Community
Participate
Working Groups
if usermatchmode is on and you assign the bug to somebody else by putting only part of the username in the assignee field a window pops up asking for a confirmation of the possible found users. hitting the confirm button will result in a "File Not Found": "The file /process_bug.cgi cannot be found. Please check the location and try again."
Created attachment 55762 [details] popup after a submit this screenshot shows the popup once I set a new assignee and submitted the ticket. In the upper left corner you see already an error message. I see this every time this kind of window pops up. No matter whether I connect to my out-of-the-box bugzilla or to our company's bugzilla. the next attachmend will show what happened once I click on "continue"
Created attachment 55763 [details] once I hit continue or any other link in the popup window
I assume that you are able to use full email addresses? For now I added a FAQ entry: http://wiki.eclipse.org/index.php/Mylar_FAQ#Bugzilla_Connector_troubleshooting
Rob, I found this bug and want to know if we should resolve this. For me it looks like this can somehow related to bug 152415. The actual problem is that we get an html page with the possible users as theresult of the post method. But the actual code only handle errors. If we fix this we should first come up with the correct workflow. Maybe we do this after bug 202879 and bring up a special version of a Content Assist. But keep in mind that we can have more then one attribute to process (Assigned to, QA Contact, add CC). regards Frank
Created attachment 84912 [details] mylyn/context/zip
(In reply to comment #4) > Rob, > > I found this bug and want to know if we should resolve this. For me it looks > like this can somehow related to bug 152415. > > The actual problem is that we get an html page with the possible users as > theresult of the post method. > But the actual code only handle errors. > > If we fix this we should first come up with the correct workflow. > Maybe we do this after bug 202879 and bring up a special version of a Content > Assist. But keep in mind that we can have more then one attribute to process > (Assigned to, QA Contact, add CC). > > regards > > Frank In the case of usermatch enabled and the user enters partial names for more than one field, the followup dialog could present all fields with suggestions returned by bugzilla, then either resubmit with the values selected or if possible complete the original submission correctly. Marking bug#202879 as a subtask of this.
Created attachment 85823 [details] Patch with without ui Patch for 202879 also included to demostrate how all fit together. http://wiki.eclipse.org/Mylyn_Bugzilla_Connector#Langugage_Configurations updated for new error_confirm_match command. I must do the cleanup of my ui imlementation before I can submit this to. (Hope that I can do this until 6. January 2008)
Created attachment 85824 [details] mylyn/context/zip
Created attachment 86228 [details] Patch with UI please verify and keep in mind that Patch for 202879 is included to.
Created attachment 86229 [details] mylyn/context/zip
Frank, can you please use convention "bug NNNN" when referring to other issues, so the hyperlink detector would be able to match on a standard pattern. Thanks.
Frank, I hope to put some time into reviewing patches this week. Sorry for the delay.
Created attachment 89687 [details] Patch This patch is include a patch for bug#210283 and the previous path. This was done because I think the are related. Patch for bug#202879 is also included to demonstrate how all fit together. All this must wait until bug#179254 is resolved. If you work on this I can separate this one patch to the corresponding bugs for a better review.
Created attachment 89688 [details] mylyn/context/zip
Need to defer: http://wiki.eclipse.org/index.php/Mylyn/3.0_Plan#Deferred_Items
Due to the 3.0 editor refactoring this patch is now stale. Before diving back in here Frank we'll need to discuss how this is going work with the new editor.
*** Bug 255684 has been marked as a duplicate of this bug. ***
Rob, we should also consider a user cache similar to the one implemented for JIRA. This would also make it easier to implement content assist that would match full names.
Created attachment 120122 [details] possible patch for 3.0 editor please review this implementation
Created attachment 120123 [details] mylyn/context/zip
Frank, I agree in principle with the Bugzilla side of the plumbing to make this work. Since it doesn't apply I didn't do a thorough UI review (i.e. how errors and suggestions manifest themselves in the editor). You may want to hold off on going to deep on UI until bug 236504 is worked out.
Created attachment 135057 [details] new Version Here is a new Patch using th API from bug#202879#c18 Other things: In BugzillaClient I add some //$NON-NLS-
Created attachment 135058 [details] mylyn/context/zip
The general idea of this patch looks good to me. I think we need some more discussion about the UI presentation of the errors. We might want to use ControlDecorations to highlight fields with errors. Rob, maybe it would make sense to apply the core part of the patch and leave the ui bits for the next milestone?
Created attachment 136649 [details] patch with ControlDecoration now we use ControlDecorations to highlight fields with errors but after the first keystroke in that field we remove the decoration. Is this OK?
Created attachment 136650 [details] mylyn/context/zip
Cool stuff Frank. Patch is a little stale but I'm not worried about merging since little else will be changing in the client for 3.2 so if you and Steffen are agreed, I'd like to apply this in is entirety immediately after 3.2 (once we've reviewed the UI).
(In reply to comment #27) > Cool stuff Frank. Patch is a little stale but I'm not worried about merging > since little else will be changing in the client for 3.2 so if you and Steffen > are agreed, I'd like to apply this in is entirety immediately after 3.2 (once > we've reviewed the UI). Yes this is OK for me, should I create an updated patch after 3.2?
That would be great!
(In reply to comment #29) > That would be great! > Should we change the Target Milestone to 3.2.1?
My sense is that this should go into 3.3. since this it is a new feature.
Is this going to make 3.3?
Unfortunately we'll have to move this to a future release since it will require extensions in the task editor.
*** Bug 210283 has been marked as a duplicate of this bug. ***
*** Bug 214529 has been marked as a duplicate of this bug. ***
(In reply to comment #33) > Unfortunately we'll have to move this to a future release since it will require > extensions in the task editor. Can include this in 3.4?
Yes, we can make API enhancements now and take that on for 3.4.
Created attachment 162614 [details] patch for Version 3.4 updated patch. The UI supports decoration for the error. Content assist is not used, only a tooltip show the leagal values.
Created attachment 162615 [details] mylyn/context/zip
Rob, can you please review Frank's patch and consider it for 3.4?
Will do.
Should I create a new patch before you review this patch? I found some unmatched hunks.
I think it would be best if we took a review pass through the current patch first. I'll try to setup a meeting with Rob to go over it but we are both pretty overloaded right now so I can't promise when we'll be able to get to that.
Frank, can we postpone this to 3.6? I'm afraid we won't get to a review for 3.5.
(In reply to comment #44) > Frank, can we postpone this to 3.6? I'm afraid we won't get to a review for 3.5. Milestone change to 3.6! Should I update the patch after 3.5 is out?
Yes, that would be great.
Created attachment 191048 [details] patch V8 Here is the patch for 3.6! I changed the UI! Now you get not only the ControlDecoration you get the values from the html page in the Proposals of the ContentAssistent
Created attachment 191049 [details] mylyn/context/zip
Created attachment 196758 [details] aaaa sssss
Created attachment 196759 [details] testAttachment comment to testAttachment
Sorry, I didn't get around to reviewing. Let's consider this early in the 3.7 cycle.
Created attachment 198194 [details] patch V9 Steffen, here the recreated patch! Hope that we can soon apply this version.
Created attachment 198195 [details] mylyn/context/zip
Steffen, should I recreate the patch now?
Sure, can you commit this feature into a branch in Git?
I create a new Git branch http://git.eclipse.org/c/mylyn/org.eclipse.mylyn.tasks.git/log/?h=bug%23168204 so you can review this.
Created attachment 200621 [details] mylyn/context/zip
Frank, can you rebase the changes against the latest master and push them into a Gerrit review?
http://review.mylyn.org/51 created!
Created attachment 203550 [details] mylyn/context/zip
*** Bug 238186 has been marked as a duplicate of this bug. ***
Frank, please split the code review into two parts: Bugzilla Core and the UI changes. For the core we need to add a proper class for the error response. It should basically have a data model with the proposals returned by Bugzilla instead of a generic map. There also need to be test cases to verify that the expected proposals are returned and parsed. Once that is in place we can take a look at the UI portion.
Steffen, patch set 2 is now in change 51
Steffen, can you please tell me if I can commit this patch set and continue with the UI part.
From a quick glance, the implementation looks good. Before pushing, can you look into breaking up the test case into several smaller ones so that each test case only verifies one concern?
Steffen, I create patch set 3. We now have 6 tests! Can you please tell me if it is OK to push this into HEAD.
Thanks! Looks good to me. Please rebase the change and verify that all tests pass before pushing.
I now start with the UI
Steffen, I create http://review.mylyn.org/214 for the UI part. Can you please review.
Created attachment 209855 [details] mylyn/context/zip
Created attachment 210155 [details] overlapping control decoration The error decoration should displayed in the lower left (see PDE MANIFEST editor for an example). We also need to increase spacing between the label and input field.
This looks pretty good. I have added a few comments to the review. Here are a few more suggestions: * Please split presentErrorToUI into several methods to make it easier to understand. * I'm not convinced that the attribute editor should hold on to the ContentAssistCommandAdapter. What if we register a second ContentAssistCommandAdapter that only knows about the error proposals. Would that override the one that's already registered?
Steffen, I create an new patchset with your comments (comment#71 and comment#72). I try to use an second ContentAssistCommandAdapter but did not find a way that this works. (My try is the BugzillaTaskEditorPage.java line 718-729).
Created attachment 210244 [details] mylyn/context/zip
Created attachment 210607 [details] truncated decoration
Thanks for the update. I have commented on the review. The UI looks pretty good now. I found one minor nits: * A few pixels of the decorations are cut off. The spacing between the label text field needs to be increased slightly. When there is a confirm match could we replace the text in the input field with the proposed value? I think this would be more intuitive than populating content assist.
I create patchset 3
Created attachment 210682 [details] mylyn/context/zip
Thanks. That looks good. Were you able to fix the cut off decorations?
(In reply to comment #79) > Thanks. That looks good. Were you able to fix the cut off decorations? No that is not included. Should i try to fix this tonight?
(In reply to comment #80) > (In reply to comment #79) > > Thanks. That looks good. Were you able to fix the cut off decorations? > > No that is not included. Should i try to fix this tonight? Yes, that would be great. It should be a simple fix. Please feel free to submit with that fix included.
Done
Created attachment 210942 [details] Screenshot
Created attachment 212020 [details] screenshot