Create SelenideElement from Suppliers

[nick318]

I suggest an idea of getting element from Supplier<? extends WebElement>.
It can be helpful in different situations:

• If we have method that return some div of block and we want to find some element from that block we can write something like that:
  WebElement el = getDiv();
  $(el).$(By.id("input").setValue("text");
  But in this case we have a problem if something goes wrong with first element 'el', I mean if that element updates, we will get StaleElementReferenceException and Selenide cannot re-init that element.

But have a look at that example:
$(this::getDiv).$(input).setValue("text").

In that case, if something goes wrong, Selenide will invoke the given supplier and can re-init first element and others elements as well.

@asolntsev What do you think?

[asolntsev]

@Icok It's an interesting idea. Yes, probably we should do it.

But can you describe a real case? How could your getDiv() method be implemented?
Can't you just use $("selector")?

[nick318]

@asolntsev To answering to your question:
-- Can't you just use $("selector")?
The problem is if someone already implemented methods that return WebElement.
Sometimes yes I can, but I think it is not necessary restriction to return everywhere only 'By' objects.

-- How could your getDiv() method be implemented?
some examples:

getDiv() {
return $$("collection").stream()
                .filter(elem -> elem.is(visible))
                .filter(elem -> elem.getText().equals("need text"))
                .findFirst().orElseThrow(() -> new AssertionError("Not found div"));
}

getDiv() {
   //some actions to click by 3-rd libraries like Robot or Sikuli.
  return driver.switchTo().activeElement(); //It helpful in IE browser, because it does not work by locators sometimes
}

[asolntsev]

@Icok The build is failing. There are some checkstyle issues. Please execute ./gradle check in your branch.

[asolntsev]

Approved, but please make the build green.

[coveralls]

Coverage decreased (-0.2%) to 64.11% when pulling a47a83f on iCok:add_supplier_init into b7fe2c7 on codeborne:master.

[codecov-io]

Codecov Report

Merging #544 into master will decrease coverage by 0.51%.
The diff coverage is 83.33%.

@@             Coverage Diff              @@
##             master     #544      +/-   ##
============================================
- Coverage     60.53%   60.02%   -0.52%     
+ Complexity      772      768       -4     
============================================
  Files           148      149       +1     
  Lines          2752     2764      +12     
  Branches        272      272              
============================================
- Hits           1666     1659       -7     
- Misses          981      992      +11     
- Partials        105      113       +8

Impacted Files	Coverage Δ	Complexity Δ
src/main/java/com/codeborne/selenide/Selenide.java	0% <0%> (ø)	0 <0> (ø)	⬇️
...a/com/codeborne/selenide/impl/ElementSupplier.java	90.9% <90.9%> (ø)	5 <5> (?)
.../com/codeborne/selenide/commands/GetInnerHtml.java	60% <0%> (-40%)	2% <0%> (-1%)
...in/java/com/codeborne/selenide/impl/Navigator.java	59.37% <0%> (-6.25%)	19% <0%> (-2%)
.../java/com/codeborne/selenide/ex/ErrorMessages.java	82.75% <0%> (-3.45%)	16% <0%> (-1%)
.../codeborne/selenide/impl/ScreenShotLaboratory.java	51.68% <0%> (-2.81%)	31% <0%> (-3%)
...ain/java/com/codeborne/selenide/impl/Describe.java	78.2% <0%> (-2.57%)	27% <0%> (-1%)
...e/selenide/impl/WebDriverThreadLocalContainer.java	80.43% <0%> (-2.18%)	30% <0%> (-1%)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b7fe2c7...a47a83f. Read the comment docs.

[nick318]

@asolntsev Done

[BorisOsipov]

I am closing it due inactivity. Feel free to start conversation again if you would mind to add such changes.