On Sat, 27 Dec 2014 10:40:30 +0100, Pablo Rodriguez wrote:

On 12/26/2014 11:32 PM, j. van den hoff wrote:

...

[...] in this example (and with this font) the problem is only occuring when toggling between pages 6/7 or 7/8 (at sufficient zoom it's easy to see), i.e.the header is shifted downward on page 7. in other docs with different fonts it happens across multiple pages. my guess would be it has to do with the digit heights in the different fonts and the fact that the page numbers are much larger than the heading, but that's only a guess.

can someone confirm the problem?

Hi Jörg,

this sample shows the issue without zooming:

thanks for confirming this. @devs: can this be considered "misbehaviour" (not to call it a bug ;-)) or is it somehow unavoidable?

\setuppapersize[A8] \setupheader [style=\itx\ss] \setupheadertexts [section] \setuppagenumbering[location={header,margin}, style=\bfd] \starttext \dorecurse{80}{\section{Introduction} bla\page[yes]} \stoptext

When the glyph 7 is involved in page number, baseline is a bit lower. As you wrote, I guess this is related with the glyph metrics.

yes. this seems to be supported by the fact that it happens differently (usually more frequently) with other fonts. joerg

Just in case it helps,

Pablo

-- Using Opera's revolutionary email client: http://www.opera.com/mail/

On Sat, 27 Dec 2014 13:36:49 +0100, Hans Hagen wrote:

On 12/27/2014 12:16 PM, j. van den hoff wrote:

...

On Sat, 27 Dec 2014 10:40:30 +0100, Pablo Rodriguez wrote:

...

On 12/26/2014 11:32 PM, j. van den hoff wrote:

...

[...] in this example (and with this font) the problem is only occuring when toggling between pages 6/7 or 7/8 (at sufficient zoom it's easy to see), i.e.the header is shifted downward on page 7. in other docs with different fonts it happens across multiple pages. my guess would be it has to do with the digit heights in the different fonts and the fact that the page numbers are much larger than the heading, but that's only a guess.

can someone confirm the problem?

Hi Jörg,

this sample shows the issue without zooming:

thanks for confirming this. @devs: can this be considered "misbehaviour" (not to call it a bug ;-)) or is it somehow unavoidable?

it's not a bug at all .. you just get what you ask for

good to know.

\showframe \showglyphs

your header is too small so what you want only works if the heights of all characters are the same which is not the case here

well this seems not to be the whole explanation AFAICS. it might be correct for pablo's example, but for my original example, which is equivalent to: \showframe\showglyphs \setupheader [style=\smallbodyfont\ss\it] \setupheadertexts [section] \setuppagenumbering[location={header,margin}, style=\bfc] \starttext \dorecurse{10}{\section{Introduction} bla \page[yes]} \stoptext I can't see that the header space is to small, no? or what am I missing? using `\showglyphs' (which I didn't know of: thanks), it still looks to me as if the problem is caused by the fact, that all positions are measured from the top and the header line's baseline is aligned with the baseline of the large page numbers glyph boxes (although only approximately so, as far as I can tell on my monitor...) where -- for the default font -- the glyphbox of the `7' has a larger vertical extension than the others (so that it extends a bit further down as measured from the top) which causes the baseline of the header to shift downwards accordingly. joerg

solutions:

- use a proper height for the header - package the header text in a box and set the height of that box

...

...

\setuppapersize[A8] \setupheader [style=\itx\ss] \setupheadertexts [section] \setuppagenumbering[location={header,margin}, style=\bfd] \starttext \dorecurse{80}{\section{Introduction} bla\page[yes]} \stoptext

When the glyph 7 is involved in page number, baseline is a bit lower. As you wrote, I guess this is related with the glyph metrics.

yes. this seems to be supported by the fact that it happens differently (usually more frequently) with other fonts.

joerg

...

Just in case it helps,

Pablo

-- Using Opera's revolutionary email client: http://www.opera.com/mail/

coming back to this again, if one uses old-style numerals for the page numbering the "problem" is aggravated: \showframe\showglyphs \definefontfeature [default] [default] [onum=yes] \definefontfamily [mainface] [serif] [garamondno8] \setupbodyfont[mainface,12pt] \setupheader [style=\smallbodyfont\ss\it] \setupheadertexts [section] \setuppagenumbering[location={header,margin}, style=\bfc] \starttext \dorecurse{97}{\section{Introduction} bla \page[yes]} \stoptext now, it also is obious that the "baseline jumps" occur in the page numbers themselves (caused mainly by the large extension of the `6') so going, e.g., from page 25 to 26 the `2' jumps downward quite a bit. I understand that this is unavoidable if the positions are determined as they seemingly are (measuring from the top and separately for each page). what I at least would think to be nicer would be a situation where the baseline is determined for the _whole_ document as the minimum (lowest position on page) of all page-wise baselines (that would be determined by the digit `6' in the present example) and then using _that_ baseline for all headers and page numbers which would eliminate the jittering of header lines and page numbers. I am aware that this will not always be the desired behaviour and that there might be documents where header layout might not be identical document wide, so it might open a can of worms, but for most book-type documents I would think this to be reasonable and desirable. my questions: -- would that (a common baseline for the text in all the page headers) be better or worse than the present behaviour from a typographical point of view? -- if desirable, could/should it be done? -- could such behaviour be achieved presently with some clever trick? joerg On Sat, 27 Dec 2014 13:36:49 +0100, Hans Hagen wrote:

solutions: - use a proper height for the header - package the header text in a box and set the height of that box

-- Using Opera's revolutionary email client: http://www.opera.com/mail/

On Sun, 28 Dec 2014 00:45:32 +0100, Hans Hagen wrote:

-- could such behaviour be achieved presently with some clever trick?

\showstruts \setuppagenumbering[style=\bfd\setcharstrut{0123456789}\strut,strut=no] - no default strut - calculate one based on to-be-used-characters (tight case) - apply that one

compare it with \setuppagenumbering[style=\bfd\setstrut\strut]

thanks a lot. that indeed fixes it.

And then wikify it

can do. although things are not yet 100% clear to me (not much to be found in the documentation). so is this right: -- I guess `\setstrut\strut' does set the strut to the maximum height of all the glyphs in the used font? the explanation in 4.6 of the ref. manual is not really telling me how the box height in the example comes about (i.e. what the mentioned `maximal height' actually is). -- `\setcharstrut{0123456789}': the command seems to be completely undocumented (I could not find it in wiki and ref. manual), but it seems clear what it does. except what do you mean by "tight case" here? -- why is it necessary to state `strut=no' to switch of the default on top of setting the strut explicitly? joerg

Hans

-- Using Opera's revolutionary email client: http://www.opera.com/mail/

On Sun, 28 Dec 2014 18:55:15 +0100, Hans Hagen wrote:

set the strut to match the given set of characters

so, I've tried to add a description to the wiki: http://wiki.contextgarden.net/Command/setcharstrut I'd appreciate if you would double-check and correct any errors/misconceptions. joerg -- Using Opera's revolutionary email client: http://www.opera.com/mail/