enough - only problem with doing it in cli.main.up is that it can't be tested.

[aanand]

(easily)

[dnephin]

Ya, that's true. At some point I think it might make sense to split up some of the larger cli functions into some smaller functions (to make it easier to test), but that doesn't need to happen now.

In this case I think everything that happens after project.up() could be in a separate function:

ifdetached:
returnprint_logs_for_up(containers, service_names, monochrome, timeout)

[dnephin]

Closed by #1931

[aanand]

When running docker-compose up A B, logs are shown for those services and all their dependencies, which is not what the user expects. This regressed in #1645.

Not wild about the variable names here - would appreciate a second pair of eyes.

Closes#1856.

[dnephin]

I'm not sure about this. I think project.up() should still return all the containers it created.

I would think the filtering should happen in the cli.main.up() (or it can call a function to do the filtering) since that's the layer that deals with user expectations.

I think it's conceivable that something else might want to know all the containers created by up (I say this because I've actually written code that has that expectation).

[aanand]

Fair enough - only problem with doing it in cli.main.up is that it can't be tested.

[aanand]

(easily)

[dnephin]

Ya, that's true. At some point I think it might make sense to split up some of the larger cli functions into some smaller functions (to make it easier to test), but that doesn't need to happen now.

In this case I think everything that happens after project.up() could be in a separate function:

ifdetached:
returnprint_logs_for_up(containers, service_names, monochrome, timeout)

[dnephin]

Closed by #1931

[aanand]

When running docker-compose up A B, logs are shown for those services and all their dependencies, which is not what the user expects. This regressed in #1645.

Not wild about the variable names here - would appreciate a second pair of eyes.

Closes#1856.

[dnephin]

I'm not sure about this. I think project.up() should still return all the containers it created.

I would think the filtering should happen in the cli.main.up() (or it can call a function to do the filtering) since that's the layer that deals with user expectations.

I think it's conceivable that something else might want to know all the containers created by up (I say this because I've actually written code that has that expectation).

[aanand]

Fair enough - only problem with doing it in cli.main.up is that it can't be tested.

[aanand]

(easily)

[dnephin]

Ya, that's true. At some point I think it might make sense to split up some of the larger cli functions into some smaller functions (to make it easier to test), but that doesn't need to happen now.

In this case I think everything that happens after project.up() could be in a separate function:

ifdetached:
returnprint_logs_for_up(containers, service_names, monochrome, timeout)

[dnephin]

Closed by #1931

[aanand]

When running docker-compose up A B, logs are shown for those services and all their dependencies, which is not what the user expects. This regressed in #1645.

Not wild about the variable names here - would appreciate a second pair of eyes.

Closes#1856.

[dnephin]

I'm not sure about this. I think project.up() should still return all the containers it created.

I would think the filtering should happen in the cli.main.up() (or it can call a function to do the filtering) since that's the layer that deals with user expectations.

I think it's conceivable that something else might want to know all the containers created by up (I say this because I've actually written code that has that expectation).

[aanand]

Fair enough - only problem with doing it in cli.main.up is that it can't be tested.

[aanand]

(easily)

[dnephin]

Ya, that's true. At some point I think it might make sense to split up some of the larger cli functions into some smaller functions (to make it easier to test), but that doesn't need to happen now.

In this case I think everything that happens after project.up() could be in a separate function:

ifdetached:
returnprint_logs_for_up(containers, service_names, monochrome, timeout)

[dnephin]

Closed by #1931

[aanand]

When running docker-compose up A B, logs are shown for those services and all their dependencies, which is not what the user expects. This regressed in #1645.

Not wild about the variable names here - would appreciate a second pair of eyes.

Closes#1856.

[dnephin]

I'm not sure about this. I think project.up() should still return all the containers it created.

I would think the filtering should happen in the cli.main.up() (or it can call a function to do the filtering) since that's the layer that deals with user expectations.

I think it's conceivable that something else might want to know all the containers created by up (I say this because I've actually written code that has that expectation).

[aanand]

Fair enough - only problem with doing it in cli.main.up is that it can't be tested.

[aanand]

(easily)

[dnephin]

Ya, that's true. At some point I think it might make sense to split up some of the larger cli functions into some smaller functions (to make it easier to test), but that doesn't need to happen now.

In this case I think everything that happens after project.up() could be in a separate function:

ifdetached:
returnprint_logs_for_up(containers, service_names, monochrome, timeout)

[dnephin]

Closed by #1931