-
Notifications
You must be signed in to change notification settings - Fork 67
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Array spread breaks ternaries due to not parenthesising them #177
Comments
chris-morgan
added a commit
to chris-morgan/buble
that referenced
this issue
Jan 10, 2019
In most cases this is unnecessary, but just occasionally it is necessary, notably in ternaries; for `a ? b : c.concat(…)` is very different from `(a ? b : c).concat(…)`. We could be cleverer about whether parentheses are needed, but Bublé doesn’t generally seem to go out of its way to *absolutely* minimise such things, and I’m also not sure whether there’s a good way to do that, or whether you have to start doing checks “is it a ParenthesizedExpression, CallExpression, NewExpression, &c.” With the opening parenthesis leaked into CallExpression.js, it’d become messier, too. I’m not clear why the insertion of the opening parenthesis had to happen in src/program/types/CallExpression.js rather than src/utils/spread.js when start === element.start; but if I did it inside spread(), if I used appendLeft or prependLeft it’d end up like `.apply((Math, a).concat(…))` instead of `.apply(Math, (a).concat(…))`, and if I used appendRight or prependRight, nothing happened. Fixes bublejs#177.
chris-morgan
added a commit
to chris-morgan/buble
that referenced
this issue
Jan 23, 2019
In most cases this is unnecessary, but just occasionally it is necessary, notably in ternaries; for `a ? b : c.concat(…)` is very different from `(a ? b : c).concat(…)`. We could be cleverer about whether parentheses are needed, but Bublé doesn’t generally seem to go out of its way to *absolutely* minimise such things, and I’m also not sure whether there’s a good way to do that, or whether you have to start doing checks “is it a ParenthesizedExpression, CallExpression, NewExpression, &c.” With the opening parenthesis leaked into CallExpression.js, it’d become messier, too. I’m not clear why the insertion of the opening parenthesis had to happen in src/program/types/CallExpression.js rather than src/utils/spread.js when start === element.start; but if I did it inside spread(), if I used appendLeft or prependLeft it’d end up like `.apply((Math, a).concat(…))` instead of `.apply(Math, (a).concat(…))`, and if I used appendRight or prependRight, nothing happened. Fixes bublejs#177.
chris-morgan
added a commit
to chris-morgan/buble
that referenced
this issue
Jan 23, 2019
In most cases this is unnecessary, but in some cases it is necessary, notably in ternaries; for `a ? b : c.concat(…)` is very different from `(a ? b : c).concat(…)`. I’m not clear why the insertion of the opening parenthesis had to happen in src/program/types/CallExpression.js rather than src/utils/spread.js when start === element.start; but if I did it inside spread(), if I used appendLeft or prependLeft it’d end up like `.apply((Math, a).concat(…))` instead of `.apply(Math, (a).concat(…))`, and if I used appendRight or prependRight, nothing happened. Fixes bublejs#177.
adrianheine
pushed a commit
that referenced
this issue
Jan 23, 2019
In most cases this is unnecessary, but in some cases it is necessary, notably in ternaries; for `a ? b : c.concat(…)` is very different from `(a ? b : c).concat(…)`. I’m not clear why the insertion of the opening parenthesis had to happen in src/program/types/CallExpression.js rather than src/utils/spread.js when start === element.start; but if I did it inside spread(), if I used appendLeft or prependLeft it’d end up like `.apply((Math, a).concat(…))` instead of `.apply(Math, (a).concat(…))`, and if I used appendRight or prependRight, nothing happened. Fixes #177.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Minimal test case:
Result:
Expected:
These two are wildly different!
The workaround is manual parenthesisation:
Given that you can only spread iterables in array spread, I think ternaries are the only realistic situation where this will matter. (Well, I guess strings could wreck it too—
...a + b
—but #166 shows spreading strings is broken anyway.)The text was updated successfully, but these errors were encountered: