1224804 - [css-align] Make ::-moz-table-outer inherit 'justify-self' too

Status: RESOLVED FIXED

(Core :: Layout: Tables, defect)

Description

[Mats Palmgren (inactive)]

I saw that ::-moz-table-outer has align-self:inherit.  Presumably because
the nsTableFrame is the primary frame and the outer frame's style inherits
from that IIRC.  So we should do the same for 'justify-self'.

Comment 1

[Mats Palmgren (inactive)]

Attachment: fix

I removed the comment on 'align-self' because these properties are
universal now.  I feel like removing the comment on 'order' too
because it's not just for tables that are flex/grid items really.
It's also needed for inheritance to work through <table> *regardless*
if the table is a flex/grid item or not.  Should we remove that
comment too?

Comment 2

[Daniel Holbert [:dholbert]]

(In reply to Mats Palmgren (:mats) from comment #1)
> I feel like removing the comment on 'order' too
> because it's not just for tables that are flex/grid items really.
> It's also needed for inheritance to work through <table> *regardless*
> if the table is a flex/grid item or not.

I don't think that's true, RE inheritance. (If it were true, then this line of thinking would be true of *all* css properties, and we'd arguably need "all: inherit" here.)

IIRC :-moz-table-outer is special because it's a wrapper for the table (the parent in the frame tree), but it's a *child* in the style-context tree.  So any properties on the table that its "real" parent cares about will need to be "inherited" up to the moz-table-outer wrapper, or else the "real" parent won't see those properties.

Comment 3

[Daniel Holbert [:dholbert]]

But anyway, getting back to inheritance in general -- these "inherit" statements don't matter for normal inheritance to the table's normal children, because ::-moz-table-outer isn't part of their style-system ancestry at all. (It's really a sibling, w/ no children, from a style-context perspective.)

Comment 4

[Daniel Holbert [:dholbert]]

Comment on attachment 8687535 [details] [diff] [review]
fix

Review of attachment 8687535 [details] [diff] [review]:
-----------------------------------------------------------------

Having said all that: I'm OK with these comments all going away (including the 'order' one) because it's clear that these are all properties that may matter to the "real" parent (for layout purposes), so it makes sense that they have to be included here.

So: r=me, with the 'order' comment removed as well if you like.

Comment 5

[Mats Palmgren (inactive)]

> I don't think that's true, RE inheritance.

Right, I was confused about the style context hierarchy in comment 0, sorry.

Comment 6

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/028619a0e9ae

Comment 7

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

Backed out for test_value_cloning.html bustage
https://hg.mozilla.org/integration/mozilla-inbound/rev/734a55000c77

Comment 8

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/2abf9c0c5c23

Comment 9

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/2abf9c0c5c23

Comment 10

[Hallvord R. M. Steen [:hallvors]]

Did this fix possibly cause bug 1231004 ? It regressed either because of this or one of the other changes going into the same nightly.

Comment 11

[Daniel Holbert [:dholbert]]

I don't think so. (I agree with Markus's guess in bug 1231004 comment 6.)