You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Based on conversation in #601 and nodejs/node#22585, I believe the following TSC members are in favor of shipping recursive mkdir() in v10.x as-is (that is, as something that can be turned on in the options object of fs.mkdir()):
I don't believe further conversation is likely to move many people from one camp to the other. And I don't believe that taking more time to decide this is in the interest of the community. I therefore would like to call for a vote. (Anyone on the TSC can stop the vote by disagreeing that it's time for a vote and indicating that they believe more conversation would be useful. I implore you not to do that, but I can't stop you from doing it if it's what you believe.)
So, vote counting: If I have you categorized correctly above, please leave a comment to that effect or else a 👍 on this comment. If I have you categorized incorrectly, please leave a comment indicating how to correct. If I have not categorized you at all, please leave a comment as to where your vote should go. That last one would apply to @rvagg@fhinkel@ChALkeR@gibfahn@TimothyGu
As I've clarified a couple of times now, I'm not opposed to shipping recursive mkdir in 10.x. I'm opposed to shipping it if we're going to change it after we settle on the feature detection. I'm perfectly fine with shipping it now so long as it's understood that the feature detection discussion is a wholly separate issue.
As I've clarified a couple of times now, I'm not opposed to shipping recursive mkdir in 10.x. I'm opposed to shipping it if we're going to change it after we settle on the feature detection. I'm perfectly fine with shipping it now so long as it's understood that the feature detection discussion is a wholly separate issue.
There is no consensus to avoid the alias which is why I listed you as opposed.
Activity
Trott commentedon Sep 21, 2018
And I'll make this comment for tallying things once people confirm their sentiments above:
YES:
NO:
ABSTAIN:
jasnell commentedon Sep 21, 2018
As I've clarified a couple of times now, I'm not opposed to shipping recursive mkdir in 10.x. I'm opposed to shipping it if we're going to change it after we settle on the feature detection. I'm perfectly fine with shipping it now so long as it's understood that the feature detection discussion is a wholly separate issue.
Trott commentedon Sep 21, 2018
There is no consensus to avoid the alias which is why I listed you as opposed.
mhdawson commentedon Sep 24, 2018
+1 for shipping as is.
fhinkel commentedon Sep 24, 2018
I abstain.
mcollina commentedon Sep 24, 2018
+1 for shipping as is.
Trott commentedon Sep 24, 2018
With 10 votes for and 1 abstention, that's enough to decide it for "ship it".