[tex-k] Why is B952. "Don't allow implicit left brace after |#| (Udo Wermuth). @476" considered a bug?

Karl Berry karl at freefriends.org
Mon Mar 1 01:10:35 CET 2021

Hi Ulrich,

    What are the exact conditions under which things that deviate
    from the behavior described in the TeXbook can be considered bugs?

I cannot give you a definition. At this point in TeX's life, nearly
anything that can be found is in a gray area.

    With many things that I have raised over the years and found
    inconsistent with the description in TeXBook I have been told by
    the luminaries at comp.text.tex/de.comp.text.tex/

All I can tell you is that if you find inconsistencies that seem worth
reporting to you, send them in, regardless of what other people say
(https://tug.org/texmfbug). The small group of us who vet the bugs for
Knuth (not to be coy; it's me, David Fuchs, Donald Arseneau, and --
still providing assistance though no longer lead "entomologist"
--Barbara Beeton) are not active on comp.text.tex, much less
de.comp.text.tex. Is anyone active on c.t.t any more :)? Last time I
looked there it was nothing but automated CTAN announcements. Anyway ...

    the "technique of deliberate lying" is praised a means of easing

The material involved in this bug report is all double-dangerous bends,
which I don't think ever has deliberate lies (although admittedly Don
never says precisely that, just "better approximations to the truth").

Anyway, we passed along this report from Udo because there was a
definite inconsistency between The TeXbook and tex.web.  TB said "<left
brace> and <right brace> are explicit character tokens", while tex.web
accepted implicit tokens for <left brace>. That is not just a matter of
sloppiness, it's a factual discrepancy.

We suggested to Don that it would be better to change The TeXbook than
to change the code, but he chose to change the program. He wrote this
comment on the report:
  DEK> ... I want a clean rule for <definition text> and I've got one,
  so I want the bottom of p.275 to be valid forever. [...] If anybody has
  used that "feature" [...] they [can] easily fix their code: instead of
    \def\cs alpha#\bgroup beta}
    \def\cs alpha\bgroup{beta \bgroup}

I admit those two TeX lines are not something I can just look at and
immediately convince myself they are equivalent, but it's what Don wrote
(well, except he wrote real \alpha and \beta Greek letters). At least
I'm pretty sure I transcribed it right.

In any event, it's been changed and he will not consider it again, so it
is what it is. --best, karl.

More information about the tex-k mailing list.