Testing the contextGenerator class

[iifawzi]

Signed-off-by: iifawzi iifawzie@gmail.com

[codecov]

Codecov Report

Base: 95.60% // Head: 97.40% // Increases project coverage by +1.79% 🎉

Coverage data is based on head (e1fc834) compared to base (a70dec0).
Patch has no changes to coverable lines.

Additional details and impacted files

@@             Coverage Diff              @@
##             master     #394      +/-   ##
============================================
+ Coverage     95.60%   97.40%   +1.79%     
  Complexity     2174     2174              
============================================
  Files            68       68              
  Lines          4617     4617              
============================================
+ Hits           4414     4497      +83     
+ Misses          203      120      -83     

Impacted Files	Coverage Δ
src/Tools/ContextGenerator.php	76.85% <0.00%> (+76.85%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[iifawzi]

can't find a proper name to test this? could you help @williamdes

sql-parser/src/Tools/ContextGenerator.php

Lines 292 to 294 in a70dec0

	if (preg_match('/([^[0-9]*)([0-9]*)/', $name, $parts) === false) {
	return $name;
	}

[iifawzi]

Most probably, this's a dead code

sql-parser/src/Tools/ContextGenerator.php

Lines 177 to 180 in a70dec0

	$len = strlen($words[$i]);
	if ($len === 0) {
	continue;
	}

it checks whether the keyword is of length zero? how could this be possible?

[iifawzi]

Multiple keywords are added here to test that the keywords will be in a new line

[williamdes]

Hi,

Thank you for this, I will have a look when I finish my refactoring upstream at mariadb-mysql-kbs. I am near to have re-implementing the scraping into Rust code. Next should follow data for keywords and if possible functions and everything else.

I was thinking it would be nice to:

• reduce the format of context files to a minimum with sections like:
  • Added
  • Removed
  • Changed

When using a vscode plugin to have the diff between files I did see that most of the file remains the same
It makes it uneasy to track changes and see missing elements.
We should rather build a diff system because it's real world about what is exactly the case.

Common is version v0 and MariaDB is v1 and so on and a version is nothing without the previous one. So it's like git.
If the contexts end up cleaned up like that we can fire up my upstream changes and ask for version x.y.z what are the known keywords and compare the lists to find missing on wrong ones.
Ideally when upstream and the parser are reconciled we can (maybe?) delete the context files and use my lib as source of truth. And still make manual updates of contexts and keywords