Fix log copy

[jreidinger]

No description provided.

[ancorgs]

Just asking. Do we have any convention on naming all these clients living under lib/installation? Like always using WhateverFinish for clients inheriting from FinishClient?

[jreidinger]

actually no convention...in bootloader it is simple Bootloader/FinishClient but here it can create collisions so I name it this way. CopyLogs without finish can confuse reader as it can be seen as common class to copy logs from one place to another.

[ancorgs]

The y2start.log returned by the previous call is not used in the expectation. Feels strange.

[jreidinger]

fair comment. I will add it.

[ancorgs]

LGTM.

BTW, if you start working on the SCR helpers, as suggested by you comment, I'm not unhappy with my last attempt. https://github.com/ancorgs/yast-security/blob/pre907907_cleanup/test/SCRStub.rb

[jreidinger]

@ancorgs yep, I plan to reuse your stub, but to be generic enough, there is some problems like .target.bash and .target.bash_output and also .local.bash for WFM and not SCR. I worry it won't be extra nice and probably some parts will be repetitive

[ancorgs]

That's why I removed things like expect_to_execute in the last incarnation and just implemented a exec_bash matcher for ".target.bash". My idea was to have a generic exec matcher that checks again every possible form and them some specific matchers like exec_bash, exec_bash_output and so on.

[jreidinger]

@ancorgs unfortunatelly this won't work as

def exec_bash(command)
receive(:Execute).with(path(".target.bash"), command)
end

work only for SCR, because WFM use .local.bash as System agent is attached to path .local and not .target. It is quite stupid and hard to handle. It also do not solve issue with .target.bash_output

[ancorgs]

@jreidinger ahm. I see.