Changelog for awwan

Previously, the "env-set" take the file argument as the last argument and optional. This changes move the file argument to the first argument so the bash completion can detect and print the completion for list of keys.

While at it, fix handling key with quoted in EnvSet and EnvGet.

Previously, the magic line "#put" detect whether the source file is encrypted or not automatically, so we did not need to put ".vault" suffix in the source path. This changes make it to be always explicit. If we want to copy encrypted file than the source file must be the path to encrypted file.

The Stop button call "DELETE /awwan/api/execute?id=" with id is the previous execution ID.

The HTTP API for stopping execution have the following signature,

If the ID is exist, the execution will be cancelled and return HTTP status 200 with the following body,

Otherwise it will return HTTP status 404 with error message.

This changes we make sure that the passed script path is not a directory to prevent error when running play with unknown host name.

Once the ExecResponse EndAt is not empty or event queue has been closed, clear the map that store the execution response and context cancellation to minimize resources usage.

If the file to be copied contains sensitive data, putting them in "/tmp" considered a security risk, even though it will be moved to destination later.

The issue is when the "#put" command failed, the plain file is left on "/tmp" directory.

This changes add additional advantage where we did not need to remove the temporary directory on remote when execution completed, since the temporary directory should be accessible by user only.

The env-get command get the value from environment files. Syntax,

The "key" argument define the key where value is stored in environment using "section:sub:name" format. The "dir" argument is optional, its define the directory where environment files will be loaded, recursively, from BaseDir to dir. If its empty default to the current directory.

The env-set command set the value of environment file. Syntax,

The "key" argument define the key to be set using "section:sub:name" format. The "value" argument define the value key. The "file" argument define path to environment file.

For example, to set the value for "name" under section "host" to "myhost" in file "awwan.env" run

To set the value for key "pass" under section "user" subsection "database" to value "s3cret" in file "awwan.env" run

In the right side of Save button we now have a button Encrypt that allow user to Encrypt opened file.

This require the workspace has been setup with private key (.ssh/awwan.key) and pass file (.ssh/awwan.pass).

In the right side of Save button we now have a button Decrypt that allow user to Decrypt file with ".vault" extension only.

This require the workspace has been setup with private key (.ssh/awwan.key) and pass file (.ssh/awwan.pass).

In case we have a script that manage local host and remote server, calling "play" on "#local" lines only always open the connection to remote server.

To minimize opening unused connections, let the "#local" command works on both commands. Its up to user which part of lines that they want to execute on remote or local.

Previously, we used to run awwan as CLI so each connection is open and closed once the command completed. Since we now use awwan WUI frequently, any command that execute Play does not close the session immediately once finished. This cause many connections open in remote server.

This changes close the SSH connections immediately once the Play command finished.

Changes,

In case awwan run with "serve" and we modify the ".ssh/config", the changes does not detected by awwan because we only read ".ssh/config" once we Awwan instance created.

This changes fix this issue by always loading SSH config every time the Play method executed so the user CLI and WUI has the same experiences.

Previously, the HTTP endpoint for "GET /awwan/api/fs" return the content of files when the requested node is a directory. This is cause unnecessary load because when requesting directory we only need list of file names not the content.

This changes reduce the response by returning only list of node child without its content, which require update on share module on [Node.JSON].

Previously, we use a quick "pop-up" to display notification for each information or error from WUI. Sometimes this is annoying, it overlap the buttons, make it hard to Save and Encrypt at the same time.

In this changes we move the nofication message to be displayed in the output, same with output of execution.

Previously, the Execute endpoint wait for command execution to finish. In case the command takes longer than proxy or server write timeout, it will return with a timeout error to client.

In this changes, we generate an execution ID for each request and return it immediately.

The new API "GET /awwan/api/execute/tail" implement Server-sent events, accept the execution ID from "/awwan/api/execute".

Once called with valid ID, it will streaming the command output to client.

By using this new API, the WUI can receive the output of command immediately without waiting for all commands to be completed.

This changes move the text input for creating new file above the list.

Filling the text field will filter the list based on the node name using regular expression.

Another changes is for node with type directory now suffixed with "/".

The magic command "#get" and "#put" now have an inline options to set the owner and permission of copied file. Example of usage are,

The $USER, $GROUP and $PERM are optionals.

If $USER and/or $GROUP is set, a copied file will have owner set to user $USER and/or group to $GROUP. If $PERM is set, a copied file will have the mode permission set to $MODE.

For each script execution, a file suffixed with ".log" will be created in the same directory with the same name as script file. For example, if the script is path is "a/b/c.aww" then the log file would named "a/b/c.aww.log".

This is to provides history and audit in the future.

Previously, the HTTP API for deleting node only remove the file but not the node in the memfs.

This changes remove the child node from memfs, so the next refresh on directory will not contains the removed file.

Previously, when the command execution failed, we check the error and return it as HTTP status code 500. In this way, user cannot view the log and actual error.

In this changes, if the command failed, we store the error in separate field "Error" and return to the caller with HTTP status code 200.

By using "mlog.MultiLogger" every output or error can be written to stdout/stderr and additional log writer that can collect both of them, buffered and returned to the caller.

This changes simplify the HttpResponse to use only single output that combine both stdout and stderr.

If a multi lines command does not have spaces or have multiple spaces, join them as is. For example,

should return the value as ab, while

should return "a<space><space>b".

Changes,

This is to prevent copying or executing command with value that are not defined or typo which make the result empty and may result in undefined behaviour.

For example if we have "app_dir =" with empty variable and command in the script that remove that directory recursively,

will result removing "/bin" entirely.

Previously, the "local" and "play" command only accept two kind of arguments: one argument for executing single line or two arguments for executing line range. There are no options to executing multiple single line, multiple line range, or combination of them.

This changes make the both commands accept list of lines or line range where each separated by comma. For example, to execute multiple, different single lines

To execute multiple line range,

Or to execute multiple lines and line range,

I use two remote repositories: GitHub and SourceHut. GitHub support rendering README using asciidoc while SourceHut not. This cause the repository that use README.adoc rendered as text in SourceHut which make the repository page less readable.

Also, the pkg.go.dev now render README but only support Markdown.

Since we cannot control the SourceHut and go.dev, the only option is to support converting Markdown in ciigo so I can write README using Markdown and the rest of documentation using Asciidoc.

With configuration we can enable or disable specific linters and customize it without touching the Makefile.

While at it, enable linter presets for bugs, comment, metalinter, module, performance, unused and fix the reported issues.

Previously, the /_www/doc is a symlink to /_adoc. To minimize symlinks in this repository, we replace the _www/doc with its actual directory.

While at it, move CHANGELOG.adoc to root directory for better view ability.

Changes,

The main.js already build with esbuild.

Using git schema will fail on user that does not have SSH key in git.sr.ht.

Since fa94025f8e, we change the local statement to be executed using "/bin/sh", "-c", "<statement>". This cause the #get! command failed to execute because

Previously, to print the help and version we use the flag package, "-h" for printing usage and "-v" for printing "version".

The flag should be used to changes the behaviour of command not terminating the command. Using and mixing flag as command make the CLI API to be inconsistent.

For example, the following argument is inconsistent,

Instead of executing script.aww, it print the usage and terminate.

Previously, we build the JavaScript files on directory _www by using TypeScript compiler (tsc).

This changes replace the tsc with esbuild, a Go module that provide API to convert TypeScript to JavaScript, and use tsc for type checking only.

The web-user interface now have button "Save" that save the edited content of file to storage.

The serve command will run a HTTP server that provide web-user interface to edit and execute script.

Currently, the web-user interface provide the following features,

By using "sh -c", the statement can contains environment variable, piping, and other shell commands that cannot be supported if executed directly using native os.Exec.

This changes replace the string ${x} or $x in the statements with the current environment variables values.

For example, statement "mkdir ${HOME}/tmp" will expand the ${HOME} to the current user home directory.

Previously, the command to put and get file from remote server depends on installed scp program on the host computer.

In this changes we add the SFTP client and use it as primary function to put and get file to/from remote when possible and use the scp as fallback.

If the mode is local there is no need to parse and load the SSH config, since the command to be executed will run on local machine anyway.

Previously, the base directory is set on current working directory. This limit the way to execute awwan only from the directory that contains the .ssh directory.

This changes set the environment BaseDir based on the first .ssh directory that we found from script path up to the "/". So, user can execute awwan script from any sub-directories.

By knowing this values, user can use it to invoke other SSH related command, for example to copy file using scp

Magic word #require: will ensure that the next statement will always executed when its skipped with start number. For example, given following script with line number

executing awwan local script.aww 3, will always execute line number 2 echo a, but not line number 5 (because its before line start 3).

In case the owner of file is not active user and it does not have read permission, the "#get!" command will fail when copying command from remote to local.

Previously, if command contains quote like,

the script will run it as ["echo", "a, b"] which is not what we will expected and may cause some command failed to run.

This changes fix the parsing of command string by detecting possible quote.