Closed
Bug 518597
Opened 15 years ago
Closed 12 years ago
[Faceted Search] GLODA search results should contain matched message fragment
Categories
(Thunderbird :: Search, defect)
Thunderbird
Search
Tracking
(Not tracked)
RESOLVED
FIXED
Thunderbird 12.0
People
(Reporter: sdwilsh, Assigned: florian)
References
Details
(Whiteboard: [no l10n impact])
Attachments
(2 files, 2 obsolete files)
77.15 KB,
image/png
|
Details | |
6.02 KB,
patch
|
asuth
:
review+
bwinton
:
ui-review+
|
Details | Diff | Splinter Review |
It'd be really helpful if the search results that we are presented with contained the part of the message that contains the matched search term[s] instead of just the start of the message.
Flags: blocking-thunderbird3?
Updated•15 years ago
|
Assignee: nobody → david.ascher
Flags: blocking-thunderbird3? → blocking-thunderbird3+
Updated•15 years ago
|
Whiteboard: [no l10n impact]
Comment 1•15 years ago
|
||
While I agree that this bug would be good, I don't think it really blocks, and I'm reluctant to do a quick fix w/o good test coverage. I know I won't have time for that.
Assignee: david.ascher → nobody
Updated•15 years ago
|
Summary: GLODA search results should contain matched message fragment → [Faceted Search] GLODA search results should contain matched message fragment
Comment 2•15 years ago
|
||
Given comment 1, I'm removing this from the blocking list.
Flags: blocking-thunderbird3+ → blocking-thunderbird3-
Updated•15 years ago
|
Flags: wanted-thunderbird3+
Updated•15 years ago
|
Target Milestone: Thunderbird 3.0rc1 → ---
Assignee | ||
Comment 3•12 years ago
|
||
This patch: - ensures the first match of the body text is visible in the displayed snippet - displays in bold all the matches that are visible in the snippet. This is important for instant messaging (bug 714733) as IM conversation search results are usually completely meaningless if the 5 first lines of a conversation are displayed.
Assignee: nobody → florian
Status: NEW → ASSIGNED
Attachment #589724 -
Flags: ui-review?(bwinton)
Attachment #589724 -
Flags: review?(bugmail)
Assignee | ||
Comment 4•12 years ago
|
||
Comment on attachment 589724 [details] [diff] [review] Patch >+ // Unicode characters with a code point > 128 are 2 bytes long. >+ // Unicode characters with a code point > 2048 are 3 bytes long. Assume these comments read > 127 and > 2047 (I've fixed that locally already).
Comment 5•12 years ago
|
||
Comment on attachment 589724 [details] [diff] [review] Patch The logic looks good, and it's great to have this feature! The caveat is that you need to handle the case where we are displaying a gloda query that is not a fulltext query, in which case stashedColumns is not going to be present. (Its initialization is predicated on query.options.stashColumns being present.) You can initiate such a query by selecting one of the auto-complete options in the global search box such as a contact or a tag name. Presumably you would still just want to display the first 5 lines or whatever we do now in such a case? My only real nit is that the rest of the file uses the Thunderbird idiom of: for each (let [, thing] in Iterator(things)) instead of this patch's use of: for each (let thing in things) The reason most Thunderbird code uses the former (or index subscripting) is out of paranoia that some jerk is going to mess with Array.prototype and break that logic. Since this code lives in its own page, it's not really a concern, but if you could make the change for consistency, that would be great. The UX review can still happen on this patch unless Blake wants something different for the non-fulltext case too.
Attachment #589724 -
Flags: review?(bugmail) → review-
Assignee | ||
Comment 6•12 years ago
|
||
Attachment #589724 -
Attachment is obsolete: true
Attachment #590160 -
Flags: ui-review?(bwinton)
Attachment #590160 -
Flags: review?(bugmail)
Attachment #589724 -
Flags: ui-review?(bwinton)
Assignee | ||
Comment 7•12 years ago
|
||
Blake, here's a screenshot, just in case it can help you for the ui-review ;).
Comment 8•12 years ago
|
||
As far as UI is concerned, may I suggest we look into what FF is doing with in-page search, ie apply a different background/foreground to the found words ? This really stands out
Comment 9•12 years ago
|
||
Comment on attachment 590160 [details] [diff] [review] Patch v2 I agree that we need to make the terms stand out a little more. I would also like to see the terms highlighted when they appear in the subject or email address. Finally, one of my matches started off with: … This is the *matched* text and I think we should get rid of the extra newline in that case. Thanks, Blake.
Attachment #590160 -
Flags: ui-review?(bwinton) → ui-review-
Assignee | ||
Comment 10•12 years ago
|
||
(In reply to Blake Winton (:bwinton - Thunderbird UX) from comment #9) > I agree that we need to make the terms stand out a little more. This new patch changes the highlights as discussed on IRC. > I would also like to see the terms highlighted when they appear > in the subject or email address. I would prefer if we could handle that in a separate bug. > Finally, one of my matches started off with: > … > > This is the *matched* text > > and I think we should get rid of the extra newline in that case. I added a special case for this.
Attachment #590160 -
Attachment is obsolete: true
Attachment #590232 -
Flags: ui-review?(bwinton)
Attachment #590232 -
Flags: review?(bugmail)
Attachment #590160 -
Flags: review?(bugmail)
Comment 11•12 years ago
|
||
Comment on attachment 590232 [details] [diff] [review] Patch v3 Woo!
Attachment #590232 -
Flags: review?(bugmail) → review+
Comment 12•12 years ago
|
||
Comment on attachment 590232 [details] [diff] [review] Patch v3 Assuming this matches the last screenshot I saw on IRC, ui-r=me! :) Later, Blake.
Attachment #590232 -
Flags: ui-review?(bwinton) → ui-review+
Assignee | ||
Comment 13•12 years ago
|
||
http://hg.mozilla.org/comm-central/rev/916cf00f9afa
Status: ASSIGNED → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Target Milestone: --- → Thunderbird 12.0
You need to log in
before you can comment on or make changes to this bug.
Description
•