Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Block Library: Avoid column width auto-adjustment when sibling width changes #19515

Merged
merged 11 commits into from
Feb 10, 2020

Conversation

aduth
Copy link
Member

@aduth aduth commented Jan 8, 2020

Partially addresses #16370

This pull request seeks to remove auto-adjustment of column widths when updating the width of another column. The intention is to remedy current confusion which can occur when trying to set each column to a specific width, where currently the auto-adjustment can yield an unexpected behavior. With these changes, the auto-adjustment behavior still exists to reallocate space when adding or removing a column from the Columns wrapper block, but specific width updates to individual columns should no longer have any side effects.

This also means there is no longer a guarantee that the sum total of widths of the columns are necessarily equal to 100%. In some cases, this can yield some benefit. For example, it is now fairly simple to create a set of columns which only occupy a fraction of the available space. Additionally, given that columns are not assigned with an explicit width, it allows for columns which are not assigned an explicit width to grow to occupy the available space, rather than forcefully being assigned an explicit width value.

To this last point, this can potentially cause some confusion of its own, since when adjusting the width of a column, you may still observe that the other columns grow or shrink automatically, but only because they are not (yet) assigned with an explicit width, and thus adjust themselves flexibly.

In the discussion of #16370, it was mentioned to consider including a warning notice when the columns were set to widths which did not sum to 100%. This has not yet been implemented in this branch. It could be included here, or as part of a separate pull request, or not at all (depending on feedback to the usability of the proposed changes).

Testing Instructions:

Repeat scenarios from #16370, #19112, or #17918, observing whether you encounter any challenges.

@aduth aduth added [Type] Enhancement A suggestion for improvement. [Block] Columns Affects the Columns Block labels Jan 8, 2020
@aduth
Copy link
Member Author

aduth commented Jan 8, 2020

You can test this live at: http://gutenberg.run/19515

@mapk
Copy link
Contributor

mapk commented Jan 8, 2020

My initial test proved this to be functional and harmless. I set a Columns block with 3 columns; 30%, 50%, and 60%. The columns stayed aligned with the content width, and when changing it to wide-width, they still worked a flowed really well.

Editor screenshots:

Screen Shot 2020-01-08 at 2 51 17 PM

and wide-width

Screen Shot 2020-01-08 at 2 52 19 PM

As a user, while I was completely out of the 100% range, the columns still looked fairly correct to me. I was quite happy with the results. I still don't like jumping around from column to column to adjust the sizing, but this end result worked wonderfully.

@aduth
Copy link
Member Author

aduth commented Jan 9, 2020

I still don't like jumping around from column to column to adjust the sizing

Yeah, in testing the behavior here, I made a remark at #19397 (comment) to a particular block selection behavior which was changed there. I'd not go so far as to say it's a regression, but especially when trying to resize columns immediately after inserting the Columns block, it can be a pain to try to select them.

@aduth
Copy link
Member Author

aduth commented Jan 21, 2020

@talldan If you have a free moment, would you be able to give this one a pass at code review? Pinging since you had given some eyes to a previous, related pull request which sought to solve the same problem in a different way (#16822 (comment)).

Copy link
Contributor

@talldan talldan left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code looks good, and in terms of usability it's now much easier to set manual width 👍

To this last point, this can potentially cause some confusion of its own, since when adjusting the width of a column, you may still observe that the other columns grow or shrink automatically, but only because they are not (yet) assigned with an explicit width, and thus adjust themselves flexibly.

@aduth I agree with that, that is potentially confusing.

What I imagine should happen is that any column with an unset width takes up the remaining space (sharing it equally with any other unset columns). That way if I had four columns, made the first three 25%, the last one would automatically be 25%.

What seems to be happening at the moment is that a column with an unset width by default has a flex-basis: calc( 50% - 16px ), so it's like its set to just under 50% by default.

I still think this PR should be merged as it's a nice improvement.

@aduth
Copy link
Member Author

aduth commented Jan 23, 2020

@aduth I agree with that, that is potentially confusing.

@mapk Do you have any thoughts on this and whether it's a problem worth worrying over? I can see arguments both ways actually, where it's either potentially confusing in how some of the columns are still flexibly arranged, vs. convenient in cases where I might want those remaining columns to be flexibly arranged.

What I imagine should happen is that any column with an unset width takes up the remaining space (sharing it equally with any other unset columns). That way if I had four columns, made the first three 25%, the last one would automatically be 25%.

This could work, and in fact is basically the same as what we have today, except that I assume we'd only want it to redistribute the width on the first time the block changes from a state of unset widths to explicit widths, and avoid further redistribution from that point forward.

@paaljoachim
Copy link
Contributor

paaljoachim commented Jan 25, 2020

I am testing the playground (very neat btw) and had a difficult time selecting a column. I had to change the tool to select and then it worked. I added 3 columns. Click the first column added 25%. Clicked the 3rd column added 25% here as well. Clicking the middle column I kinda expected to see the percentage width numbers of 50%, but no numbers are listed.

It would be great to have an edit icon one can click to have the various visual initial column options show up again, so one can click a new column layout. For instance this time selecting a bigger column on the right and two smaller on the left.

Btw I also look forward to being able to drag the sides of a column to resize it.

I could of course mention a lot of other column features as well that do not belong in this issue. Such as gutter between columns, background color and on and on...:)

@aduth
Copy link
Member Author

aduth commented Jan 29, 2020

I am testing the playground (very neat btw) and had a difficult time selecting a column

I think there might have been some unrelated issues at the time the branch was created. It might be better in master? I know it's not a great experience in general. That said, the changes here don't really (and aren't intending to) have an impact for better or worse on the column selection experience, but rather the width adjustment behavior.

Click the first column added 25%. Clicked the 3rd column added 25% here as well. Clicking the middle column I kinda expected to see the percentage width numbers of 50%, but no numbers are listed.

I think this reinforces what @talldan and myself expressed as possible concern, where perhaps we should force all columns to have a width, at least the first time the blocks change from "fully flexible" to any of the blocks having a specific width.

@timhibberd
Copy link

Hi @aduth I ran a few tests via your tester at http://gutenberg.run/19515 and I can see the benefits of the non-100% approach. If you set percents in all the boxes then you can achieve a variety of effects. 10% | 10% | 10% gives you a total width of 30% with a 70% right blank area. 50% | 50% | 50% gives you the equivalent of 100% in equal thirds.

There are a few anomalies though from a U.X. perspective:

  1. If I create a new default 3-column instance then click the up-arrow on the width control in the column settings it changes the width to 0.1 rather than the default width + 0.1 (so 50.1% as an example) as I would have expected.

  2. When I now adjust the first block from 0.1% to 90% I don't get the effect I would expect...the other two boxes do not appear to be 5% wide each. In fact, I don't know what their widths are since this information is not displayed. I also don't know what the total non-100% width is as a reference point.

  3. When I then adjust the first block to 100%, I have no idea what size the other two blocks are now.

  4. When I then adjust the first block to 140%...the first block gets really small!!!

  5. When you set the width to 30% | 30% | 30% you get the equivalent of 100% even though you should get the equivalent of 90%. If you change the widths to 29% | 29% | 29% you get a total width of 87% as expected.

  6. If you have two column blocks on the page one after the other, the second column block's editor selector field overlaps the first column block's selector field. So if you hover below a block in the first column block you will actually select a block from the second column block. An end-user who is used to clicking below a column to select it in the editor will be frustrated if they do not know they can also click above the column box to select its focus.

  7. If you set the blocks to 1% | 1% | 1% you only have a small sliver of an encapsulating editor box to click on to set a specific block's focus.

Although the dynamic total %width approach looks promising I'm not sure how well it will be received by the average end-user or average designer who is used to working to a pre-set columns proportion and a mental reference point of 100% total width. The lack of visibility of the current widths and the conceptual floating width % leaves the end user / designer uncertain as to what the widths are unless they set all of them manually. If certainty can only be achieved by setting ALL of the boxes manually, then the simplest U.I. is one that re-inforces that convention...use the default or set them all.

As per [my previous suggestion](#17918 (comment) and #17918 (comment).) I would humbly suggest we follow historic convention and stick with:

  1. User accepts the default or sets the fields manually

  2. All widths are visible. Default widths are visible as fractions.

  3. Total width is 100%

  4. The units can be toggled from fractions to percent to support 1/3 | 1/3 | 1/3 correctly.

  5. Adjusting the width of a block reduces the right sibling block proportionally (with the exception that the far right block reduction adjusts its left sibling block proportionally)

image

@mapk
Copy link
Contributor

mapk commented Feb 3, 2020

@mapk Do you have any thoughts on this and whether it's a problem worth worrying over?

Sorry for a late response here. I messed around with this again and agree with @talldan's comment here:

What I imagine should happen is that any column with an unset width takes up the remaining space (sharing it equally with any other unset columns). That way if I had four columns, made the first three 25%, the last one would automatically be 25%.

While setting the widths currently, the undetermined width on the columns seems to be very out of relation to the other widths.


When setting each column width, the last column was abnormally wide.

columns 2020-02-03 14_40_53


When "reseting" the last column's width, I expected it to remain at the left over percentage amount, but instead it went wider than expected again.

reset 2020-02-03 14_44_13

@timhibberd
Copy link

Has anyone run this past the accessibility team? The new approach proposed here does not make visible the actual default column widths or the notional total width% in this new floating width approach. How is a user / designer with a screen reader going to adjust column widths?

@jasmussen
Copy link
Contributor

Took this branch for a quick spin and inserted a 3 column block. Here's the behavior I noticed:

  • By default, no column(s) have percentage widths defined
  • I type in an explicit number in one of the columns, it does not cause the other columns to get an explicit size
  • I type in explicit numbers in the other columns, and those are respected also

I am possibly in the minority, but I wanted to explicitly voice support for this approach. It respects explicitly input numbers, and makes it more predictable how columns are sized, when done so manually. As part of a recent self imposed empathy challenge I tried creating some block patterns, and the following is a group containing two column blocks in succession:

Screenshot 2020-02-04 at 09 12 27

The first column block has 2 columns, the next one has 3, and it was exceptionally painful to get the 33% value I explicitly input in the 3rd columns of both, to stick.

As far as I can tell, there's only one downside to this PR. Unless all three column widths are somehow defined, only those actually defined are correctly sized. Here, the first column is set to 80% width, and the 2 remaining columns are unset:

Screenshot 2020-02-04 at 09 21 53

But obviously the sizes aren't correct. Is this a limitation of the flex based nature of the columns block, or can flex-shrink make it behave more like a table?

@jasmussen
Copy link
Contributor

Hi @timhibberd, can you elaborate on this?

Has anyone run this past the accessibility team? The new approach proposed here does not make visible the actual default column widths or the notional total width% in this new floating width approach. How is a user / designer with a screen reader going to adjust column widths?

If a column does not have a set width, barring what's possibly a bug, an empty text input field should be correct. As for actualy setting these values using a screen-reader, this should be no different from what the behavior is in master.

cols

@timhibberd
Copy link

timhibberd commented Feb 4, 2020

@jasmussen - my comment about accessibility is related to the combination of no display of the actual column widths coupled with a floating concept of total width as a percentage being some number other than 100%. I have eyesight and I'm not sure what the widths are. I'm not sure how a screen-reader would figure out what I can't.

Please have a look at my post that preceded yours. This lists all the issues with the current implementation and my strong concern that this approach, though clever, will not be understood by the end-users. This is a significant departure from conventional UI for a Columns editor with no end-user focus testing and no accessibility team input. I proposed a conventional approach that addresses all the outstanding issues using conventional tactics. No one has commented on it pro / con though. Happy for you to be the first :-)

@jasmussen jasmussen added the Needs Accessibility Feedback Need input from accessibility label Feb 4, 2020
@jasmussen
Copy link
Contributor

Please have a look at my post that preceded yours. This lists all the issues with the current implementation and my strong concern that this approach, though clever, will not be understood by the end-users.

Andrew is skilled in best accessibility practices, I did not get a sense that any concessions were suggested, nor should any be made. I added an "Needs Accessibility Feedback" label to give this visibility and sanity check this further.

I think we may be talking past each other, though: what I'm suggesting is that the the clever approach — automatically calculate and explicitly set column-widths — is possibly a source of confusion, as it overrides user-set values.

And instead, that a simpler approach, behaving like a standard HTML table, would probably be preferable. That is — if a column has an unset value, it literally does not have a width that we can surface visually, because that width is calculated by the browser.

@timhibberd
Copy link

@jasmussen - I agree with your point that the browser would choose the width so displaying a visible width is potentially inaccurate. The point I was trying to make is that this implementation has an arbitrary and unstated concept of what the total width is. This will weird out end-users who think in terms of 100% (except footballers who believe 110% is possible) and designers who want to specify a verifiable percentage. I have dealt with some very pedantic clients who require pixel-perfect. renderings. I also pointed out that at various settings I have no idea what the widths of the unspecified columns are. Set the first column to 100% and you tell me what the widths are of the other two. You and I may not care as long as it looks good, but I know some clients who will want to know exactly what the widths are. This may also become more of an issue when we allow the inter-column margins to be adjusted. So my suggestion was...make the UI re-inforce that if you override the default you set all three columns or if auto-set then once you set the first column the other two columns split the difference.

@jasmussen
Copy link
Contributor

jasmussen commented Feb 4, 2020

this implementation has an arbitrary and unstated concept of what the total width is

I may be giving too much creedence to the table block. It truly can get you into trouble if you really want to watch the world burn:

table

The equivalent here is to set 80% in all three columns, which obviously is more than a hundred. What happens then? In the table example, this breaks it. It seems like we could let the columns block be 240% wide, causing a horizontal scrollbar. It might not be ideal, but it would be what the user asked for.

It becomes a bit of a question of how much handholding to do. Take the paragraph block, you can make this atrocity happen:

Screenshot 2020-02-04 at 11 10 37

The text is barely legible, no-one should do this. While we don't prevent the user, we do show a warning note about the contrast.

Would a similar warning note, added to a Columns block with more than 100% total width as defined by the user, mitigate the trouble users can get themselves into?

@aduth aduth force-pushed the remove/16370-column-auto-adjust branch from 0b2df9f to 92b8fd4 Compare February 7, 2020 01:23
@aduth
Copy link
Member Author

aduth commented Feb 7, 2020

Last-ditch effort, though see note below about potential blocker:

  1. Display radio control for user to choose between flexible and explicit percentage widths, in order to address concerns about ambiguity of overloading the single range control to represent both flexible and percentage widths. Only show range if "Manual". Avoids ever showing an empty input when width unassigned.

Screen Shot 2020-02-06 at 8 09 20 PM

Screen Shot 2020-02-06 at 8 09 35 PM

  1. Show messaging for error state where sum total of percentage widths is not equal to 100%.

Screen Shot 2020-02-06 at 8 09 47 PM

Can be tested at: http://gutenberg.run/19515
Contrast differences against current editor: http://gutenberg.run/20077

Important Note: I don't think the front-end styles of columns are currently correct to support the use-cases we're targeting in allowing mix of flexible and fixed-width columns. Notably, these styles:

flex-basis: calc(50% - #{$grid-size-large});
flex-grow: 0;

This has strange results when trying something like 20 | auto | 20, where one would expect the middle to fill to occupy the remaining 60%, but this doesn't happen.

I might expect something like flex-basis: 100%; flex-shrink: 1; instead, but I'm not sure how the column gutters affect this.


You'll have to forgive my brevity, as it is the end of a long day. I'll plan to revisit when I am of a fresher state of mind.

@timhibberd
Copy link

Nice "Down to the Wire" save :-)

I am good with change 1 and change 2. By forcing the user to switch from auto to manual it reinforces that the user does not need to do anything if they are content with the default. Most of my clients will be content with the block level default choices and won't feel the need to press the "manual button". I also like that change 1 allows the default column units to now be displayed in the input box when in manual mode.

Minor Adjustment: You might consider changing "Manual width" to "Manually adjust width" to be consistent in verbiage with the other radio button option.

NOTE: Your mixed mode seemed to work fine. Even the 20 | auto | 20 at least on the latest Firefox.

Minor Adjustment: If you could make the width input box wider so the default two decimal place numbers (e.g. 33.33) do not look squashed. And at present double-click selects only the integral part rather than the whole number including the decimal part. You probably already had that on your to-do list since you got this revision done lightning fast (kudos for the quick pivot :-)

image

Minor Adjustment: If you set the columns to 25 | 50 | 25 and then increment the 3rd column:

No message pops up for 25.1 through 25.4

image

but an error message will popup at 25.5

image

Side-Note: It would be great to get rid of the double vertical scroll bar as you see in the screenshots if possible.

@jasmussen
Copy link
Contributor

This is impressive work. Thank you all for staying on this. Today is the day, the last day to get a fix in before the beta period. But it's important to remember that work won't stop today. The only constant is change!

To review, the problem is that in the shipping product, users can't set the width of more than 1 column. As soon as they set a single column, the other columns are auto-distributed the remaining space.

This is what I see in the branch:

col settings

This is a good suggestion for adding clarity. Other solutions could be to offer a ButtonGroup that defaults to Auto but can be set to Manual, unhiding the slider. I'm sure there are many other options also.

But this is also a lot of UI complexity to introduce in the last minute. Bugs around scrollbars, finding a labels that wrap well in multiple languages — these all add up to making this change risky to merge at this point.

In hopes of finding the tiniest possible change to what is currently shipping, I created this mock:

col settings alt

The only change is that it sets the placeholder text to "Auto" when unset.

Despite the impressive work that's gone in, I would ship the latter. This would give us the time to polish and iterate the former, based off the work that's already done. Instead of crunching, the changes suggested can be discussed in a relaxed setting, while still allowing us to strongly mitigate the presently shipping bug.

This is not to diminish any of the opinions offered, or any of the work already done.

@aduth
Copy link
Member Author

aduth commented Feb 7, 2020

@jasmussen Reflecting on this, I'm inclined to agree that the recent revisions -- while worthy improvements explorations -- are too much too late to be trying to include. I agree that scaling back to a more modest iteration over the current behavior of the editor is the best option forward. The placeholder is at least a minimal step to absorb some possible misinterpretations of the empty field.

I pushed this in the latest batch of commits.

We still need to address the issue of front-end styling for automatic column width growth mentioned in #19515 (comment).

aduth added 2 commits February 7, 2020 11:05
Avoid situation where adding or removing a Column in a Columns would forcefully assign width to a block which did not previously have a width. Only opt to redistribute widths to explicit numbers if all blocks already have widths.
@aduth
Copy link
Member Author

aduth commented Feb 7, 2020

I pushed another set of changes which aims to try to resolve the issue with front-end styling (it also affected the editor, which I was not previously aware of). The implementation is such that, if a block is not assigned a width, it will maximally grow to occupy whatever space is available. Columns with assigned widths should use that width, and grow no further (same as it was before).

I've been trying a handful of different adaptations and I believe it's working correctly. The gutters can sometimes be puzzling to decipher their impact, though I think these are all correct when you consider the impact that more or fewer gutters should have on each individual column:

image

It's much easier to assess when you normalize to remove the gutters:

image

@timhibberd
Copy link

@jasmussen I agree that setting the placeholder text to "Auto" when unset would be a valuable change for the release.

I would also ask that the second change @aduth proposed which is to display an error message when the total width goes over 100% be added if possible. This is a "must have" to compensate for having a variable percentage associated with the combined width.

I would also strongly recommend that the Reset button be re-labelled to represent what it really does which is to reset all the columns in the block. So "Reset Block" or "Reset All" or "Reset Columns" or "Reset all Columns".

I leave it yours and @aduth capable hands. we can revisit post release 4.5. I liked the auto / manual per-column selection that @aduth came up with in the end and I like the explicit switch from auto to manual.

Let us know if you close this PR and open a new one for post 4.5 discussion. Cheers guys :-)

Copy link
Contributor

@youknowriad youknowriad left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For me, this works way better than what we currently have. I think we should consider landing this and continuing on any other improvements as follow-ups.

@aduth
Copy link
Member Author

aduth commented Feb 10, 2020

Given the choice between allowing the current behavior to persist through another major version vs. the changes as presented here, I too consider it to be an improvement over the current function of the block.

I grant that there remain some potential points of confusion, as discussed and explored for improvement through the history of this pull request. Despite that and in weighing these with the current situation, I believe that the experience of editing columns' widths will be on-the-whole improved.

I propose to proceed by:

  • Merging this pull request
  • Keeping Column width value changes. It's hard to set width. #16370 open as indicator that the experience remains imperfect
  • Identify specific areas of improvement and consider creating individual issues tailored toward addressing those problems
    • These may be the same as discussed and explored here (e.g. "warning notice when widths exceed 100%", "toggle between auto and manual width"), allowing for more time to consider the design and functional implications of these changes.

I recognize the impact that these changes can have, and am very appreciative of all the feedback this has drawn as a result. I'm eager to continue those discussions as we proceed with further improvements to this experience.

I would also strongly recommend that the Reset button be re-labelled to represent what it really does which is to reset all the columns in the block. So "Reset Block" or "Reset All" or "Reset Columns" or "Reset all Columns".

The "Reset" button only does affect the width assigned to the currently selected column, so the current label would be accurate. It was a behavior prior to this pull request that pressing "Reset" in a block could have caused the widths of adjacent blocks to have been reset as well, which is no longer the case after these changes. Interestingly, with the previous behavior, you could still have arrived in a scenario of having the same sorts of mixed set of explicitly-assigned and flexible-width column blocks we have talked through here. This would depend on the order in which a user assigns widths, and the position of the block at which "Reset" is pressed (e.g. 3 columns, set 25% on first, then "Reset" on second, result would have been equivalent of 25% | auto | auto). It would be my hope that by removing those direct side-effects of a width change, we avoid more of these unpredictable scenarios.

@aduth aduth merged commit 26e2bf6 into master Feb 10, 2020
@aduth aduth deleted the remove/16370-column-auto-adjust branch February 10, 2020 14:18
@github-actions github-actions bot added this to the Gutenberg 7.5 milestone Feb 10, 2020
@aduth
Copy link
Member Author

aduth commented Feb 10, 2020

For posterity's sake, I was intentional in creating the history as retaining original commits of the explorations implemented here (later reversed as revert commits), fully expecting that they may serve as valuable reference for future implementations.

See:

  • e25b28b - Show Notice when total width is not 100%
  • 92b8fd4 - Show Columns block range control only when "Manual" width

In reflecting on the implementation of e25b28b, I observe that it may need additional improvements to handle cases where there is a mixed set of assigned/unassigned widths, where a combination of a subset of those columns would exceed 100% (e.g. 60% | auto | 60%). The proposed implementation would bail out of displaying the notice if any of the columns was flexible width, thus not showing a warning for this invalid scenario.

aduth added a commit that referenced this pull request Feb 11, 2020
…columns (#20169)

* Block Library: Columns: Fix equal column growth for unassigned-width columns

* Block Library: Columns: Restore mid-range viewport styles

* Block Library: Columns: Keep margin resets in mid-range viewports

Reference from prior to #19515 : https://github.com/WordPress/gutenberg/blob/7997bf66490bab0f6984a114c22ad125f04b9e89/packages/block-library/src/columns/editor.scss#L69-L75

* Block Library: Columns: Update comment to reference mid-range 2 columns
ellatrix pushed a commit that referenced this pull request Feb 12, 2020
…columns (#20169)

* Block Library: Columns: Fix equal column growth for unassigned-width columns

* Block Library: Columns: Restore mid-range viewport styles

* Block Library: Columns: Keep margin resets in mid-range viewports

Reference from prior to #19515 : https://github.com/WordPress/gutenberg/blob/7997bf66490bab0f6984a114c22ad125f04b9e89/packages/block-library/src/columns/editor.scss#L69-L75

* Block Library: Columns: Update comment to reference mid-range 2 columns
jorgefilipecosta pushed a commit that referenced this pull request Feb 17, 2020
…columns (#20169)

* Block Library: Columns: Fix equal column growth for unassigned-width columns

* Block Library: Columns: Restore mid-range viewport styles

* Block Library: Columns: Keep margin resets in mid-range viewports

Reference from prior to #19515 : https://github.com/WordPress/gutenberg/blob/7997bf66490bab0f6984a114c22ad125f04b9e89/packages/block-library/src/columns/editor.scss#L69-L75

* Block Library: Columns: Update comment to reference mid-range 2 columns
jorgefilipecosta pushed a commit that referenced this pull request Feb 17, 2020
…columns (#20169)

* Block Library: Columns: Fix equal column growth for unassigned-width columns

* Block Library: Columns: Restore mid-range viewport styles

* Block Library: Columns: Keep margin resets in mid-range viewports

Reference from prior to #19515 : https://github.com/WordPress/gutenberg/blob/7997bf66490bab0f6984a114c22ad125f04b9e89/packages/block-library/src/columns/editor.scss#L69-L75

* Block Library: Columns: Update comment to reference mid-range 2 columns
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
[Block] Columns Affects the Columns Block Needs Accessibility Feedback Need input from accessibility [Type] Enhancement A suggestion for improvement.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

7 participants