-
Notifications
You must be signed in to change notification settings - Fork 253
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Remove duplication in clearing command stdout / stderr #235
Remove duplication in clearing command stdout / stderr #235
Conversation
Try to protect against duplicate log lines caused by the interaction between the backend, command and formatter
I believe issue capistrano#224 also affects the SimpleText formatter. I’ve added a failing test (test_command_lifecycle_logging) to expose the problem.
This commit removes duplication in the SimpleText and Pretty formatters by better encapsulating the logic to split and clear stdout/stderr buffers on the Command object.
Factored out duplicated logic from SimpleText and Pretty formatters for formatting a log line from one of the standard streams (ie stdout/stderr).
Looked good to me @robd , but there's failures on travis, could you take a look and then I'd be more than happy to merge. |
Sorry @leehambley - my bad, I didn't test on ruby 1.9. Just out of interest - do you have any easy way to run the tests against multiple rubies locally? |
string.lines returns array in ruby 2, but enumerable in 1.9. Convert the enumarable to array - can remove this when we stop supporting ruby 1.9.
@leehambley I fixed the 1.9 compatibility issues, so hopefully this is good to go. |
I use
Thanks, merged with my gratitude! |
Remove duplication in clearing command stdout / stderr
@leehambley Sorry build breakage is my fault due to a stupid timing assumption. - I am travelling so have dodgy internet, but will get a fix PR to you ASAP |
Acknowledged Rob, it also passed locally for me, didn't have time to
|
Fixed in #236. Sorry about that. |
## 1.8.1 * Change license to MIT, thanks to all the patient contributors who gave their permissions. ## 1.8.0 * add SSHKit::Backend::ConnectionPool#close_connections [PR #285](capistrano/sshkit#285) @akm * Clean up rubocop lint warnings [PR #275](capistrano/sshkit#275) @cshaffer * Prepend unused parameter names with an underscore * Prefer “safe assignment in condition” * Disambiguate regexp literals with parens * Prefer `sprintf` over `String#%` * No longer shadow `caller_line` variable in `DeprecationLogger` * Rescue `StandardError` instead of `Exception` * Remove useless `private` access modifier in `TestAbstract` * Disambiguate block operator with parens * Disambiguate between grouped expression and method params * Remove assertion in `TestHost#test_assert_hosts_compare_equal` that compares something with itself * Export environment variables and execute command in a subshell. [PR #273](capistrano/sshkit#273) @kuon * Introduce `log_command_start`, `log_command_data`, `log_command_exit` methods on `Formatter` [PR #257](capistrano/sshkit#257) @robd * Deprecate `@stdout` and `@stderr` accessors on `Command` * Add support for deprecation logging options. [README](README.md#deprecation-warnings), [PR #258](capistrano/sshkit#258) @robd * Quote environment variable values. [PR #250](capistrano/sshkit#250) @Sinjo - Chris Sinjakli * Simplified formatter hierarchy. [PR #248](capistrano/sshkit#248) @robd * `SimpleText` formatter now extends `Pretty`, rather than duplicating. * Hide ANSI color escape sequences when outputting to a file. [README](README.md#output-colors), [Issue #245](capistrano/sshkit#245), [PR #246](capistrano/sshkit#246) @robd * Now only color the output if it is associated with a tty, or the `SSHKIT_COLOR` environment variable is set. * Removed broken support for assigning an `IO` to the `output` config option. [Issue #243](capistrano/sshkit#243), [PR #244](capistrano/sshkit#244) @robd * Use `SSHKit.config.output = SSHKit::Formatter::SimpleText.new($stdin)` instead * Added support for `:interaction_handler` option on commands. [PR #234](capistrano/sshkit#234), [PR #242](capistrano/sshkit#242) @robd * Removed partially supported `TRACE` log level. [2aa7890](capistrano/sshkit@2aa7890) @robd * Add support for the `:strip` option to the `capture` method and strip by default on the `Local` backend. [PR #239](capistrano/sshkit#239), [PR #249](capistrano/sshkit#249) @robd * The `Local` backend now strips by default to be consistent with the `Netssh` one. * This reverses change [7d15a9a](capistrano/sshkit@7d15a9a) to the `Local` capture API to remove stripping by default. * If you require the raw, unstripped output, pass the `strip: false` option: `capture(:ls, strip: false)` * Simplified backend hierarchy. [PR #235](capistrano/sshkit#235), [PR #237](capistrano/sshkit#237) @robd * Moved duplicate implementations of `make`, `rake`, `test`, `capture`, `background` on to `Abstract` backend. * Backend implementations now only need to implement `execute_command`, `upload!` and `download!` * Removed `Printer` from backend hierarchy for `Local` and `Netssh` backends (they now just extend `Abstract`) * Removed unused `Net::SSH:LogLevelShim` * Removed dependency on the `colorize` gem. SSHKit now implements its own ANSI color logic, with no external dependencies. Note that SSHKit now only supports the `:bold` or plain modes. Other modes will be gracefully ignored. [#263](capistrano/sshkit#263) * New API for setting the formatter: `use_format`. This differs from `format=` in that it accepts options or arguments that will be passed to the formatter's constructor. The `format=` syntax will be deprecated in a future release. [#295](capistrano/sshkit#295) * SSHKit now immediately raises a `NameError` if you try to set a formatter that does not exist. [#295](capistrano/sshkit#295)
## 1.8.1 * Change license to MIT, thanks to all the patient contributors who gave their permissions. ## 1.8.0 * add SSHKit::Backend::ConnectionPool#close_connections [PR #285](capistrano/sshkit#285) @akm * Clean up rubocop lint warnings [PR #275](capistrano/sshkit#275) @cshaffer * Prepend unused parameter names with an underscore * Prefer “safe assignment in condition” * Disambiguate regexp literals with parens * Prefer `sprintf` over `String#%` * No longer shadow `caller_line` variable in `DeprecationLogger` * Rescue `StandardError` instead of `Exception` * Remove useless `private` access modifier in `TestAbstract` * Disambiguate block operator with parens * Disambiguate between grouped expression and method params * Remove assertion in `TestHost#test_assert_hosts_compare_equal` that compares something with itself * Export environment variables and execute command in a subshell. [PR #273](capistrano/sshkit#273) @kuon * Introduce `log_command_start`, `log_command_data`, `log_command_exit` methods on `Formatter` [PR #257](capistrano/sshkit#257) @robd * Deprecate `@stdout` and `@stderr` accessors on `Command` * Add support for deprecation logging options. [README](README.md#deprecation-warnings), [PR #258](capistrano/sshkit#258) @robd * Quote environment variable values. [PR #250](capistrano/sshkit#250) @Sinjo - Chris Sinjakli * Simplified formatter hierarchy. [PR #248](capistrano/sshkit#248) @robd * `SimpleText` formatter now extends `Pretty`, rather than duplicating. * Hide ANSI color escape sequences when outputting to a file. [README](README.md#output-colors), [Issue #245](capistrano/sshkit#245), [PR #246](capistrano/sshkit#246) @robd * Now only color the output if it is associated with a tty, or the `SSHKIT_COLOR` environment variable is set. * Removed broken support for assigning an `IO` to the `output` config option. [Issue #243](capistrano/sshkit#243), [PR #244](capistrano/sshkit#244) @robd * Use `SSHKit.config.output = SSHKit::Formatter::SimpleText.new($stdin)` instead * Added support for `:interaction_handler` option on commands. [PR #234](capistrano/sshkit#234), [PR #242](capistrano/sshkit#242) @robd * Removed partially supported `TRACE` log level. [2aa7890](capistrano/sshkit@2aa7890) @robd * Add support for the `:strip` option to the `capture` method and strip by default on the `Local` backend. [PR #239](capistrano/sshkit#239), [PR #249](capistrano/sshkit#249) @robd * The `Local` backend now strips by default to be consistent with the `Netssh` one. * This reverses change [7d15a9a](capistrano/sshkit@7d15a9a) to the `Local` capture API to remove stripping by default. * If you require the raw, unstripped output, pass the `strip: false` option: `capture(:ls, strip: false)` * Simplified backend hierarchy. [PR #235](capistrano/sshkit#235), [PR #237](capistrano/sshkit#237) @robd * Moved duplicate implementations of `make`, `rake`, `test`, `capture`, `background` on to `Abstract` backend. * Backend implementations now only need to implement `execute_command`, `upload!` and `download!` * Removed `Printer` from backend hierarchy for `Local` and `Netssh` backends (they now just extend `Abstract`) * Removed unused `Net::SSH:LogLevelShim` * Removed dependency on the `colorize` gem. SSHKit now implements its own ANSI color logic, with no external dependencies. Note that SSHKit now only supports the `:bold` or plain modes. Other modes will be gracefully ignored. [#263](capistrano/sshkit#263) * New API for setting the formatter: `use_format`. This differs from `format=` in that it accepts options or arguments that will be passed to the formatter's constructor. The `format=` syntax will be deprecated in a future release. [#295](capistrano/sshkit#295) * SSHKit now immediately raises a `NameError` if you try to set a formatter that does not exist. [#295](capistrano/sshkit#295)
This is really preparatory work for supporting pty handlers - #234
This PR grew from me not understanding why these lines that clear stdout/stderr are needed. These were added to as part of #224 to get round a duplicate messages issue, but I wanted to backfill with a failing test so I don't break it when I start fiddling with that code in #234.
I wrote a test case for the
Pretty
formatter in 61e03af which fails with duplicate messages if the clearing lines are deleted, I also noticed that there was no test at all for theSimpleText
formatter, so I backfilled this in c13b10b. This highlighted that the duplicate messages #224 issue is not fixed for the SimpleText formatter. So I fixed this in cb06ab2.The remaining commits remove duplication within and between the
Pretty
andSimpleText
formatters.Let me know what you think - I would welcome feedback on code style and so-on.