Community
Participate
Working Groups
As requested in bug 35068 I am raising this as an enhancement. Pressing an appropriate symbol doesn't always complete the current code-assist selection. eg int offset, length = 10; int i = offset + le<ctrl_space>+ Pressing + should complete "length". eg2 int[] array = new int[le<ctrl_space>] Pressing ] should complete "length" Affected symbols + - / * < > & | ? : ^ % ) } ] ! A fix is attached to bug 35068.
Probably I'm out of the loop, nevertheless two issues: 1) int i = offset + le<ctrl_space><space> Pressing <space> should complete "length" => Does this work? As far as I know, the SPACE did NOT have any effect up till now. 2) int i = offset + leng<ctrl_space> Pressing <ctrl_space> should complete "length" without any FURTHER requests from the user. i.e. when no further possibilities from the list possible AUTO-Complete should be done. If both don't fit exactly into this Bug then I'm willing to open another bug. But then somebody should give me more info if this is required. ALSO, frankly said: Don't know if in the current built both issues were already fixed or not.
Created attachment 27049 [details] Complete on symbols R3_1 This is an update to the patch attached to bug 35068 against R3_1. It fixes this bug and bug 39667.
I agree that e.g. operators make sense when entering an expressionbut but since we currently have no means to get the context (e.g. whether we're entering an expression) we don't want to add more triggers which are used in all cases.
Created attachment 27078 [details] Complete on symbols (inc casts) R3_1 We can generalize this to 2 situations: A) In a situation where the user has chosen to press an invalid symbol (eg "extends Clonea<ctrl-space>+") 1) We could trigger and produce invalid Java or, 2) We could not trigger and produce invalid Java. And you want to choose 2, because the symbol is invalid. B) In a situation where the user has chosen to press a valid symbol (eg "a = new int[leng<ctrl-space>]) 1) We could trigger and produce valid Java or, 2) We could not trigger and produce invalid Java. And you want to choose 2 to behave the same as A. You appear to be trying to fix the behaviour when the cat jumps on the keyboard at the expense of breaking it when the user is trying to write valid Java. The cat doesn't care whether we trigger or not, it just wants to be stroked :-)
Created attachment 27232 [details] symbol triggers
*** Bug 139239 has been marked as a duplicate of this bug. ***
Is there any chance to have a fix in the next release? Since in 3.2 this is still an issue.
Get rid of deprecated state.