497: Easy to forget to save/publish comments

jcasares
chipx86
chipx86
Dec. 27, 2008
This could be broken into multiple issues, but the overarching goal is to
ensure that comments are not lost or go unpublished. 

FORGET SAVING DIFF COMMENT
Write a long comment for a line of code, switch to publish tab without
saving, click Ship It and Publish. Comment is lost forever.
Idea: always save comments unless user explicitly discards.

FORGET PUBLISHING DIFF COMMENT
Write a long comment for a line of code, save it, close window. Comment was
not published.
Idea: ask when leaving/closing page if user wants to publish.
Idea: after a comment is saved, show a Publish button right there.

FORGET PUBLISHING MAIN REQUEST PAGE COMMENT 
Add a comment to a long thread in the review request page. Click OK to
save. Publish button at top of thread appears but is not visible in
browser.  Close window. Comment was not published.
Idea: Ask when leaving/closing page if user wants to publish. 
Idea: Place a Publish button next to comment and/or end of thread, not just
at beginning of thread.
jcasares
#1 jcasares
Another case:

CLOSING BROWSER [TAB]
NAVIGATING TO ANOTHER PAGE
Write a long comment. Close tab/browser or click on random link/back button.
Idea: Confirmation dialog when closing or leaving review board. 
#2 cor****@gmai***** (Google Code) (Is this you? Claim this profile.)
It's also extremely tedious and non-intuitive to make changes to the Summary field. 
I repeatedly made changes, but every time I did anything it would discard them. 
Finally I found hitting enter at the end of the line would save the change as a
draft.  Maybe add standard green check (save) and red cross (discard) at the end of
the field to help.

I agree that it's much too easy to lose comments without them being saved as drafts,
however so far I'm not so worried about having the saved drafts stay unpublished.
chipx86
#3 chipx86
Regarding the buttons, see also bug 600.
  • +Component-Reviews
    +Component-DiffViewer
    +Usability
david
#4 david
  • +Confirmed
chipx86
#5 chipx86
These are being addressed in my UI rewrite.
  • -Confirmed
    +Started
  • +chipx86
chipx86
#6 chipx86
  • +Milestone-Release1.0
chipx86
#7 chipx86
There's a lot in here, but the core issues about losing changes should be fixed in
r1644. Review Board now lets you know you might lose things and gives you a chance to
cancel moving away from the page.
  • -Started
    +Fixed