1124344 - css letter-spacing does not affect mono-space tabs

Status: RESOLVED INVALID

(Core :: Layout, defect)

Description

[rui.damas]

Attachment: letter-spacing.png

User Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Ubuntu Chromium/39.0.2171.65 Chrome/39.0.2171.65 Safari/537.36

Steps to reproduce:

there is an image in attach.


Actual results:

mono-space tabs have wrong size with letter-spacing.


Expected results:

mono-space tabs should have size complying with letter-spacing.

Comment 1

[:Gijs (he/him)]

Can you provide an HTML/CSS testcase to illustrate what you mean (e.g. the bits you used to generate the screenshot), as well as provide more detail as to what exactly you think is going wrong here?

In particular, you say it has the "wrong size" - are you saying it's getting spacing it shouldn't have, or misses spacing it should have, or has spacing but should have different spacing? It's not clear to me which you're saying is happening, by looking at your screenshot.

I am not a spec reader, but a quick look at CSS3 suggests that letter-spacing applies between typographic character units, and that for the purposes of the spec, that includes things that unicode thinks are either Letters or Numbers, which a spacing character like space or tab isn't.

Comment 2

[rui.damas]

what you see in the screenshot are the first 2 lines of a text editor that i'm coding... they are generated by javascript​... each letter is inside its own span element, all inside a div (__display)... i apply the css to all childs... i bumped into this when, at first, using padding for letter spacing, leads to this problem... seams logic to me by using letter-spacing should solve this specially if is a mono-spaced font.

.__editor .__text {
  letter-spacing: 1px;
}

those are 2 screenshots... i glued them with gimp (chromium does the same: screenshot in attach).

what i'am saying is that the tab size does to take in account the letter-spacing (which may apply the letter-spacing at tab boundaries, but the "internal space" seams to be wrong, as if each space inside the tab is wrong)... you can see the difference in the screenshot i sent first.

"letter-spacing applies between typographic character units" ... while this may make some sense in "typographic" theory... in practice makes no sense for mono-spaced text loose its mono-spaciness... right?

Comment 3

[rui.damas]

* correction of last message: where is __display should be _text

Comment 4

[rui.damas]

*... and the chromium screenshot is here: https://code.google.com/p/chromium/issues/detail?id=450720

Comment 5

[rui.damas]

<!-- test case -->
<pre style="letter-spacing: 5px;">
123456789
&#9;9
</pre>

<pre>
123456789
&#9;9
</pre>

Comment 6

[rui.damas]

please don't tell me that i should implement tab algorithms (with individual spaces) for my editor: that would be ridiculous... letter-spacing should comply to mono-space rules (tabs are actually made of typographic units (spaces)... right?... coders... common-sense... :)

Comment 7

[:Gijs (he/him)]

Unless I'm reading the spec wrong (quite possible!) I expect this would need a spec change. I don't have the expertise to determine whether your suggestion is as sensible as you... suggest. :-)
Boris?

Comment 8

[Boris Zbarsky [:bzbarsky]]

Yep, as far as I can tell we're doing exactly what the spec says.  The right place to talk about spec changes is [e-mail].

Comment 9

[rui.damas]

can you leave here the link (or copy/past) for the spec (or part) you read... where explains how to calculate tabs (i presume).

i would like to read give it a look before sending an email to that [e-mail] referring this page...

weeee: i found a bug in the spec! :D

in the meanwhile you should fix this... since it is wrong.
mono-space is mono-space, if does not apply to typo-rules make an exception for it...
aka: can you suggest any other workaround?..

ty, r.

Comment 10

[Boris Zbarsky [:bzbarsky]]

http://www.w3.org/TR/CSS21/text.html#white-space-model says:

  All tabs (U+0009) are rendered as a horizontal shift that lines up the start edge of
  the next glyph with the next tab stop. Tab stops occur at points that are multiples of
  8 times the width of a space (U+0020) rendered in the block's font from the block's
  starting content edge. 

Note that this says nothing about converting tabs to spaces or letter-spacing affecting tab stops.

Also note that this depends on the _block'_s font, while letter-spacing can be applied to inlines inside the block and so forth.

> in the meanwhile you should fix this... since it is wrong.

It's not clear to me that it's wrong is the point, so we'd rather not start randomly violating the spec unless there is a clear indication it will change.

Comment 11

[rui.damas]

... i cannot read in that text nothing that prevents letter-spacing from applying to the "multiples of
  8" (that actually depends on the chars before the tab) " times the width of a space (U+0020) rendered in the block's font", and a space next to another or mores spaces (a tab) must comply to it's letter-space context... presume that it must no?... really?... after reading that i still expect that letter-spacing is taken in account in tabs, and i see no objection... by reading, it would be easier for me to think that all tabs render 8 chars than to think that letter-spacing does not apply... does not talk about tab size (in characters) variation, yet you know how tabs are rendered in different sizes depending on the position... come on, run the test case and admit: that is ridiculous... because it simply is :P

"Also note that this depends on the _block'_s font, while letter-spacing can be applied to inlines inside the block and so forth." ... sorry, did not understand what you meant.

Note that the units of tab character size is in characters.
  https://developer.mozilla.org/en-US/docs/Web/CSS/tab-size
... if the tab is in a letter-spaced context, part of the text, how can we say that mono-space complies to mono-space?

to me is quite clear that it is, somehow, wrong... if not for all fonts, than for mono-space it is... if this is wrong, then there should exist a solution to allow this (though i don't see any other)... is only natural that i want to use 1px of letter-spacing and keep the tab sizes right, right coders: that is the usefulness of mono-space.

... so: i need this working... i'm sent bellow to [e-mail]

subject: spec complain: css letter-spacing in tabs not clear

​please help us on this bug, we are having spec interpretation issues:

  ​https://bugzilla.mozilla.org/show_bug.cgi?id=1124344

Comment 12

[Boris Zbarsky [:bzbarsky]]

> i cannot read in that text nothing that prevents letter-spacing from applying

The text is pretty clear: the tab stops are spaced at 8*(width of U+0020 in block's font).  There's no letter-spacing involved.

> and a space next to another or mores spaces

There are no actual spaces.  There are tab stops, at certain distances apart from each other.

> it would be easier for me to think that all tabs render 8 chars

I have no idea how you would get that from the spec language here.

> yet you know how tabs are rendered in different sizes depending on the position

Yes, it says "next tab stop".  Which is a different distance away depending on where you are, obviously, because the tab stop positions are fixed.

> sorry, did not understand what you meant

Consider the rendering of this case:

<pre>
<span style="letter-spacing: 1em">123456789</span>
<span style="letter-spacing: 1em">&#9;9</span>
<span>&#9;9</span>
</pre>

I think the only one having spec interpretation issues is you.  The spec is very clear, though it doesn't say what you want.  But I can't tell what you'd even want in the testcase with letter-spacing on the <span>s.

Comment 13

[rui.damas]

you really look at our mono-space test cases and feels it's right?... i do not care about the "stupid" spec... if it does not work right, should change.

give-me a workaround then!... i can do "wonders" css/html... the whole chabang, and i cannot put 1px between letters because the tabs will be screwed?... something is wrong here, if not your interpretation, then the spec. 

Boris... sorry: but will someone else with more common-sense of reality talk to me... ty.

Comment 14

[Boris Zbarsky [:bzbarsky]]

> you really look at our mono-space test cases and feels it's right?

It's following the spec.  That may or may not be the same thing, depending.

> if it does not work right, should change

Defining "right" in all circumstances here is not simple.  But again, the spec may want changing, depending on how important people judge this use case and how complicated things get when one tries to address it.  This is why I suggested you contact the standards group.

Until then, I'm going to stop wasting my time on this.  This bug is invalid unless the spec changes.

> give-me a workaround then!

Use JS to expand out tabs into the right number of spaces?  There is no pure-CSS workaround; CSS proper is not very good with working with tabs, because tabs don't make much sense in situations when multiple different fonts at multiple different sizes are involved.... Sad, but that's life.

> but will someone else with more common-sense of reality talk to me

You know, insulting people is not a good way to get them to help you.

Comment 15

[rui.damas]

"It's following the spec.  That may or may not be the same thing, depending."

you speak like a soldier... i'm just a common civilian... and kidding :P

"Until then, I'm going to stop wasting my time on this."

yes, please, ty.

... and yes: use js to fill with spaces what "your" tab rendering should already do right... any other suggestion?... love to hear it... how many others will "you" make do the same?... how many people have given up using letter-spacing because of one stupid tab? (...and did not complain... you just have no idea... just because you don't need this working right, you seam not to care)

"Sad, but that's life."

true: it is only very sad that i have to simulate what is right... "when in deed there is no need"... because "you" already almost got it right (is just missing the letter-spacing).

...
i'm not insulting you... there is something wrong, i do not care for who to blame, i want it fixed... you already showed that it does not bother you personally (what matters to you is complying to the spec under "your" interpretation.. and i understand your "state of being")... what matters to me is to be able to do nice and logic stuff... makes no logic see that all mono-space chars works fine with letter-spacing... except tab?... which is measured in spaces?... should i presume that?... (these are not to answer).

again: please look at the test case and see: that is wrong, makes no-sense.

Comment 16

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Given the model in:
  http://dev.w3.org/csswg/css-text/#white-space-phase-2
  http://dev.w3.org/csswg/css-text/#tab-size-property
I think it's reasonably clear that if letter-spacing did apply, it would be the letter-spacing of the block.

Comment 17

[rui.damas]

sure... "you" are all pseudo-"right"... in practical uselessness... unfortunately: no one with good taste and more than 2 neurons agrees with "your" mono-space corruption... but please insist: i'm just starting to have "fun".

your "reasonably clear" is rather disappointing.

... any volunteers to change the spec?

and... your "see also" was supposed to bring any light on the subject?... or just link back here?.. didn't get it... (just kidding: don't bother answering these)

Comment 18

[rui.damas]

... now seriously, real people need stuff that works as real people expected... specs are made to please us, not us to please the specs.

... [e-mail] takes so much time to answer...

Comment 19

[rui.damas]

... and my apologies if i have a "hard time" talking with people that believe in mono-space corruption... anyone believing so, is off my realm of understanding and off good-sense...

... spec guys, please answer my email.

Comment 20

[Boris Zbarsky [:bzbarsky]]

Please stop spamming this bug.

Your mail to the list is presumably in the moderation queue, since you didn't subscribe before sending.  It hasn't come through to the list yet.