I updated my Salesforce IDE to the last version (Summer 14 / APIv31) this week and I noticed some issues when saving apex classes that contains characters with accentuation. For example, strings like "não" where saved "n?o", but I only noticed that when I saw the code in browser or updated the class in the IDE via refresh from server.
Doing some research, I discovered a new feature related to the tooling API: https://developer.salesforce.com/page/Force.com_IDE
Set the Tooling API as the default mechanism used for Apex classes and triggers, and for Visualforce pages and components.
When disabling this feature, the issue with accentuation does not occur but the IDE continuously syncs with the server. When re-enabling the feature, the syncs stop and the issue with accentuation comes back.
Anyone had issues with accented characters with the new SFDC IDE?
Update:
I reproduced the issue using a new installation of eclipse + SFDC IDE plugin. Here are the steps:
1) Download de eclipse IDE Kepler version
2) Install the lasted version of the SFDC IDE (v.31)
3) Create a project pointing to a sandbox server. In my case was at CS16.
4) Create a class Test.class
5) Add a comment:
public with sharing class Test {
// não
}
6) Save the class. As the IDE is configured to work offline, I need to save the class using the menu Force.com -> Save to Server.
7) I opened the class via browser, and the comments are ok (no issue with accentuation)
8) Change the project to work on-line
9) Modify the class
public with sharing class Test {
// não aáoó
}
10) Save the class. Now the class is saved automatically to the server.
11) Checking the class again via web, the accentuation is lost.
public with sharing class Test {
// n?o a?o?
}
//não
comment and it round trips with the correct encoding. It might be related to the string encoding Eclipse is using to set the Body.