SIL is a programming language for automating actions in Atlassian Jira and Confluence. You can find more information about SIL here.
I often meet code written in SIL and I would like to share with you my thoughts on how to make SIL code clean.
First I will provide my rules on SIL clean code and then I will take an example of SIL code and refactor it to make it clean using the rules.
Rules
Here is an example of SIL code.
string USER = currentUser();
// Response
struct returnData {
string status;
}
// Project
struct space {
string key;
}
// Inner part with content
struct storage {
string value;
string representation;
}
// Part for storage
struct body {
storage storage;
}
// Main entity for sending to Confluence
struct reqData {
string type;
string title;
space space;
body body;
}
reqData data;
data.type = "page";
data.title = "Page for issue " + key + " " + summary + ".";
data.space.key = project;
data.body.storage.value = "<p> Author:"+userFullName(USER) + " description: " + description + "</p>";
data.body.storage.representation = "storage";
// Create request
HttpRequest request;
HttpHeader header = httpCreateHeader("Content-Type", "application/json");
HttpHeader authHeader = httpBasicAuthHeader("admin", "admin");
request.headers += header;
request.headers += authHeader;
logPrint("WARN", request);
//POST
string JSONData = toJson(data);
logPrint("WARN", JSONData);
returnData result = httpPost("http://192.168.54.203:8090/rest/api/content/", request, JSONData);
string errMsg = httpGetErrorMessage();
logPrint("ERROR", "Last error message: " + errMsg);
logPrint("WARN", result);
string COMMENT = "Page created in Confluence " + updated + " by " + userFullName(USER) + " Status : " +result.status + ".";
addComment(key, USER, COMMENT);
//Return Status
return result.status;
This code is difficult to understand fast and comments do not help with it. Let s try to modify this code to make it more readable using the rules.
Have a look at the structs. Here is an example:
struct body {
storage storage;
}
You see "storage storage" looks a bit confusing here. It would be more meaningful if it were like this:
struct body {
Storage storage;
}
Now it is obvious that we define the storage variable with the Storage type. To be able to make this change we need to start the names of all structs with the upper case:
// Response
struct ReturnData {
string status;
}
// Project
struct Space {
string key;
}
// Inner part with content
struct Storage {
string value;
string representation;
}
// Part for storage
struct Body {
Storage storage;
}
// Main entity for sending to Confluence
struct ReqData {
string type;
string title;
Space space;
Body body;
}
And let's name our structs so, that we would not need comments:
struct CreateConfluencePageResponse {
string status;
}
struct Space {
string key;
}
struct Storage {
string value;
string representation;
}
struct Body {
Storage storage;
}
struct CreateConfluencePageRequest {
string type;
string title;
Space space;
Body body;
}
In my opinion these names would be obvious enough to understand the structure.
Also we should put the struct declaration at the start of the code.
We completed Rule 1, 5, 6 from the rules.
Now if to look at the rest of the code we will see the following code:
string USER = currentUser();
CreateConfluencePageRequest data;
data.type = "page";
data.title = "Page for issue " + key + " " + summary + ".";
data.space.key = project;
data.body.storage.value = "<p> Author:"+userFullName(USER) + " description: " + description + "</p>";
data.body.storage.representation = "storage";
// Create request
HttpRequest request;
HttpHeader header = httpCreateHeader("Content-Type", "application/json");
HttpHeader authHeader = httpBasicAuthHeader("admin", "admin");
request.headers += header;
request.headers += authHeader;
logPrint("WARN", request);
//POST
string JSONData = toJson(data);
logPrint("WARN", JSONData);
CreateConfluencePageResponse result = httpPost("http://192.168.54.203:8090/rest/api/content/", request, JSONData);
string errMsg = httpGetErrorMessage();
logPrint("ERROR", "Last error message: " + errMsg);
logPrint("WARN", result);
string COMMENT = "Page created in Confluence " + updated + " by " + userFullName(USER) + " Status : " +result.status + ".";
addComment(key, USER, COMMENT);
//Return Status
return result.status;
We notice that we can not grasp the idea of the code fast. Let's try to split the code into blocks:
getNewConfluencePageContent();
createHttpRequest();
createConfluencePage();
addCommentToJiraIssue();
That would be the major blocks in our SIL code. Let's try to create all these functions:
function getNewConfluencePageContent() {
CreateConfluencePageRequest data;
data.type = "page";
data.title = "Page for issue " + key + " " + summary + ".";
data.space.key = project;
data.body.storage.value = "<p> Author:"+userFullName(currentUser()) + " description: " + description + "</p>";
data.body.storage.representation = "storage";
return toJson(data);
}
function createHttpRequest() {
HttpRequest request;
HttpHeader header = httpCreateHeader("Content-Type", "application/json");
HttpHeader authHeader = httpBasicAuthHeader("admin", "admin");
request.headers = addElement(request.headers, header);
request.headers += addElement(request.headers, authHeader);
logPrint("WARN", request);
return request;
}
function createConfluencePage(string pageJson) {
HttpRequest request = createHttpRequest();
CreateConfluencePageResponse result = httpPost("http://192.168.54.203:8090/rest/api/content/", request, pageJson);
logPrint("ERROR", "Last error message: " + httpGetErrorMessage());
logPrint("WARN", result);
return result;
}
function addCommentToJiraIssue(string resultStatus) {
string COMMENT = "Page created in Confluence " + updated + " by " + userFullName(currentUser()) + " Status : " +resultStatus + ".";
addComment(key, currentUser(), COMMENT);
}
string pageJson = getNewConfluencePageContent();
CreateConfluencePageResponse result = createConfluencePage(pageJson);
addCommentToJiraIssue(result.status);
return result.status;
So now we have a couple of functions (we completed Rule 3) and a set of events, which much easier to understand:
string pageJson = getNewConfluencePageContent();
CreateConfluencePageResponse result = createConfluencePage(pageJson);
addCommentToJiraIssue(result.status);
return result.status;
So we can see that first we get the contents of the new page. If we want to dig down to the exact contents we can look at the getNewConfluencePageContent function and find out what the contents are. Then we create a Confluence page from a json, then we add a comment to the Jira issues and return the status of the response from Confluence.
Then we can suggest that we need to create a Confluence page in more than one SIL script. We should move out the Confluence create page function into a reusable module. Let's do it. We can create a file called confluence_helper.incl with contents like this:
struct CreateConfluencePageResponse {
string status;
}
struct Space {
string key;
}
struct Storage {
string value;
string representation;
}
struct Body {
Storage storage;
}
struct CreateConfluencePageRequest {
string type;
string title;
Space space;
Body body;
}
function createHttpRequest() {
HttpRequest request;
HttpHeader header = httpCreateHeader("Content-Type", "application/json");
HttpHeader authHeader = httpBasicAuthHeader("admin", "admin");
request.headers = addElement(request.headers, header);
request.headers += addElement(request.headers, authHeader);
logPrint("WARN", request);
return request;
}
function createConfluencePage(string confluenceUrl, string pageJson) {
HttpRequest request = createHttpRequest();
CreateConfluencePageResponse result = httpPost(confluenceUrl, request, pageJson);
logPrint("ERROR", "Last error message: " + httpGetErrorMessage());
logPrint("WARN", result);
return result;
}
We moved out structs and functions, which are needed to create a Confluence page and made the createConfluencePage function more generic by adding a new parameter (confluenceUrl) to it. We completed Rule 4.
Now our SIL code would look like this:
include "confluence_helper.incl";
function getNewConfluencePageContent() {
CreateConfluencePageRequest data;
data.type = "page";
data.title = "Page for issue " + key + " " + summary + ".";
data.space.key = project;
data.body.storage.value = "<p> Author:"+userFullName(currentUser()) + " description: " + description + "</p>";
data.body.storage.representation = "storage";
return toJson(data);
}
function addCommentToJiraIssue(string resultStatus) {
string comment = "Page created in Confluence " + updated + " by " + userFullName(currentUser()) + " Status : " +resultStatus + ".";
addComment(key, currentUser(), comment);
}
const string CONFLUENCE_URL = "http://192.168.54.203:8090/rest/api/content/";
string pageJson = getNewConfluencePageContent();
CreateConfluencePageResponse result = createConfluencePage(CONFLUENCE_URL, pageJson);
addCommentToJiraIssue(result.status);
return result.status;
I saved the Confluence url in a constant variable and modified variable names according to Google Java notation style. We completed Rule 7.
I think we can stop here with our refactoring. Now we can easily reuse our function to create a Confluence page in other SIL routines and our code is easier to grasp.
Alexey Matveev
software developer
MagicButtonLabs
Philippines
1,574 accepted answers
2 comments