Community
Participate
Working Groups
In Eclipse Proton Version: 2018-09 (4.9.0) Build id: 20180917-1800 the order of Java completion proposals is completely unreasonable and hinders workflow. Given the following snippet: NavigableMap<...> map; map.ent Invoking the content assist on the variable 'map' with a char sequence of 'ent' results in the following order of proposals: ceilingEntry computeIfAbsent computeIfPresent entrySet Of course, I'm looking for the proposal entrySet, and it takes too many keystrokes to select that as it is only the fourth proposal. My settings for the content assist is sorting by relevance (using alphabetically results in the same proposal order). Camel case matches are enabled, substring matches are enabled. I'd like to keep came case matches and substring matches enabled, as those are awesome features. I believe that any proposal that starts with the currently invoked character sequence should be ordered first in the resulting list. Using Eclipse Oxygen Version: Oxygen.3a (4.7.3a) Build id: M20180330-0640 The proposals are in the expected order as in: entrySet ceilingEntry computeIfAbsent computeIfPresent
Sorry, I misspoke in my initial report. When you invoke the content assist on the phrase 'map.ent', then the proposals are in the expected order. Same in Proton, as in Oxygen. The bug appears, when you type 'map.', then open the content assist, and type 'ent' while it is still open. This way the visible proposals are not reordered to present 'entrySet' first, but in the unexpected order I descripbed previously.
I don't see this problem in Eclipse 4.9 plain SDK downloaded from http://download.eclipse.org/eclipse/downloads/drops4/R-4.9-201809060745/. Please check if you can reproduce it with the build mentioned above. Make sure that no other plug-ins are installed.
After downloading Eclipse from: http://download.eclipse.org/eclipse/downloads/drops4/R-4.9-201809060745/ I've tested, and the bug doesn't appear. Version: 2018-09 (4.9) Build id: I20180906-0745 However, please note that as far as I can tell, this is an older build compared to the one I used, which has the build id 20180917-1800 (I assume the build id means the build date). The one the I'm using is this one (where the bug is present): https://www.eclipse.org/downloads/download.php?file=/technology/epp/downloads/release/2018-09/R/eclipse-java-2018-09-win32-x86_64.zip The Eclipse IDE for Java Developers on the downloads page. https://www.eclipse.org/downloads/packages/
I see the same annoying behavior in Version: 2019-03 (4.11.0) Build id: 20190314-1200
(In reply to Bence Sipka from comment #3) > The Eclipse IDE for Java Developers on the downloads page. > https://www.eclipse.org/downloads/packages/ That package contains Mylyn which changes sorting. This might cause the problem. You can try the latest SDK release: https://download.eclipse.org/eclipse/downloads/drops4/R-4.11-201903070500/ or the latest milestone: https://download.eclipse.org/eclipse/downloads/drops4/S-4.12M1-201904110625/ or try that latest Java package (which contains Mylyn): https://www.eclipse.org/downloads/packages/release/2019-06/m1/eclipse-ide-java-developers