Summary:
Fixes T3814. Broadly, remarkup tables in inline comments did not work properly. I ran into several messes here and cleaned up some of them:
- Some of this code is doing `JX.$N('div', {}, JX.$H(response.markup))`, to turn an HTML response into a node, passing that around, and then doing junk with it. This is super old and gross.
- The slightly more modern pattern is `JX.$H(response.markup).getFragment().firstChild`, but this is kind of yuck too and not as safe as it could be.
- Introduce `JX.$H(response.markup).getNode()`, which actually expresses intent here. We have a bunch of `getFragment().firstChild` callsites which should switch to this, but I didn't clean those up yet because I don't want to test them all.
- Switch the `JX.$N('div', {}, JX.$H(response.markup))`-style callsites to `JX.$H(response.markup).getNode()`.
- `copyRows()` is too aggressive in finding `<tr />` tags. This actually causes the bug in T3814. We only want to find these tags at top level, not all tags. Don't copy `<tr />` tags which belong to some deeper table.
- Once this is fixed, there's another bug with mousing over the cells in tables in inline comments. We select the nearest `<td />`, but that's the cell in the remarkup table. Instead, select the correct `<td />`.
- At this point, these last two callsites were looking ugly. I provided `findAbove()` to clean them up.
Test Plan: Created, edited, deleted, moused over, and reloaded a revision with inline comments including remarkup tables. Used "Show more context" links.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T3814
Differential Revision: https://secure.phabricator.com/D6924
Summary: If `jsxmin` is not available, use a pure PHP implementation instead (JsShrink).
Test Plan:
- Ran `arc lint --lintall` on all JS and fixed every relevant warning.
- Forced minification on and browsed around the site using JS behaviors. Didn't hit anything problematic.
Reviewers: vrana, btrahan
Reviewed By: vrana
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D5670
Summary:
This is straightforward, except that `form.submit()` does not call onsubmit handlers. Create a `didSyntheticSubmit` event and have everything which listens for form submits listen for it too.
Fixes T704.
Test Plan: Hit control + enter in inline comments, main commetns, Pholio, conpherence. Verified it triggered appropriate JS (workflow / special behaviors).
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T704
Differential Revision: https://secure.phabricator.com/D4704
Summary:
This lets you delete inlines from the preview at the bottom of
the page, instead of hunting for them through the diffs.
There is not yet a keyboard shortcut.
The mechanism for updating the inlines in the diffs is kind of a hack
and I'm sure I'm special-casing way too much, but at least it works.
Test Plan:
Load revision with many diffs. Create inlines all over the
place. Delete them all. Mwahaha.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T1433, T1431
Differential Revision: https://secure.phabricator.com/D3131
Test Plan:
Edit inline comment.
Clear its text.
Hit Cancel.
Try to find Undo link on the usual place with no success.
Repeat for same text.
Repeat for different text, this time with success.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: https://secure.phabricator.com/D1739
Summary:
The current approach of using a modal overlay dialog to create/edit inline
comments is pretty silly. Use an inline textarea instead.
This element isn't perfect and we have some mild modalness issues, but I think
it's better than the silly thing we've got going on right now. We can keep
poking it as people break it.
Test Plan:
- Created comments; submitted and undid them in empty and nonempty states.
Used undo for nonempty states + cancel.
- Edited comments; saved and canceled them. Used undo for changed state.
- Replied to comments; yada yada as above.
- Deleted comments.
- Did various modal trickery where I clicked "Reply" on something else with a
dialog already up, this very mildly glitches but I think it's not a big issue.
Reviewers: vrana, btrahan, Makinde, nh
Reviewed By: vrana
CC: aran, epriestley
Maniphest Tasks: T431
Differential Revision: https://secure.phabricator.com/D1716
Summary: See D939. Regardless of what we do there, these will break, and they're
pretty silly anyway (see the giant caveat comments in the second one).
Test Plan: Clicked a direct-jump comment link, did save/cancel for inline
comments.
Reviewers: phil, cpojer, tomo, mroch
Reviewed By: phil
CC: aran, phil
Differential Revision: 940
Summary:
We use a 'null' row to indicate the element should be appended to the end of the
table (otherwise, it is prepended to the row in question), but also derive the
table from the row. This needs more cleanup in general but fix the immediate
issue at least.
Test Plan:
Added an inline comment to the last line of a file.
Reviewed By: jungejason
Reviewers: tuomaspelkonen, jungejason, aran
CC: aran, jungejason
Differential Revision: 425
Summary:
When a user hits 'cancel' on a 'new', 'edit', or 'reply' operation, add a little
"Changes discarded. __Undo__" insert so they can get their change back. No undo
for delete since there's an explicit prompt. Once this lands we can make
'escape' work again to close dialogs.
This change started feeling really good when I was merging all the duplicate
code and making things more consistent, but by the time I started writing client
rendering it felt gross. I'm not really thrilled with it but I guess it's a step
forward? The feature seems pretty OK in practice. Let me know how much barfing
this causes and I can try to remedy the most acute concerns.
This also fixes a bug where replies always (?) appear on the 'new' side of the
diff (I think?).
Test Plan:
Applied 'new', 'edit', 'delete' and 'reply' operations, pressed 'cancel' and
'okay' in each case, with and without changing text where relevant. All
behaviors seem to conform with expectations, except that canceling out of 'edit'
without changing the text gives you an option to undo when it shouldn't really.
There's no super easy way to get at the original text right now.
Reviewed By: aran
Reviewers: aran, jungejason, tuomaspelkonen
CC: simpkins, aran, epriestley
Differential Revision: 406
