added filter for statement statistics that contain null entries #40
Conversation
| statementMeta = fileCoverage.statementMap; | ||
| Object.keys(statementStats).forEach(function (stName) { | ||
| var statementStats = Object.keys(fileCoverage.s).filter(function (stName) { | ||
| return fileCoverage.s[stName] !== null; |
There was a problem hiding this comment.
would it be overly poor form to use != null to capture any ones that are undefined as well? or would it be overkill to do && !== undefined?
| var statementStats = Object.keys(fileCoverage.s).filter(function (stName) { | ||
| return fileCoverage.s[stName] !== null; | ||
| }), | ||
| statementMeta = fileCoverage.statementMap; |
There was a problem hiding this comment.
I think this should be a var statementMeta line instead of the comma-list-declaration syntax, it's kind of hard to grok imho.
|
@JaKXz I'm pretty sure that in normal circumstances, getting a @SamNelson are you able to share the codebase that you were bumping into issues on? I think if we switch the order of |
|
@bcoe wouldn't it be good to have this passing anyways? in case there are other scenarios where there are null entries in statement statistics? |
|
@JaKXz the problem is it would result in misreporting test coverage information -- except in bizarre situations, like the interaction we're seeing between |
JaKXz
dismissed
their stale review
Based on @bcoe's point, @SamNelson I think you should just re-order the plugins. I believe @hzoo started a thread about figuring out how to better approach the babel plugin order problem, but I don't know where that is.
|
@SamNelson please feel free to reopen this discussion if you're still bumping into issues 👍 we've rolled out a few changes to nyc/istanbul-lib-instrument that I think should solve your problem. |
This should fix issue #39