Series changes (a version of #1396) #1583

FrankMittelbach · 2024-12-13T22:34:54Z

Internal housekeeping

Status of pull request

Ready to merge

Checklist of required changes before merge will be approved

[n/a] Test file(s) added
Version and date string updated in changed source files
Relevant \changes entries in source included
Relevant changes.txt updated
Rollback provided (if necessary)?
ltnewsX.tex (and/or latexchanges.tex) updated

This is a slightly adjusted version of #1396 because that pull request would be difficult to merge after all several changes to ltfssaxes.dtx. It contains the pull request #1582 as the two are somewhat interwined.

# Conflicts: # base/ltfssaxes.dtx

…hanges # Conflicts: # base/ltfssaxes.dtx

base/ltfssaxes.dtx

cfr42 · 2024-12-14T05:17:23Z

Is it possible to predict the chosen rules? In most cases, the reasoning is obvious, but it is not always clear why weight is given priority over width in some cases but not in others (or the other way around).

I also wonder how easy it would be to enable users to \Show whether a rule applies and, if so, what it is. (Most users wouldn't use it, but the same is true for most aids to diagnosis.)

[I am somewhat tempted to remove code which subverts the series change rules from nfssext-cfr if the format is new enough, but I don't understand the reasons for the rule choices at the moment.]

FrankMittelbach · 2024-12-14T13:10:31Z

Is it possible to predict the chosen rules? In most cases, the reasoning is obvious, but it is not always clear why weight is given priority over width in some cases but not in others (or the other way around).

it is pretty difficult to come up with consistent defaults when a falback is required. Could be that during all this editing some leftovers are still there.

My basic guiding principle for the few cases with the forth argument non-empty was

bold is something that is typically used in headings so if some form of bold is requested you don't want to loose it altogether, so for semibold, extrabold and ultrabold if that doesn't exist we try just bold as a nearby neighbor
bx is swapped with b if not existing to account for \bfdefault being bx while only CM and LM and a few others have bx rather than b.
of course, one could argue, that something like {c}{bx} -> {bx} or {b} could be better {c}{bx} -> {bc}`, i.e., always interpreting "bx" as a bad 2e name for "b" but then you loose the ability to really specify "bx" meaning go to "bold extended"
using bx as a fallback for b makes little sense for anything other than "m" because if we are in, say, "c" then this font family is most certainly having a "b" if it has any bold at all.

I guess that's all behind the current choices.

I also wonder how easy it would be to enable users to \Show whether a rule applies and, if so, what it is. (Most users wouldn't use it, but the same is true for most aids to diagnosis.)

maybe. but it is already logged if a substitution happens ie if the 4th argument is used rather than the 3rd if I remember correctly, but I can have a look to see if some "debugging" cna be easily implemented.

[I am somewhat tempted to remove code which subverts the series change rules from nfssext-cfr if the format is new enough, but I don't understand the reasons for the rule choices at the moment.]

user227621

Firstly, a few minor comments (another review will follow)

base/ltfssaxes.dtx

user227621

Two more minor comments (another review will follow)

base/changes.txt

base/ltfssaxes.dtx

Co-authored-by: user227621 <[email protected]>

user227621

To fully implement the "stay with some sort of bold if a request for sb, eb or ub cannot be fulfilled" logic, the following rules are missing.

(another review will follow)

base/ltfssaxes.dtx

Co-authored-by: user227621 <[email protected]>

base/ltfssaxes.dtx

FrankMittelbach · 2024-12-14T20:40:01Z

I also wonder how easy it would be to enable users to \Show whether a rule applies and, if so, what it is. (Most users wouldn't use it, but the same is true for most aids to diagnosis.)

something like this?

\documentclass{article}

%\input{test2e}

\usepackage[debugshow]{tracefnt}

\makeatletter


\DeclareRobustCommand\selectfont
        {%
    \ifx\f@linespread\baselinestretch \else
      \set@fontsize\baselinestretch\f@size\f@baselineskip \fi
    \ifx\delayed@f@adjustment\@empty
    \else
      \typeout{-----> Try changing font series in \noexpand\selectfont}%
      \let\f@shape@saved\f@shape
      \let\f@series@saved\f@series
      \delayed@f@adjustment
      \maybe@load@fontshape
      \ifcsname \f@encoding/\f@family/\f@series/\f@shape \endcsname
      \else
        \typeout{>>>> Font face \f@encoding/\f@family/\f@series/\f@shape\space
	   does not exit ^^J%
	   \@spaces\@spaces Retry with
	   \noexpand\f@series = '\f@series@saved' and
	   \noexpand\f@shape = '\f@shape@saved'
	   }%
        \let\f@shape\f@shape@saved
        \let\f@series\f@series@saved
        \let\delayed@merge@font@shape\merge@font@shape
        \let\delayed@merge@font@series\merge@font@series
        \delayed@f@adjustment
        \let\delayed@merge@font@shape\merge@font@shape@without@substitution
        \let\delayed@merge@font@series\merge@font@series@without@substitution
      \fi
      \let\delayed@f@adjustment\@empty
    \fi
    \@forced@seriesfalse
    \xdef\font@name{%
      \csname\curr@fontshape/\f@size\endcsname}%
    \pickup@font
    \font@name
    \UseHook{selectfont}%
    \size@update
    \enc@update
    \typeout{-----> Series after \noexpand\selectfont is '\f@series'}%
    }


\DeclareRobustCommand\fontseries[1]{%
  \expandafter \let \expandafter \reserved@a
        	      \csname series@\f@series @#1\endcsname
  \typeout{-----> \noexpand\fontseries request with '#1'^^J%
   	      \@spaces\@spaces
               Current font series: '\f@series'^^J%
  	      \@spaces\@spaces
	      \ifx  \reserved@a \relax
 	         No rule for '\f@series' + '#1'
		 --> try '#1' unconditionally
	      \else
  	         Rule: '\f@series' + '#1'
		 --> 
	         '\expandafter \@firstoftwo \reserved@a'
	         \if!\expandafter \@secondoftwo \reserved@a!\else
	            \space (\expandafter \@secondoftwo \reserved@a)
		 \fi
	       \fi
	}%
    \@forced@seriesfalse
    \expandafter\def\expandafter\delayed@f@adjustment\expandafter
        {\delayed@f@adjustment\delayed@merge@font@series{#1}}}

\DeclareRobustCommand\fontseriesforce[1]{%
  \expandafter \let \expandafter \reserved@a
        	      \csname series@\f@series @#1\endcsname
  \typeout{-----> \noexpand\fontseriesforce request with '#1'^^J%
  	      \@spaces\@spaces
               Current font series: '\f@series'^^J%
  	      \@spaces\@spaces
	      \ifx  \reserved@a \relax
 	         No rule for '\f@series' + '#1'
		 --> try '#1' unconditionally
	      \else
  	         Rule: '\f@series' + '#1'
		 --> 
	         '\expandafter \@firstoftwo \reserved@a'
	         \if!\expandafter \@secondoftwo \reserved@a!\else
	            \space (\expandafter \@secondoftwo \reserved@a)
		 \fi
	       \fi
	}%
    \@forced@seriestrue
    \expandafter\def\expandafter\delayed@f@adjustment\expandafter
      {\delayed@f@adjustment\edef\f@series{#1}}}


\let\delayed@merge@font@series\merge@font@series@without@substitution


\def\merge@font@series#1{%
  \typeout{---> Attempt to switch with fallback for \f@series\space + #1}%
  \expandafter\expandafter\expandafter
  \merge@font@series@
    \csname series@\f@series @#1\endcsname
    {#1}%
    \@nil
}
\def\merge@font@series@#1#2#3\@nil{%
  \def\reserved@a{#3}%
  \ifx\reserved@a\@empty
    \typeout{-->> Attempt to switch to '#2' (ignore current series - no rule)}%
    \set@target@series{#2}%
  \else
    \typeout{-->> Attempt to switch to '#1'}%
    \edef\reserved@a{\f@encoding /\f@family /#1/\f@shape}%
     \ifcsname \reserved@a \endcsname
       \set@target@series{#1}%
    \else
       \typeout{-->> ... failed; attempt to switch to '#2' (fallback)}%
       \ifcsname \f@encoding /\f@family /#2/\f@shape \endcsname
         \set@target@series{#2}%
         \@font@shape@subst@warning
       \else
         \typeout{-->> ... failed; attempt to switch to '#3' (ignore current series)}%
         \set@target@series{#3}%
         \@font@shape@subst@warning
       \fi
    \fi
  \fi
  \typeout{->>> \noexpand\f@series now '\f@series'}%
}


\def\merge@font@series@without@substitution#1{%
  \typeout{---> Attempt to switch without using fallback for \f@series\space + #1}%
  \expandafter\expandafter\expandafter
  \merge@font@series@without@substitution@
    \csname series@\f@series @#1\endcsname
    {#1}%
    \@nil
}


\def\merge@font@series@without@substitution@#1#2#3\@nil{%
  \def\reserved@a{#3}%
  \ifx\reserved@a\@empty
    \typeout{-->>  Switch unconditionally to '#2' (ignore current series)}%
    \set@target@series{#2}%
  \else
    \typeout{-->> Switch unconditionally to '#1'}%
    \set@target@series{#1}%
  \fi
  \typeout{->>> \noexpand\f@series now '\f@series'}%
}


\makeatother

\begin{document}

%\START

\bfseries A
\fontseries{c}\selectfont B
\fontseries{zzz}\selectfont C
\bfseries D\tracingnone

\newpage
%\OMIT
\end{document}

After cleanup and some integration with tracefnt perhaps.

user227621

Some rewordings

base/ltfssaxes.dtx

user227621 · 2024-12-14T20:48:53Z

Three further comments:

The copyright notice in ltfssaxes.dtx is outdated: Copyright (C) 2019-2020 Frank Mittelbach
An incomplete sentence at the beginning of ltfssaxes.dtx: "This file contains the implementation for handling extra axes splitting the series and the values into sub-categories. selection commands."
The rollback data needs to be regenerated because of added/deleted/changed rules.

user227621

Improve alignment

base/ltfssaxes.dtx

cfr42 · 2024-12-15T16:30:21Z

something like this?

That's more elaborate than I had in mind, but, yes, that would be useful, I think.

On line 24: exit -> exist.

Thanks.

car222222

Not checked: approx. one million? Change Rules . . .
so I cannot guarantee that all the alignments are spot on!

A few wording changes are either essential, or recommended.

base/ltfssaxes.dtx

cfr42 · 2024-12-15T17:38:31Z

I think using a more probable series in the fourth argument makes more sense than repeating the series from the first, unless there is some guarantee that the first actually exists.

Probably I'm missing something here, but there is certainly no general guarantee that the current font name reflects the actual font series (or shape etc.). After \fontseries{zzz}\selectfont, the current font series is zzz. So if you add a rule like

\DeclareFontSeriesChangeRule {zzz}{eb}{eb}{zzz}

because zzz is a better approximation for eb than b, that doesn't mean you will actually get a better approximation for eb in the typeset result.

Co-authored-by: user227621 <[email protected]>

FrankMittelbach · 2024-12-15T18:19:57Z

I think using a more probable series in the fourth argument makes more sense than repeating the series from the first, unless there is some guarantee that the first actually exists.

Probably I'm missing something here, but there is certainly no general guarantee that the current font name reflects the actual font series (or shape etc.). After \fontseries{zzz}\selectfont, the current font series is zzz. So if you add a rule like
\DeclareFontSeriesChangeRule {zzz}{eb}{eb}{zzz}
because zzz is a better approximation for eb than b, that doesn't mean you will actually get a better approximation for eb in the typeset result.

Given that one doesn't know what 'zzzz' it is hard to say if that is a good rule or not, but if the user wanted 'zzz' and now wants to switch to 'eb' and 'eb' doesn't exist then your rule says stay with the series you are in, so it is not an unreasonable rule even for that case. Of course, if you want to ensure that there is always some "boldness" as the result and using 'b' is a better choice (which is what I wrote earlier).

But the statement wasn't that any arbitray zzzz is good approximation for eb, the argument was that ebuc is a better approroximation to ubuc than b or buc.

cfr42 · 2024-12-15T20:48:28Z

But the statement wasn't that any arbitray zzzz is good approximation for eb, the argument was that ebuc is a better approroximation to ubuc than b or buc.

Yes, but the current series being ebuc doesn't guarantee this, does it? Because you don't know whether it is actually ebuc or not.

cfr42 · 2024-12-15T21:05:40Z

But, thinking about this again, maybe the rules do now guarantee the current font is the best available substitute, even if it is not what it says on the tin?

FrankMittelbach · 2024-12-15T21:52:35Z

But the statement wasn't that any arbitray zzzz is good approximation for eb, the argument was that ebuc is a better approroximation to ubuc than b or buc.

Yes, but the current series being ebuc doesn't guarantee this, does it? Because you don't know whether it is actually ebuc or not.

No it does not, i.e., if somebody did \fontseries{ebuc} in one go an this didn't exist, but as I tried to explain in my other reply, if ubuc and ebuc doesn't exist trying the latter one more time doesn't really hurt. Any such fallback is only helping so much and if you are making an important document then you better make sure that all that is requested as font faces is really to your liking and if not do some adjustments.

cfr42 · 2024-12-16T05:05:11Z

No it does not, i.e., if somebody did \fontseries{ebuc} in one go an this didn't exist, but as I tried to explain in my other reply, if ubuc and ebuc doesn't exist trying the latter one more time doesn't really hurt.

Sorry. I do not really understand how to navigate GitHub replies as the link always just dumps me at the top of the page.

Any such fallback is only helping so much and if you are making an important document then you better make sure that all that is requested as font faces is really to your liking and if not do some adjustments.

Well, yes. I don't really have a view on what the rule should be. I just wasn't convinced by the reason for changing your initial proposal. Indeed, there is a limit to what you can do. I keep track of fake fonts and it still isn't always clear what to do.

FrankMittelbach · 2024-12-16T08:41:19Z

No it does not, i.e., if somebody did \fontseries{ebuc} in one go an this didn't exist, but as I tried to explain in my other reply, if ubuc and ebuc doesn't exist trying the latter one more time doesn't really hurt.

Sorry. I do not really understand how to navigate GitHub replies as the link always just dumps me at the top of the page.

you are not the only one. I find that difficult too and largely rely on the additional mails I get as well

Any such fallback is only helping so much and if you are making an important document then you better make sure that all that is requested as font faces is really to your liking and if not do some adjustments.

Well, yes. I don't really have a view on what the rule should be. I just wasn't convinced by the reason for changing your initial proposal. Indeed, there is a limit to what you can do. I keep track of fake fonts and it still isn't always clear what to do.

I'm not 100% convinced either, but my current feeling is that is not a bad idea the way it is now. Maybe it has to be revisited at some point if mre real life usage shows what works best.

@user227621 any more comments from you (you still request changes)?

base/ltfssaxes.dtx

user227621 · 2024-12-16T15:20:29Z

And the (hopefully) final comment: In line 244, replace

%    Only a few entries have ``alternative'' values and perhaps most of
%    them should get dropped. Or maybe not \ldots{} needs some thought
%    perhaps.

with

%    Only a few entries have ``alternative'' values.

(I think we've thought about it enough now.)

Co-authored-by: user227621 <[email protected]>

…ything that hasn't changed between 2024 and 2020)

FrankMittelbach · 2024-12-16T17:00:53Z

unless I overlooked something I think this PR is done from my side

user227621

You could remove the unnecessary <----- comments in the rollback section (they just produce overfull boxes). Otherwise, I think this PR is ready.

base/ltfssaxes.dtx

Co-authored-by: user227621 <[email protected]>

FrankMittelbach · 2024-12-17T14:11:46Z

thanks @car222222, @cfr42 and @car222222 for helping to get this done

FrankMittelbach · 2024-12-17T14:30:04Z

thanks @car222222, @cfr42 and @car222222 for helping to get this done

correction: thanks @user227621 ... !

user227621 · 2024-12-17T19:19:56Z

@FrankMittelbach You're welcome.

cfr42 · 2024-12-17T21:39:01Z

You don't need to thank me for sticking my nose in where it wasn't wanted. ;)

FrankMittelbach · 2024-12-17T21:59:50Z

You don't need to thank me for sticking my nose in where it wasn't wanted. ;)

any constructive feedback is wanted (by me at least). Doesn't necessarily me I follow a suggestion (though often enough I do), but just tihnk about it or clarify, if something is not clear, is a Good Thing (TM). And it resulted, for example, in some draft code for debugging which eventually wil be useful, once integrated

FrankMittelbach added 10 commits May 18, 2022 10:13

typo

d3e2650

Merge branch 'develop' into WIP/ssc-mappings

24743e3

# Conflicts: # base/ltfssaxes.dtx

rollback tests and documentation for #1581

0e87160

correct docs after review

6057f09

ltfssaxes from #1396

e60a658

Merge commit '6057f09f891b6d01017cd64ad93a973a53dfd2d1' into series-c…

186d460

…hanges # Conflicts: # base/ltfssaxes.dtx

ltnews41 and changes

32ec910

align series change rule arguments

8856bcd

more rule cleanup and start of rollback support

f821ef7

Full support

ef0cf69

FrankMittelbach requested review from josephwright, car222222 and davidcarlisle December 13, 2024 22:38

FrankMittelbach mentioned this pull request Dec 13, 2024

Add numerous \DeclareFontSeriesChangeRule entries #1396

Closed

6 tasks

cfr42 reviewed Dec 14, 2024

View reviewed changes

base/ltfssaxes.dtx Outdated Show resolved Hide resolved

wording

4abab39

user227621 suggested changes Dec 14, 2024

View reviewed changes

base/changes.txt Outdated Show resolved Hide resolved

base/ltfssaxes.dtx Outdated Show resolved Hide resolved

FrankMittelbach and others added 2 commits December 14, 2024 19:02

Apply suggestions from code review

b9633cc

Co-authored-by: user227621 <[email protected]>

more suggestions from review

e858194

FrankMittelbach requested a review from user227621 December 14, 2024 18:16

otf files do not end in .ttf :-)

987e70b

user227621 suggested changes Dec 14, 2024

View reviewed changes

Apply suggestions from code review

71e5638

Co-authored-by: user227621 <[email protected]>

FrankMittelbach commented Dec 14, 2024

View reviewed changes

base/ltfssaxes.dtx Outdated Show resolved Hide resolved

base/ltfssaxes.dtx Outdated Show resolved Hide resolved

base/ltfssaxes.dtx Show resolved Hide resolved

user227621 suggested changes Dec 14, 2024

View reviewed changes

base/ltfssaxes.dtx Outdated Show resolved Hide resolved

base/ltfssaxes.dtx Outdated Show resolved Hide resolved

base/ltfssaxes.dtx Outdated Show resolved Hide resolved

base/ltfssaxes.dtx Outdated Show resolved Hide resolved

user227621 suggested changes Dec 15, 2024

View reviewed changes

base/ltfssaxes.dtx Outdated Show resolved Hide resolved

car222222 approved these changes Dec 15, 2024

View reviewed changes

Update base/ltfssaxes.dtx

cac3aef

Co-authored-by: user227621 <[email protected]>

FrankMittelbach added 4 commits December 15, 2024 19:59

corrections by Chris

ac2edfd

re-add a rule

6555f31

grumble

408d9e8

and, of course, my typical typo

84c5e3e

user227621 suggested changes Dec 16, 2024

View reviewed changes

FrankMittelbach and others added 3 commits December 16, 2024 16:38

Apply suggestions from code review

cf887ec

Co-authored-by: user227621 <[email protected]>

some further minor changes

2e90da0

regen and cleanup of the rollback data (we don't need to roll back an…

3e1821b

…ything that hasn't changed between 2024 and 2020)

user227621 approved these changes Dec 17, 2024

View reviewed changes

Apply suggestions from code review

199f03f

Co-authored-by: user227621 <[email protected]>

FrankMittelbach merged commit 176699b into develop Dec 17, 2024
86 checks passed

FrankMittelbach deleted the series-changes branch December 17, 2024 14:12

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Series changes (a version of #1396) #1583

Series changes (a version of #1396) #1583

FrankMittelbach commented Dec 13, 2024 •

edited

Loading

cfr42 commented Dec 14, 2024

FrankMittelbach commented Dec 14, 2024

user227621 left a comment

user227621 left a comment •

edited

Loading

user227621 left a comment

FrankMittelbach commented Dec 14, 2024

user227621 left a comment

user227621 commented Dec 14, 2024

user227621 left a comment

cfr42 commented Dec 15, 2024

car222222 left a comment

cfr42 commented Dec 15, 2024

FrankMittelbach commented Dec 15, 2024

cfr42 commented Dec 15, 2024

cfr42 commented Dec 15, 2024

FrankMittelbach commented Dec 15, 2024

cfr42 commented Dec 16, 2024

FrankMittelbach commented Dec 16, 2024

user227621 commented Dec 16, 2024

FrankMittelbach commented Dec 16, 2024

user227621 left a comment

FrankMittelbach commented Dec 17, 2024

FrankMittelbach commented Dec 17, 2024

user227621 commented Dec 17, 2024

cfr42 commented Dec 17, 2024

FrankMittelbach commented Dec 17, 2024

Series changes (a version of #1396) #1583

Series changes (a version of #1396) #1583

Conversation

FrankMittelbach commented Dec 13, 2024 • edited Loading

Internal housekeeping

Status of pull request

Checklist of required changes before merge will be approved

cfr42 commented Dec 14, 2024

FrankMittelbach commented Dec 14, 2024

user227621 left a comment

Choose a reason for hiding this comment

user227621 left a comment • edited Loading

Choose a reason for hiding this comment

user227621 left a comment

Choose a reason for hiding this comment

FrankMittelbach commented Dec 14, 2024

user227621 left a comment

Choose a reason for hiding this comment

user227621 commented Dec 14, 2024

user227621 left a comment

Choose a reason for hiding this comment

cfr42 commented Dec 15, 2024

car222222 left a comment

Choose a reason for hiding this comment

cfr42 commented Dec 15, 2024

FrankMittelbach commented Dec 15, 2024

cfr42 commented Dec 15, 2024

cfr42 commented Dec 15, 2024

FrankMittelbach commented Dec 15, 2024

cfr42 commented Dec 16, 2024

FrankMittelbach commented Dec 16, 2024

user227621 commented Dec 16, 2024

FrankMittelbach commented Dec 16, 2024

user227621 left a comment

Choose a reason for hiding this comment

FrankMittelbach commented Dec 17, 2024

FrankMittelbach commented Dec 17, 2024

user227621 commented Dec 17, 2024

cfr42 commented Dec 17, 2024

FrankMittelbach commented Dec 17, 2024

FrankMittelbach commented Dec 13, 2024 •

edited

Loading

user227621 left a comment •

edited

Loading