bug#24870: 26.0.50; parse-partial-sexp ignores comment-end

[Matt Armstrong <marmstrong@google.com>]

Andreas Röhler <andreas.roehler@easy-emacs.de> writes:

> On 30.11.2016 10:10, Matt Armstrong wrote:
>> Andreas Röhler <andreas.roehler@easy-emacs.de> writes:
>>> haskell-mode, at EOB: --- {- To explore this file: -} asdf = --- 
>>> parse-partial-sexp thinks being inside a paren - see attachment. GNU 
>>> Emacs 26.0.50.1 (i686-pc-linux-gnu, GTK+ Version 3.14.5) of 2016-10-11 
>> Hi Andreas, Emacs does not have a haskell-mode, so this bug is 
>> difficult to reproduce. It may be more appropriate to report this to 
>> the haskell-mode maintainers for triage. They can figure out if it is 
>> a problem that should be fixed in haskell-mode itself, or a problem 
>> with Emacs. Alternatively, can you provide a series of clear 
>> instructions to reproduce the problem in a fresh Emacs started without 
>> your customizations? For example, begin by running "emacs -Q" and go 
>> from there. Your attached .png presents a buffer called 
>> *parse-partial-sexp-output*, but it is not clear how this was 
>> generated. Thanks
> Hi Matt,
>
> checked that with help of the haskell-mode folks already.
> https://github.com/haskell/haskell-mode/issues/1459

Thanks.  For reference, this is what you have said on the github bug:

Seems a bug of
GNU Emacs 26.0.50.1 (i686-pc-linux-gnu, GTK+ Version 3.14.5) of 2016-11-15
Does not exist at
GNU Emacs 25.1.90.1 (i686-pc-linux-gnu, GTK+ Version 3.14.5) of 2016-11-29


> As it's nice at current pretest Emacs, concluded a bug in trunk.
> Here a shortened recipe. Put code below in a buffer:
> ||
> |{- Just a comment: -}|
>
> M-x haskell-mode RET

Note that haskell-mode is not part of Emacs.  Ideally, your steps to
reproduce that begin with running "emacs -Q".  Emacs maintainers that
might not also be Haskell hackers will appreciate it.  :)

Also, can you describe the visible symptom that caused you to begin
looking at syntax-ppss and parse-partial-sexp?  That description may
help me or others spot similarities with other reported bugs.

(I must say that I am not an Emacs expert, and I do not usually reply to
Emacs bugs.  I looked at a few bugs last night as a way to help
maintainers triage the "easy" bugs.  It does not look like this bug is
easy!)