bug#47505: 25.1; ediff thinks words changed when filling prefixed

bug#47505: 25.1; ediff thinks words changed when filling prefixed

[Samuel Wales]

to reproduce, please put this in file a:

# asdfjkansdkljfan ksd fnak nsdflkadsnfkansdkfnakdsnf
# aklsfdnakljfnsdkasdnfja sdn fkljad nsfkj ansdkfn aksdjfna
# ksjfndk ajdnsf kajsndkfjansdkfjn ja nsdkjfn askdj fnakjs fdnakj
# nsdfkaj dns

then set fill-column to 60.  then fill.  then save to file b.

you now have a and b with the same text, but a filled commented
paragraph that differs between them.

now open in diff mode.  you will notice that it thinks only the comments
changed.  this is expected behavior.

now open with m-x ediff RET.  it thinks the first word in the last line
in b is different.  but it is not different.

best i can do for a bug report.  thanks.


On 3/28/21, Eli Zaretskii <eliz@gnu.org> wrote:
>> From: Samuel Wales <samologist@gmail.com>
>> Date: Sun, 28 Mar 2021 21:48:18 -0700
>> Cc: help-gnu-emacs@gnu.org
>>
>> for purpoes of the test do something like this:
>>
>> # asdfjkansdkljfan ksd fnak nsdflkadsnfkansdkfnakdsnf
>> # aklsfdnakljfnsdkasdnfja sdn fkljad nsfkj ansdkfn aksdjfna
>> # ksjfndk ajdnsf kajsndkfjansdkfjn ja nsdkjfn askdj fnakjs fdnakj
>> # nsdfkaj dns
>>
>> then fill to a narrow fill column.
>
> Please submit a full bug report with the details, TIA.
>
>


-- 
The Kafka Pandemic

Please learn what misopathy is.
https://thekafkapandemic.blogspot.com/2013/10/why-some-diseases-are-wronged.html

bug#47505: 25.1; ediff thinks words changed when filling prefixed

[Andreas Schwab]

On Mär 30 2021, Samuel Wales wrote:

> now open with m-x ediff RET.  it thinks the first word in the last line
> in b is different.  but it is not different.

I don't see how that is a bug.  The word moved around, and ediff
correctly marked it as such.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."

bug#47505: 25.1; ediff thinks words changed when filling prefixed

[Samuel Wales]

doyou think diff's behavior is a bug then?  i choose ediff's.

On 3/31/21, Andreas Schwab <schwab@linux-m68k.org> wrote:
> On Mär 30 2021, Samuel Wales wrote:
>
>> now open with m-x ediff RET.  it thinks the first word in the last line
>> in b is different.  but it is not different.
>
> I don't see how that is a bug.  The word moved around, and ediff
> correctly marked it as such.
>
> Andreas.
>
> --
> Andreas Schwab, schwab@linux-m68k.org
> GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
> "And now for something completely different."
>


-- 
The Kafka Pandemic

Please learn what misopathy is.
https://thekafkapandemic.blogspot.com/2013/10/why-some-diseases-are-wronged.html

bug#47505: 25.1; ediff thinks words changed when filling prefixed

[Michael Heerdegen]

Samuel Wales <samologist@gmail.com> writes:

> doyou think diff's behavior is a bug then?  i choose ediff's.

Depends on what you think what "fine diffs" are.  They are actually word
wise diffs, and comment starts are a just a word.  That how it's
implemented.  In the case you describe I find that sometimes annoying,
too.  But I guess it's not easy to change that.

It might be possible to tell Ediff to ignore certain words (like comment
starts) but it's then not trivial to decide whether some text got
commented that was not commented before.

This is all not trivial.  There can be lines that are partly code and
partly comment; in Elisp, the comment start can have changed (e.g. ;; ->
;;;), you can have "doubled comments" like ;;; ;;; where uncommenting
would leave comments ... how would you define when two regions
should count as similar?

Regards,

Michael.

bug#47505: 25.1; ediff thinks words changed when filling prefixed

[Samuel Wales]

i don't know.  i just know what seems right and seems wrong.  not too
useful is it?  but it is better than not filing the bug perhaps.

diff-mode does it right so ediff can.  filladapt can intuit prefixes.
comment syntaxes are known.  somewhere in there lies a fix maybe.

commented paragraphs in org/shell/elisp style are always commented at
bol.  idk what diff-mode does.


On 3/31/21, Michael Heerdegen <michael_heerdegen@web.de> wrote:
> Samuel Wales <samologist@gmail.com> writes:
>
>> doyou think diff's behavior is a bug then?  i choose ediff's.
>
> Depends on what you think what "fine diffs" are.  They are actually word
> wise diffs, and comment starts are a just a word.  That how it's
> implemented.  In the case you describe I find that sometimes annoying,
> too.  But I guess it's not easy to change that.
>
> It might be possible to tell Ediff to ignore certain words (like comment
> starts) but it's then not trivial to decide whether some text got
> commented that was not commented before.
>
> This is all not trivial.  There can be lines that are partly code and
> partly comment; in Elisp, the comment start can have changed (e.g. ;; ->
> ;;;), you can have "doubled comments" like ;;; ;;; where uncommenting
> would leave comments ... how would you define when two regions
> should count as similar?
>
> Regards,
>
> Michael.
>


-- 
The Kafka Pandemic

Please learn what misopathy is.
https://thekafkapandemic.blogspot.com/2013/10/why-some-diseases-are-wronged.html

bug#47505: 25.1; ediff thinks words changed when filling prefixed

[Michael Heerdegen]

Samuel Wales <samologist@gmail.com> writes:

> i don't know.  i just know what seems right and seems wrong.  not too
> useful is it?  but it is better than not filing the bug perhaps.
>
> diff-mode does it right so ediff can.  filladapt can intuit prefixes.
> comment syntaxes are known.  somewhere in there lies a fix maybe.
>
> commented paragraphs in org/shell/elisp style are always commented at
> bol.  idk what diff-mode does.

I don't know how diff-mode does it, but I wonder if it then doesn't sort
out changes that are significant.  "I don't know" is not meant ironic,
maybe someone else can help further.

Regards,

Michael.

bug#47505: 25.1; ediff thinks words changed when filling prefixed

[Lars Ingebrigtsen]

Michael Heerdegen <michael_heerdegen@web.de> writes:

> Depends on what you think what "fine diffs" are.  They are actually word
> wise diffs, and comment starts are a just a word.  That how it's
> implemented.  In the case you describe I find that sometimes annoying,
> too.  But I guess it's not easy to change that.

I think this is basically working as intended (even if it's different in
the two cases, because the interfaces emphasise different things).

So I'm closing this bug report.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#47505: 25.1; ediff thinks words changed when filling prefixed

[Samuel Wales]

i don't understand what you mean by the interfaces emphasize different
things.  to me, they should have the same goal here.  in any case,
thanks fort considering it.  if anybody has a fix, please let me know.
i might incorporate it into something useful.

On 5/8/21, Lars Ingebrigtsen <larsi@gnus.org> wrote:
> Michael Heerdegen <michael_heerdegen@web.de> writes:
>
>> Depends on what you think what "fine diffs" are.  They are actually word
>> wise diffs, and comment starts are a just a word.  That how it's
>> implemented.  In the case you describe I find that sometimes annoying,
>> too.  But I guess it's not easy to change that.
>
> I think this is basically working as intended (even if it's different in
> the two cases, because the interfaces emphasise different things).
>
> So I'm closing this bug report.
>
> --
> (domestic pets only, the antidote for overdose, milk.)
>    bloggy blog: http://lars.ingebrigtsen.no
>


-- 
The Kafka Pandemic

Please learn what misopathy is.
https://thekafkapandemic.blogspot.com/2013/10/why-some-diseases-are-wronged.html

bug#47505: 25.1; ediff thinks words changed when filling prefixed

[Samuel Wales]

do you think diff-mode's behavior is a bug?

i would say that ediff's is a bug because moving around is not what
seems significant.  comments should be honorary whitespace.


On 3/31/21, Andreas Schwab <schwab@linux-m68k.org> wrote:
> On Mär 30 2021, Samuel Wales wrote:
>
>> now open with m-x ediff RET.  it thinks the first word in the last line
>> in b is different.  but it is not different.
>
> I don't see how that is a bug.  The word moved around, and ediff
> correctly marked it as such.
>
> Andreas.
>
> --
> Andreas Schwab, schwab@linux-m68k.org
> GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
> "And now for something completely different."
>


-- 
The Kafka Pandemic

Please learn what misopathy is.
https://thekafkapandemic.blogspot.com/2013/10/why-some-diseases-are-wronged.html

bug#47505: acknowledged by developer (control message for bug #47505)

[Samuel Wales]

thanks for considering it.  if anybody has a fix for this, even as a
hook value, please let me know.  i might incorporate it into something
useful.


On 5/8/21, GNU bug Tracking System <help-debbugs@gnu.org> wrote:
> This is an automatic notification regarding your bug report
> #47505: 25.1; ediff thinks words changed when filling prefixed,
> which was filed against the emacs package.
>
> Thank you for your report, which has now been closed.
> You can view the full report at
> http://debbugs.gnu.org/cgi/bugreport.cgi?bug=47505
>
> If you require further information, please followup to
> 47505@debbugs.gnu.org.
>
> debbugs.gnu.org maintainers
> (administrator, GNU bugs database)
>
>


-- 
The Kafka Pandemic

Please learn what misopathy is.
https://thekafkapandemic.blogspot.com/2013/10/why-some-diseases-are-wronged.html