Hi,
I notice a regression in JLine2.13 for AggregateCompleter, when the cursor is in the middle of an argument. It is a bit difficult to reproduce right now:
List<Completer> completors = new LinkedList<Completer>();
completors.add(
new ArgumentCompleter(
new StringsCompleter("foo11", "foo12", "foo13"),
new StringsCompleter("foo21", "foo22", "foo23")));
Then, type these keys exactly, | indicates cursor position:
Type: f, o, o, 1, <left>
Result: foo|1
Type: <tab>
Result: foo11|1
Type: <right>, <backspace>, <backspace>, <left>
Result: foo|1
Type <tab>
Result: foo1o|o
# should be: foo11|1
At the end, the cursor cannot go back to delete the first 'f'.
In JLine 2.12.0 and 2.12.1, the behavior is different, with a harmless glitch (additional virtual '1' appears behind the last valid char of the buffer), but else consistent.
Using git bisect, I believe the regression appears with 3b0c474 : Support for infocmp capabilities.
Hi,
I notice a regression in JLine2.13 for AggregateCompleter, when the cursor is in the middle of an argument. It is a bit difficult to reproduce right now:
Then, type these keys exactly,
|indicates cursor position:At the end, the cursor cannot go back to delete the first 'f'.
In JLine 2.12.0 and 2.12.1, the behavior is different, with a harmless glitch (additional virtual '1' appears behind the last valid char of the buffer), but else consistent.
Using git bisect, I believe the regression appears with 3b0c474 : Support for infocmp capabilities.