Add --path option to 'rustup override set'

[pickfire]

Same as --path in 'rustup override unset'

[kinnison]

Hi @pickfire

Are you still interested in persuing this change? If so, you need to rebase it against the current master (We've had a bit of a reformat of code so it's conflicting). Also it'd be nice if you could add a test to cover the new --path functionality.

[pickfire]

@kinnison I am interested but I thought rustup.rs might not accept the changes since it's kinda not accepting changes last time IIRC.

If the patch could be accepted (I could fix the stuff), I am still interested in this but probably need some time since I've been a bit busy lately.

[kinnison]

I think it's fair to say that merges will not be fast into this repo because the maintainers are busy elsewhere a lot of the time, but I at least am interested in sorting improvements out, and @nrc is able to look in and merge stuff if it's clean and clear :-D I guess step one would be to rebase onto current master and organise a test or two. Once that's done, I'm prepared to help get it mergeable if @nrc wants anything else doing first.

[bors]

☔ The latest upstream changes (presumably #1576) made this pull request unmergeable. Please resolve the merge conflicts.

[dwijnand]

Same as --path in 'rustup override unset'

If this is fixing feature parity then I'd expect it to be accepted. I agree it would be good for this to come with a test (as well as needing another rebase.. 😕)

[pickfire]

I have rebased against master but not sure what to test though, the current unset does not have tests for path as well.

[kinnison]

Thanks for the rebase, and I'm glad to see the green tick. I'll give some thought to what testing might be appropriate, given you say there's also a gap in the override unset testing too.

[bors]

☔ The latest upstream changes (presumably #1704) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

@kinnison What do I do now? Resolve the merge conflict again?

[kinnison]

I've been unable to give any time to thinking about this particular PR, which is sad. Re-reading it I'd say that we agreed on what was missing (tests) but not on what those tests would look like or who would be responsible for them. If you lack time to speculatively address these issues then please just let the PR sit for now.

It's possible someone else will come along who wants to work on Rustup and has the time to work it through. A rebase ought not to be too painful (though Dale has rearranged the codebase now so it's not going to be entirely trivial), but the tests remain the sticking point.

[pickfire]

@dwijnand Thanks a lot for helping out.

@kinnison I have no issues rebasing, I am just not sure what tests to write for this.

[kinnison]

Regarding tests, the main thing is that we should verify that rustup override {un,}set --path ... works correctly. This probably means creating a test which sets up a directory with an override, and then verifies that it works, and then unsets it and verifies the override is gone.

However, if you're really stuck with how to do that, just hold the rebased branch for now until someone can help you with the tests. Sadly I don't have enough time right now to provide an example.

[bors]

☔ The latest upstream changes (presumably #1754) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

I added test for both set and unset only for --path.

[kinnison]

Hi @pickfire

Thanks for updating the branch, but the Windows test (Appveyor) failed. From a glance it looks like you maybe hard-coded an architecture name in your test? I will try and review the code later to help you out if you can't spot the bug.

D.

[pickfire]

@kinnison Could you please help me to check what to change? I am not so rich to buy a windows laptop, I don't have one to test it out (as well I am not very familiar with windows).

[kinnison]

Other than the test issue, this LGTM.

[pickfire]

Same as --path in 'rustup override unset'

[kinnison]

Hi @pickfire

Are you still interested in persuing this change? If so, you need to rebase it against the current master (We've had a bit of a reformat of code so it's conflicting). Also it'd be nice if you could add a test to cover the new --path functionality.

[pickfire]

@kinnison I am interested but I thought rustup.rs might not accept the changes since it's kinda not accepting changes last time IIRC.

If the patch could be accepted (I could fix the stuff), I am still interested in this but probably need some time since I've been a bit busy lately.

[kinnison]

I think it's fair to say that merges will not be fast into this repo because the maintainers are busy elsewhere a lot of the time, but I at least am interested in sorting improvements out, and @nrc is able to look in and merge stuff if it's clean and clear :-D I guess step one would be to rebase onto current master and organise a test or two. Once that's done, I'm prepared to help get it mergeable if @nrc wants anything else doing first.

[bors]

☔ The latest upstream changes (presumably #1576) made this pull request unmergeable. Please resolve the merge conflicts.

[dwijnand]

Same as --path in 'rustup override unset'

If this is fixing feature parity then I'd expect it to be accepted. I agree it would be good for this to come with a test (as well as needing another rebase.. 😕)

[pickfire]

I have rebased against master but not sure what to test though, the current unset does not have tests for path as well.

[kinnison]

Thanks for the rebase, and I'm glad to see the green tick. I'll give some thought to what testing might be appropriate, given you say there's also a gap in the override unset testing too.

[bors]

☔ The latest upstream changes (presumably #1704) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

@kinnison What do I do now? Resolve the merge conflict again?

[kinnison]

I've been unable to give any time to thinking about this particular PR, which is sad. Re-reading it I'd say that we agreed on what was missing (tests) but not on what those tests would look like or who would be responsible for them. If you lack time to speculatively address these issues then please just let the PR sit for now.

It's possible someone else will come along who wants to work on Rustup and has the time to work it through. A rebase ought not to be too painful (though Dale has rearranged the codebase now so it's not going to be entirely trivial), but the tests remain the sticking point.

[pickfire]

@dwijnand Thanks a lot for helping out.

@kinnison I have no issues rebasing, I am just not sure what tests to write for this.

[kinnison]

Regarding tests, the main thing is that we should verify that rustup override {un,}set --path ... works correctly. This probably means creating a test which sets up a directory with an override, and then verifies that it works, and then unsets it and verifies the override is gone.

However, if you're really stuck with how to do that, just hold the rebased branch for now until someone can help you with the tests. Sadly I don't have enough time right now to provide an example.

[bors]

☔ The latest upstream changes (presumably #1754) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

I added test for both set and unset only for --path.

[kinnison]

Hi @pickfire

Thanks for updating the branch, but the Windows test (Appveyor) failed. From a glance it looks like you maybe hard-coded an architecture name in your test? I will try and review the code later to help you out if you can't spot the bug.

D.

[pickfire]

@kinnison Could you please help me to check what to change? I am not so rich to buy a windows laptop, I don't have one to test it out (as well I am not very familiar with windows).

[kinnison]

Other than the test issue, this LGTM.

[pickfire]

Same as --path in 'rustup override unset'

[kinnison]

Hi @pickfire

Are you still interested in persuing this change? If so, you need to rebase it against the current master (We've had a bit of a reformat of code so it's conflicting). Also it'd be nice if you could add a test to cover the new --path functionality.

[pickfire]

@kinnison I am interested but I thought rustup.rs might not accept the changes since it's kinda not accepting changes last time IIRC.

If the patch could be accepted (I could fix the stuff), I am still interested in this but probably need some time since I've been a bit busy lately.

[kinnison]

I think it's fair to say that merges will not be fast into this repo because the maintainers are busy elsewhere a lot of the time, but I at least am interested in sorting improvements out, and @nrc is able to look in and merge stuff if it's clean and clear :-D I guess step one would be to rebase onto current master and organise a test or two. Once that's done, I'm prepared to help get it mergeable if @nrc wants anything else doing first.

[bors]

☔ The latest upstream changes (presumably #1576) made this pull request unmergeable. Please resolve the merge conflicts.

[dwijnand]

Same as --path in 'rustup override unset'

If this is fixing feature parity then I'd expect it to be accepted. I agree it would be good for this to come with a test (as well as needing another rebase.. 😕)

[pickfire]

I have rebased against master but not sure what to test though, the current unset does not have tests for path as well.

[kinnison]

Thanks for the rebase, and I'm glad to see the green tick. I'll give some thought to what testing might be appropriate, given you say there's also a gap in the override unset testing too.

[bors]

☔ The latest upstream changes (presumably #1704) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

@kinnison What do I do now? Resolve the merge conflict again?

[kinnison]

I've been unable to give any time to thinking about this particular PR, which is sad. Re-reading it I'd say that we agreed on what was missing (tests) but not on what those tests would look like or who would be responsible for them. If you lack time to speculatively address these issues then please just let the PR sit for now.

It's possible someone else will come along who wants to work on Rustup and has the time to work it through. A rebase ought not to be too painful (though Dale has rearranged the codebase now so it's not going to be entirely trivial), but the tests remain the sticking point.

[pickfire]

@dwijnand Thanks a lot for helping out.

@kinnison I have no issues rebasing, I am just not sure what tests to write for this.

[kinnison]

Regarding tests, the main thing is that we should verify that rustup override {un,}set --path ... works correctly. This probably means creating a test which sets up a directory with an override, and then verifies that it works, and then unsets it and verifies the override is gone.

However, if you're really stuck with how to do that, just hold the rebased branch for now until someone can help you with the tests. Sadly I don't have enough time right now to provide an example.

[bors]

☔ The latest upstream changes (presumably #1754) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

I added test for both set and unset only for --path.

[kinnison]

Hi @pickfire

Thanks for updating the branch, but the Windows test (Appveyor) failed. From a glance it looks like you maybe hard-coded an architecture name in your test? I will try and review the code later to help you out if you can't spot the bug.

D.

[pickfire]

@kinnison Could you please help me to check what to change? I am not so rich to buy a windows laptop, I don't have one to test it out (as well I am not very familiar with windows).

[kinnison]

Other than the test issue, this LGTM.

[pickfire]

Same as --path in 'rustup override unset'

[kinnison]

Hi @pickfire

Are you still interested in persuing this change? If so, you need to rebase it against the current master (We've had a bit of a reformat of code so it's conflicting). Also it'd be nice if you could add a test to cover the new --path functionality.

[pickfire]

@kinnison I am interested but I thought rustup.rs might not accept the changes since it's kinda not accepting changes last time IIRC.

If the patch could be accepted (I could fix the stuff), I am still interested in this but probably need some time since I've been a bit busy lately.

[kinnison]

I think it's fair to say that merges will not be fast into this repo because the maintainers are busy elsewhere a lot of the time, but I at least am interested in sorting improvements out, and @nrc is able to look in and merge stuff if it's clean and clear :-D I guess step one would be to rebase onto current master and organise a test or two. Once that's done, I'm prepared to help get it mergeable if @nrc wants anything else doing first.

[bors]

☔ The latest upstream changes (presumably #1576) made this pull request unmergeable. Please resolve the merge conflicts.

[dwijnand]

Same as --path in 'rustup override unset'

If this is fixing feature parity then I'd expect it to be accepted. I agree it would be good for this to come with a test (as well as needing another rebase.. 😕)

[pickfire]

I have rebased against master but not sure what to test though, the current unset does not have tests for path as well.

[kinnison]

Thanks for the rebase, and I'm glad to see the green tick. I'll give some thought to what testing might be appropriate, given you say there's also a gap in the override unset testing too.

[bors]

☔ The latest upstream changes (presumably #1704) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

@kinnison What do I do now? Resolve the merge conflict again?

[kinnison]

I've been unable to give any time to thinking about this particular PR, which is sad. Re-reading it I'd say that we agreed on what was missing (tests) but not on what those tests would look like or who would be responsible for them. If you lack time to speculatively address these issues then please just let the PR sit for now.

It's possible someone else will come along who wants to work on Rustup and has the time to work it through. A rebase ought not to be too painful (though Dale has rearranged the codebase now so it's not going to be entirely trivial), but the tests remain the sticking point.

[pickfire]

@dwijnand Thanks a lot for helping out.

@kinnison I have no issues rebasing, I am just not sure what tests to write for this.

[kinnison]

Regarding tests, the main thing is that we should verify that rustup override {un,}set --path ... works correctly. This probably means creating a test which sets up a directory with an override, and then verifies that it works, and then unsets it and verifies the override is gone.

However, if you're really stuck with how to do that, just hold the rebased branch for now until someone can help you with the tests. Sadly I don't have enough time right now to provide an example.

[bors]

☔ The latest upstream changes (presumably #1754) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

I added test for both set and unset only for --path.

[kinnison]

Hi @pickfire

Thanks for updating the branch, but the Windows test (Appveyor) failed. From a glance it looks like you maybe hard-coded an architecture name in your test? I will try and review the code later to help you out if you can't spot the bug.

D.

[pickfire]

@kinnison Could you please help me to check what to change? I am not so rich to buy a windows laptop, I don't have one to test it out (as well I am not very familiar with windows).

[kinnison]

Other than the test issue, this LGTM.

[pickfire]

Same as --path in 'rustup override unset'

[kinnison]

Hi @pickfire

Are you still interested in persuing this change? If so, you need to rebase it against the current master (We've had a bit of a reformat of code so it's conflicting). Also it'd be nice if you could add a test to cover the new --path functionality.

[pickfire]

@kinnison I am interested but I thought rustup.rs might not accept the changes since it's kinda not accepting changes last time IIRC.

If the patch could be accepted (I could fix the stuff), I am still interested in this but probably need some time since I've been a bit busy lately.

[kinnison]

I think it's fair to say that merges will not be fast into this repo because the maintainers are busy elsewhere a lot of the time, but I at least am interested in sorting improvements out, and @nrc is able to look in and merge stuff if it's clean and clear :-D I guess step one would be to rebase onto current master and organise a test or two. Once that's done, I'm prepared to help get it mergeable if @nrc wants anything else doing first.

[bors]

☔ The latest upstream changes (presumably #1576) made this pull request unmergeable. Please resolve the merge conflicts.

[dwijnand]

Same as --path in 'rustup override unset'

If this is fixing feature parity then I'd expect it to be accepted. I agree it would be good for this to come with a test (as well as needing another rebase.. 😕)

[pickfire]

I have rebased against master but not sure what to test though, the current unset does not have tests for path as well.

[kinnison]

Thanks for the rebase, and I'm glad to see the green tick. I'll give some thought to what testing might be appropriate, given you say there's also a gap in the override unset testing too.

[bors]

☔ The latest upstream changes (presumably #1704) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

@kinnison What do I do now? Resolve the merge conflict again?

[kinnison]

I've been unable to give any time to thinking about this particular PR, which is sad. Re-reading it I'd say that we agreed on what was missing (tests) but not on what those tests would look like or who would be responsible for them. If you lack time to speculatively address these issues then please just let the PR sit for now.

It's possible someone else will come along who wants to work on Rustup and has the time to work it through. A rebase ought not to be too painful (though Dale has rearranged the codebase now so it's not going to be entirely trivial), but the tests remain the sticking point.

[pickfire]

@dwijnand Thanks a lot for helping out.

@kinnison I have no issues rebasing, I am just not sure what tests to write for this.

[kinnison]

Regarding tests, the main thing is that we should verify that rustup override {un,}set --path ... works correctly. This probably means creating a test which sets up a directory with an override, and then verifies that it works, and then unsets it and verifies the override is gone.

However, if you're really stuck with how to do that, just hold the rebased branch for now until someone can help you with the tests. Sadly I don't have enough time right now to provide an example.

[bors]

☔ The latest upstream changes (presumably #1754) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

I added test for both set and unset only for --path.

[kinnison]

Hi @pickfire

Thanks for updating the branch, but the Windows test (Appveyor) failed. From a glance it looks like you maybe hard-coded an architecture name in your test? I will try and review the code later to help you out if you can't spot the bug.

D.

[pickfire]

@kinnison Could you please help me to check what to change? I am not so rich to buy a windows laptop, I don't have one to test it out (as well I am not very familiar with windows).

[kinnison]

Other than the test issue, this LGTM.

[pickfire]

Same as --path in 'rustup override unset'

[kinnison]

Hi @pickfire

Are you still interested in persuing this change? If so, you need to rebase it against the current master (We've had a bit of a reformat of code so it's conflicting). Also it'd be nice if you could add a test to cover the new --path functionality.

[pickfire]

@kinnison I am interested but I thought rustup.rs might not accept the changes since it's kinda not accepting changes last time IIRC.

If the patch could be accepted (I could fix the stuff), I am still interested in this but probably need some time since I've been a bit busy lately.

[kinnison]

I think it's fair to say that merges will not be fast into this repo because the maintainers are busy elsewhere a lot of the time, but I at least am interested in sorting improvements out, and @nrc is able to look in and merge stuff if it's clean and clear :-D I guess step one would be to rebase onto current master and organise a test or two. Once that's done, I'm prepared to help get it mergeable if @nrc wants anything else doing first.

[bors]

☔ The latest upstream changes (presumably #1576) made this pull request unmergeable. Please resolve the merge conflicts.

[dwijnand]

Same as --path in 'rustup override unset'

If this is fixing feature parity then I'd expect it to be accepted. I agree it would be good for this to come with a test (as well as needing another rebase.. 😕)

[pickfire]

I have rebased against master but not sure what to test though, the current unset does not have tests for path as well.

[kinnison]

Thanks for the rebase, and I'm glad to see the green tick. I'll give some thought to what testing might be appropriate, given you say there's also a gap in the override unset testing too.

[bors]

☔ The latest upstream changes (presumably #1704) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

@kinnison What do I do now? Resolve the merge conflict again?

[kinnison]

I've been unable to give any time to thinking about this particular PR, which is sad. Re-reading it I'd say that we agreed on what was missing (tests) but not on what those tests would look like or who would be responsible for them. If you lack time to speculatively address these issues then please just let the PR sit for now.

It's possible someone else will come along who wants to work on Rustup and has the time to work it through. A rebase ought not to be too painful (though Dale has rearranged the codebase now so it's not going to be entirely trivial), but the tests remain the sticking point.

[pickfire]

@dwijnand Thanks a lot for helping out.

@kinnison I have no issues rebasing, I am just not sure what tests to write for this.

[kinnison]

Regarding tests, the main thing is that we should verify that rustup override {un,}set --path ... works correctly. This probably means creating a test which sets up a directory with an override, and then verifies that it works, and then unsets it and verifies the override is gone.

However, if you're really stuck with how to do that, just hold the rebased branch for now until someone can help you with the tests. Sadly I don't have enough time right now to provide an example.

[bors]

☔ The latest upstream changes (presumably #1754) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

I added test for both set and unset only for --path.

[kinnison]

Hi @pickfire

Thanks for updating the branch, but the Windows test (Appveyor) failed. From a glance it looks like you maybe hard-coded an architecture name in your test? I will try and review the code later to help you out if you can't spot the bug.

D.

[pickfire]

@kinnison Could you please help me to check what to change? I am not so rich to buy a windows laptop, I don't have one to test it out (as well I am not very familiar with windows).

[kinnison]

Other than the test issue, this LGTM.

[pickfire]

Same as --path in 'rustup override unset'

[kinnison]

Hi @pickfire

Are you still interested in persuing this change? If so, you need to rebase it against the current master (We've had a bit of a reformat of code so it's conflicting). Also it'd be nice if you could add a test to cover the new --path functionality.

[pickfire]

@kinnison I am interested but I thought rustup.rs might not accept the changes since it's kinda not accepting changes last time IIRC.

If the patch could be accepted (I could fix the stuff), I am still interested in this but probably need some time since I've been a bit busy lately.

[kinnison]

I think it's fair to say that merges will not be fast into this repo because the maintainers are busy elsewhere a lot of the time, but I at least am interested in sorting improvements out, and @nrc is able to look in and merge stuff if it's clean and clear :-D I guess step one would be to rebase onto current master and organise a test or two. Once that's done, I'm prepared to help get it mergeable if @nrc wants anything else doing first.

[bors]

☔ The latest upstream changes (presumably #1576) made this pull request unmergeable. Please resolve the merge conflicts.

[dwijnand]

Same as --path in 'rustup override unset'

If this is fixing feature parity then I'd expect it to be accepted. I agree it would be good for this to come with a test (as well as needing another rebase.. 😕)

[pickfire]

I have rebased against master but not sure what to test though, the current unset does not have tests for path as well.

[kinnison]

Thanks for the rebase, and I'm glad to see the green tick. I'll give some thought to what testing might be appropriate, given you say there's also a gap in the override unset testing too.

[bors]

☔ The latest upstream changes (presumably #1704) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

@kinnison What do I do now? Resolve the merge conflict again?

[kinnison]

I've been unable to give any time to thinking about this particular PR, which is sad. Re-reading it I'd say that we agreed on what was missing (tests) but not on what those tests would look like or who would be responsible for them. If you lack time to speculatively address these issues then please just let the PR sit for now.

It's possible someone else will come along who wants to work on Rustup and has the time to work it through. A rebase ought not to be too painful (though Dale has rearranged the codebase now so it's not going to be entirely trivial), but the tests remain the sticking point.

[pickfire]

@dwijnand Thanks a lot for helping out.

@kinnison I have no issues rebasing, I am just not sure what tests to write for this.

[kinnison]

Regarding tests, the main thing is that we should verify that rustup override {un,}set --path ... works correctly. This probably means creating a test which sets up a directory with an override, and then verifies that it works, and then unsets it and verifies the override is gone.

However, if you're really stuck with how to do that, just hold the rebased branch for now until someone can help you with the tests. Sadly I don't have enough time right now to provide an example.

[bors]

☔ The latest upstream changes (presumably #1754) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

I added test for both set and unset only for --path.

[kinnison]

Hi @pickfire

Thanks for updating the branch, but the Windows test (Appveyor) failed. From a glance it looks like you maybe hard-coded an architecture name in your test? I will try and review the code later to help you out if you can't spot the bug.

D.

[pickfire]

@kinnison Could you please help me to check what to change? I am not so rich to buy a windows laptop, I don't have one to test it out (as well I am not very familiar with windows).

[kinnison]

Other than the test issue, this LGTM.

[pickfire]

Same as --path in 'rustup override unset'

[kinnison]

Hi @pickfire

Are you still interested in persuing this change? If so, you need to rebase it against the current master (We've had a bit of a reformat of code so it's conflicting). Also it'd be nice if you could add a test to cover the new --path functionality.

[pickfire]

@kinnison I am interested but I thought rustup.rs might not accept the changes since it's kinda not accepting changes last time IIRC.

If the patch could be accepted (I could fix the stuff), I am still interested in this but probably need some time since I've been a bit busy lately.

[kinnison]

I think it's fair to say that merges will not be fast into this repo because the maintainers are busy elsewhere a lot of the time, but I at least am interested in sorting improvements out, and @nrc is able to look in and merge stuff if it's clean and clear :-D I guess step one would be to rebase onto current master and organise a test or two. Once that's done, I'm prepared to help get it mergeable if @nrc wants anything else doing first.

[bors]

☔ The latest upstream changes (presumably #1576) made this pull request unmergeable. Please resolve the merge conflicts.

[dwijnand]

Same as --path in 'rustup override unset'

If this is fixing feature parity then I'd expect it to be accepted. I agree it would be good for this to come with a test (as well as needing another rebase.. 😕)

[pickfire]

I have rebased against master but not sure what to test though, the current unset does not have tests for path as well.

[kinnison]

Thanks for the rebase, and I'm glad to see the green tick. I'll give some thought to what testing might be appropriate, given you say there's also a gap in the override unset testing too.

[bors]

☔ The latest upstream changes (presumably #1704) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

@kinnison What do I do now? Resolve the merge conflict again?

[kinnison]

I've been unable to give any time to thinking about this particular PR, which is sad. Re-reading it I'd say that we agreed on what was missing (tests) but not on what those tests would look like or who would be responsible for them. If you lack time to speculatively address these issues then please just let the PR sit for now.

It's possible someone else will come along who wants to work on Rustup and has the time to work it through. A rebase ought not to be too painful (though Dale has rearranged the codebase now so it's not going to be entirely trivial), but the tests remain the sticking point.

[pickfire]

@dwijnand Thanks a lot for helping out.

@kinnison I have no issues rebasing, I am just not sure what tests to write for this.

[kinnison]

Regarding tests, the main thing is that we should verify that rustup override {un,}set --path ... works correctly. This probably means creating a test which sets up a directory with an override, and then verifies that it works, and then unsets it and verifies the override is gone.

However, if you're really stuck with how to do that, just hold the rebased branch for now until someone can help you with the tests. Sadly I don't have enough time right now to provide an example.

[bors]

☔ The latest upstream changes (presumably #1754) made this pull request unmergeable. Please resolve the merge conflicts.

[pickfire]

I added test for both set and unset only for --path.

[kinnison]

Hi @pickfire

Thanks for updating the branch, but the Windows test (Appveyor) failed. From a glance it looks like you maybe hard-coded an architecture name in your test? I will try and review the code later to help you out if you can't spot the bug.

D.

[pickfire]

@kinnison Could you please help me to check what to change? I am not so rich to buy a windows laptop, I don't have one to test it out (as well I am not very familiar with windows).

[kinnison]

Other than the test issue, this LGTM.