Community
Participate
Working Groups
Created attachment 79321 [details] Screen shot I got the "Hot Code Replace Failed" dialog as pictured in the screen shot. The wording needs to be refined. See the screen shot for suggestions.
Good points... it is a complex dialog :-) Note, when you select the option to not show the error dialog when HCR fails, you are effectively choosing "continue". We should make this clear. I don't think it's fair to always terminate or terminate/restart without prompting.
overlapping with bug 224245
This is a trivial fix. We should do this for 3.7.
*** Bug 224245 has been marked as a duplicate of this bug. ***
Created attachment 194644 [details] patch Here is a shot at making it a bit easier on the eyes. The move the relevant details to the details section and changes the dialog message to ask the user a question. I also changed the wording on the 'don't nag me again' option to better convey that checking the option implies you want to always continue.
Created attachment 194645 [details] screen shot Here is a screen shot of the dialog you see when you try to hot swap change in a method that cannot be popped off the stack - the main method in a Java program for example. I'm not sure how I feel about the '(always continue)' part on the preference, perhaps better wording would be "Always continue when obsolete methods remain"
The 'Details' button is at the wrong location. It should be a a little bit separated from the choices. The question is wrong. For such a question I would expect a Yes/No dialog. Showing the reason doesn't add much value if it's already in the title (can be put into the Details section). Maybe better check box message: Do not show this &message again but always continue instead
(In reply to comment #7) > The 'Details' button is at the wrong location. It should be a a little bit > separated from the choices. > Agreed, it is currently in the default location ErrorDialog places it > The question is wrong. For such a question I would expect a Yes/No dialog. > > Showing the reason doesn't add much value if it's already in the title (can be > put into the Details section). > We also get this *for free* from ErrorDialog, we could override some extra bits and probably make this go away. > Maybe better check box message: > Do not show this &message again but always continue instead Thats not bad, but I think the 'Always continue when *' is better, since we tailor the preference wording to the reason the hot code replaced failed. For example here are the three messages we currently use: 'Do not show error &when hot code replace is not supported' 'Do not show error &when obsolete methods remain' 'Do not show error &when hot code replace fails' the updated messages would be: 'Always continue &when hot code replace is not supported' 'Always continue &when obsolete methods remain' 'Always continue &when hot code replace fails' each of those three preferences are separate and un-doable on the Java > Debug preference page
The problem when only naming the checkbox message 'Always continue...' is that there is no direct relation to the dialog/message and it's not 100% clear what happens when I press another button than 'Continue'. Obviously the best choice would be: [ ] Remember my selection and don't show this message in the future
I'm OK moving this to 3.8. We could then consider a solution along comment 9.
Does the "Details >>" button ever reveal additional information? If not, then it should just be removed again. The actual information from the dialog should not be moved into the "Details >>" section. It's OK to have some explanations on the page, but the really interesting items on the dialog are: - VM - Launch - Reason This information should not be embedded in a lengthy paragraph, but it should be rendered as the last three lines before the "don't show again" checkbox, so that an experienced user can quickly consume this information without reading the rest of the dialog.