bug#35885: 25.2; Few mistakes in Emacs Manual (+ proposals)

[Sebastian Urban]

Thanks for the fixes, but I don't think closing this bug was good
decision.  Even if we leave quotes behind (but we won't, right?),
Unicode code and name pairs bug will still be there.

> I believe you saw these in the Emacs 25 manual.

No, my reference is version updated for 26.2, downloaded from official
Emacs website with manuals.  For this e-mail I also looked into HTML
version.

> ... checkout the current emacs-26 branch...

Well after you said it, I did that - I downloaded basic.texi,
display.texi, search.texi, text.texi from:
http://git.savannah.gnu.org/cgit/emacs.git/tree/doc/emacs?h=emacs-26
and texinfo.tex from:
http://git.savannah.gnu.org/cgit/emacs.git/tree/doc/misc?h=emacs-26

Unfortunately I didn't make PDF out of them, but I looked into theirs
code/text.

> 1.  In PDF 4.1 Inserting Text:
> in paragraph starting with "A few common Unicode characters can be
> inserted (...)" (near word "respectively"), "left double quotation
> mark" (curved) is displayed as "backslash" and "right double quotation
> mark" (curved) is displayed as "straight double quotes".  In INFO it
> looks ok.

Well for this I have no answer...

In BASIC.TEXI (L119):
# curved quotes @t{’}, @t{“} and @t{”}, respectively.  Also, a working
While @t{...} works for
  single quotes - both curved (#x2018 & #x2019), probably including x2
  grave accent, including x2
  apostrophe, including x2
making all of them curved and in typewriter shape in PDF, it fails to
show LEFT (#x201c) and RIGHT (#x201d) DOUBLE QUOTATION MARK, and
displays instead BACKSLASH and QUOTATION MARK (#x22).  It also works
for QUOTATION MARK - @t{"}.  An ugly way to fix it would be @t{``} and
@t{''}, but I think it's not an option.

BTW This looks good in HTML version.

> ALSO (addition)
>
> In point 1 (previous message) in the same paragraph, line below, "(...)
> C-x 8 [ and inserts (...)" and curved opening quote follows, which
> is good, but it has shape of normal latex text, while it should has
> typewriter/verbatim shape (bolder one).

In BASIC.TEXI - LINE 121:
- and inserts `.  To see which characters have @kbd{C-x 8} ...
+ and inserts @t{‘}.  To see which characters have @kbd{C-x 8} ...
Just like in L115.  This bug is present in HTML version as well.

> 2.  In PDF 4.1 Inserting Text:
> in paragraph starting with "In some contexts, if you type a quotation
> (...)", display of quotes is messed up, quotes surrounding Nth
> occurrence of "like this" should be:
>    - 1st - "grave accent" and "apostrophe",
>    - 2nd - is ok,
>    - 3rd - 2x "grave accent" and 2x "apostrophe",
>    - 4th - "left double quotation mark" (curved) and "right double
>      quotation mark" (curved)
> Or just look into INFO, it's ok there.

As for 1st occurrence in BASIC.TEXI - L149:
# accent and apostrophe @t{`like this'}, it is converted to a form
It could be corrected with @kbd{`}@t{like this}@kbd{'}.  Looks good
in HTML.

As for 2nd - it also looks good in HTML.

As for 3rd occurrence in BASIC.TEXI - L151:
# commands.  Similarly, typing a quotation @t{``like this''} using
As above - @kbd{``}@t{like this}@kbd{''}...  Looks bad in HTML.

As for 4th occurrence, just like in first point - I don't know.
Looks good in HTML.

> 3.  In PDF 11.19 How Text Is Displayed:
> in last paragraph, first line, "left double quotation mark" (curved)
> is displayed as "backslash" and "right double quotation mark" (curved)
> is displayed as "straight double quotes".  In INFO it looks ok.

In DISPLAY.TEXI - L1560:
#  If the curved quotes @samp{‘}, @samp{’}, @samp{“}, and @samp{”} are
Well here we have @samp{...} instead of @t{...}, which also fails to
show “ and ”, displaying instead \ and " (just like @t{...}).  But
it looks good in HTML.

> In PDF 22.5 Quotation Marks:

First little bonus from TEXT.TEXI (L424-425):
"The funny quoting below is to make the printed version look correct.
FIXME."

>    In 1st paragraph, 2nd line: 1st "like this" is surrounded with single
> curved quotes, while they should be single straight quotes.

In TEXT.TEXI - L427:
# using straight apostrophes @t{'like this'} or double-quotes @t{"like
Similar to above example, @kbd{'}@t{like this}@kbd{'}.  Looks good in
HTML.

>    In 1st paragraph, 3rd&4th line: both "like this" are surrounded with
> good quotes, but they have bad shape (normal text), while should be
> typewriter/verbatim.

In TEXT.TEXI - L429:
# left and right single or double quotation marks `@t{like this}' or
Switch to - @t{‘like this’} - with #x2018 & #x2019.

The second "like this" is L/R double quotation marks, so again no answer.

Both look bad in HTML.

>    In 2nd paragraph, 2nd line: there should be straight quotes
> followed by their curved types, but unfortunately straight quotes are
> curved as well.  Also curved have bad shape (normal text), should be
> typewriter/verbatim.

In TEXT.TEXI - L442-443:
- type characters it optionally converts @t{`} to ‘, @t{'} to ',
- @t{``} to ``, and @t{''} to ''.  It's possible to change the
+ type characters it optionally converts @kbd{`} to @t{‘}, @kbd{'} to @t{’},
+ @kbd{``} to ??, and @kbd{''} to ??.  It's possible to change the
Of course I don't know what to put for “ and ”, so I put ?? there.
Also it looks bad in HTML.

>    In 2nd paragraph, "value" at the end: curved quotes have bad shape
> (normal text), should be typewriter/verbatim.

In TEXT.TEXI - L448:
# default value is @code{'(?@r{`} ?@r{'} ?@r{``} ?@r{''})}.
Perhaps first two could be changed to normal @t{‘} and @t{’}.  Last
two - a mystery.  Also it looks bad in HTML.

>    In 4th paragraph, 5th line: curved quotes have bad shape (normal
> text), should be typewriter/verbatim.

In TEXT.TEXI - L469:
# @t{’}, @kbd{C-x 8 @{} for ``, and @kbd{C-x 8 @}} for ''.
Again L/R double quotation mark.  Also it looks bad in HTML.

====================

Now about Unicode code & name pairs.

> I don't see anything wrong with the current typeface, so I left it
> alone.

In BASIC.TEXI - L116 we have:
   @code{U+2018} LEFT SINGLE QUOTATION MARK
In SEARCH.TEXI - L1313-1314 and L1319-1320 we have:
   @sc{u+249c parenthesized latin small letter a}
   @sc{u+2100 account of}
   @sc{u+fb00 latin small ligature ff}
In TEXT.TEXI - L430 (inside footnote) we have:
   U+2018 LEFT SINGLE QUOTATION MARK
   U+2018 RIGHT SINGLE QUOTATION MARK <--- this has wrong code!
   U+201C LEFT DOUBLE QUOTATION MARK
   U+201D RIGHT DOUBLE QUOTATION MARK
So, 3 different styles.  I think @code{...} around Unicode code and
uppercase "U" is a must.  This is how they are displayed in many other
places.  Style of name - I don't know, I would pick normal uppercase,
because of simplicity, but it's up to you.

====================

> > 6.  In INFO 15.1.1 Basics of Incremental Search:
> > - ‘<ESC> <ESC> <ESC>’ (‘isearch-cancel’) or ‘C-g C-g’ (‘isearch-abort’).
> > + ‘<ESC> <ESC> <ESC>’ (‘isearch-cancel’) or ‘C-g’ (‘isearch-abort’).
> > Because `view-lossage' and `describe-bindings' and the last paragraph
> > of 15.1.4 say: `C-g'.
>
> I left this unaltered, because in some cases you do need to type C-g
> twice, so doing it twice always is safer.

Well I think the last paragraph of 15.1.4 pointed by me explains this
behaviour.  It exactly says that sometimes C-g needs to be typed twice
to exit search.  That's why I'm sticking to my version, unless you had
other cases in mind.  And "isearch-abort" is literally binded to "C-g"
so it may rise questions.