bug#35287: dired-do-copy English beef-up

bug#35287: dired-do-copy English beef-up

[積丹尼 Dan Jacobson]

Copy: 11 files

Actually "Copied 11 files" is proper English and doesn't leave us
somewhat worrying...

bug#35287: dired-do-copy English beef-up

[Juri Linkov]

> Copy: 11 files
>
> Actually "Copied 11 files" is proper English and doesn't leave us
> somewhat worrying...

Other messages that use the same operation name:

  "Copy: 2 of 3 files skipped"

  "Copy failed for 2 of 3 files"

Shouldn't the latter be "Copying failed"?
So for every operation three verb forms are needed:
"Copy", "Copying", "Copied", correct?

bug#35287: dired-do-copy English beef-up

[Drew Adams]

> > Copy: 11 files
> > Actually "Copied 11 files" is proper English and doesn't leave us
> > somewhat worrying...
> 
> Other messages that use the same operation name:
> 
>   "Copy: 2 of 3 files skipped"
>   "Copy failed for 2 of 3 files"
> 
> Shouldn't the latter be "Copying failed"?
> So for every operation three verb forms are needed:
> "Copy", "Copying", "Copied", correct?

My own advice is to leave it alone.

If you want, for "Copy" read "Copy operation" or
"Operation Copy".

The text could even be changed to that, but
you shouldn't try for more changes than that, IMO.

(And I don't think that change is necessary or
would be very helpful.)

bug#35287: dired-do-copy English beef-up

[積丹尼 Dan Jacobson]

In my brain
>> "Copy: 2 of 3 files skipped"
>> "Copy failed for 2 of 3 files"
are good enough as is, but
>> > Copy: 11 files
makes us think maybe it meant
>> > Copy: 11 files failed
but the last part failed to appear too. That in fact is probably what is
making me nervous.
(That's why we also add a period at the end of sentences, except if we
are a cellphone chat kiddie.)
So it should say something like
>> > Copied: 11 files. # or
>> > Copy: 11 files, succeeded.
>> > Copy: 11 files, succees. # etc.

bug#35287: dired-do-copy English beef-up

[Juri Linkov]

>> > Copy: 11 files
>> > Actually "Copied 11 files" is proper English and doesn't leave us
>> > somewhat worrying...
>>
>> Other messages that use the same operation name:
>>
>>   "Copy: 2 of 3 files skipped"
>>   "Copy failed for 2 of 3 files"
>>
>> Shouldn't the latter be "Copying failed"?
>> So for every operation three verb forms are needed:
>> "Copy", "Copying", "Copied", correct?
>
> My own advice is to leave it alone.

This is important in light of our efforts for i18n.

> If you want, for "Copy" read "Copy operation" or
> "Operation Copy".

"Operation Copy" is a catchy name.

bug#35287: dired-do-copy English beef-up

[Drew Adams]

> >> > Copy: 11 files
> >> > Actually "Copied 11 files" is proper English and doesn't leave us
> >> > somewhat worrying...
> >>
> >> Other messages that use the same operation name:
> >>
> >>   "Copy: 2 of 3 files skipped"
> >>   "Copy failed for 2 of 3 files"
> >>
> >> Shouldn't the latter be "Copying failed"?
> >> So for every operation three verb forms are needed:
> >> "Copy", "Copying", "Copied", correct?
> >
> > My own advice is to leave it alone.
> 
> This is important in light of our efforts for i18n.
> 
> > If you want, for "Copy" read "Copy operation" or
> > "Operation Copy".
> 
> "Operation Copy" is a catchy name.

If you want to play with this then perhaps try to
keep the operation name as is (e.g. "copy") - don't
conjugate it or try to make a sentence/phrase.

What we have traditionally done with messages is this:

  Copying...
  Copying...done

But the `dired-do-*' framework really calls always for
an operation name (e.g. "copy", not "copying").  You
could perhaps use this:

  Copy * [2 files] to: ~/foobar/
  Copy: 2 files done

And in case of a failure, what we have now is fine:

  Copy: 1 of 2
  Copy ‘d:/foo/a.el’ to ‘d:/foo/b.el’ failed:
  (file-already-exists File already exists d:/foo/a.el)

Yes, the first line can be confusing on its own, but it's
OK with the second line.  And we definitely don't want
to say "Copy: 1 of 2 done" - the entire attempt to copy
is done, not just the attempt of one file.  We could
alternatively say "Copy: 1 of 2 succeeded".

IOW, just add "done" or "failed" or whatever to the
operation name to make clear: (1) that the operation
attempt is finished and (2) what the status of the
attempt is.

The point is to stick with "copy".  Dunno what that
would mean for translation, e.g. using the (translated)
operation name and trying to tack on a (translated)
status word like "succeeded"/"failed".

I really wouldn't mess with the `dired-do-*' framework.
And I wouldn't worry too much about translation here.

I do agree with Dan that the way it is now (without a
"done" or "succeeded") it might be a bit confusing.
But I don't think the answer is to try using "Copying"
- conjugating etc.

Keep it simple and standard (in the sense of what Emacs
has long used).  Try to fix any real ambiguity that
could be harmful, but don't go beyond that.

Just one opinion.

bug#35287: dired-do-copy English beef-up

[積丹尼 Dan Jacobson]

>>>>> "DA" == Drew Adams <drew.adams@oracle.com> writes:

DA> IOW, just add "done" or "failed" or whatever to the
DA> operation name to make clear: (1) that the operation
DA> attempt is finished and (2) what the status of the
DA> attempt is.

Good!

bug#35287: dired-do-copy English beef-up

[Juri Linkov]

>> >> > Copy: 11 files
>> >> > Actually "Copied 11 files" is proper English and doesn't leave us
>> >> > somewhat worrying...
>> >>
>> >> Other messages that use the same operation name:
>> >>
>> >>   "Copy: 2 of 3 files skipped"
>> >>   "Copy failed for 2 of 3 files"
>> >>
>> >> Shouldn't the latter be "Copying failed"?
>> >> So for every operation three verb forms are needed:
>> >> "Copy", "Copying", "Copied", correct?
>> >
>> > My own advice is to leave it alone.
>>
>> This is important in light of our efforts for i18n.
>
> You could perhaps use this:
>
>   Copy * [2 files] to: ~/foobar/
>   Copy: 2 files done

Then a complete set of messages:

  Copy: 3 files done
  Copy: 2 of 3 files skipped
  Copy: 2 of 3 files failed
  Copy: 2 files in 3 requests failed

bug#35287: dired-do-copy English beef-up

[Drew Adams]

> Then a complete set of messages:
> 
>   Copy: 3 files done
>   Copy: 2 of 3 files skipped
>   Copy: 2 of 3 files failed
>   Copy: 2 files in 3 requests failed

What is the difference between the last two?

What about the current existing messages when
there are failures, telling you to use `?' to
see the details (the `dired-log' output)?

bug#35287: dired-do-copy English beef-up

[積丹尼 Dan Jacobson]

>>>>> "DA" == Drew Adams <drew.adams@oracle.com> writes:

DA> What about the current existing messages when
DA> there are failures, telling you to use `?' to
DA> see the details (the `dired-log' output)?

You know even though the "?" works, but I never thought it is a good
idea.

The user is probably typing fast into a 1000 different windows and there
probably will be a lot of extraneous characters sent by him first to the
thing that is expecting the question mark, before he notices he is being
asked something.

So I would have the whole error message presented to him, with an 'Got
that, Holmes? (type a full "yes" to acknowledge)' style showstopper,
before letting him proceed, as failed copies are a serious matter.

bug#35287: dired-do-copy English beef-up

[Drew Adams]

> DA> What about the current existing messages when
> DA> there are failures, telling you to use `?' to
> DA> see the details (the `dired-log' output)?
> 
> You know even though the "?" works, but I never thought it is a good
> idea.
> 
> The user is probably typing fast into a 1000 different windows and there
> probably will be a lot of extraneous characters sent by him first to the
> thing that is expecting the question mark, before he notices he is being
> asked something.
> 
> So I would have the whole error message presented to him, with an 'Got
> that, Holmes? (type a full "yes" to acknowledge)' style showstopper,
> before letting him proceed, as failed copies are a serious matter.

The user is _not_ being asked anything.  She can type
`?' if/when she wants, to see the details - or not.

The `dired-do-*' operations are synchronous.  The user
cannot be typing fast into 1000 different windows while
it is working.  When it is done the user gets a message
about the possibility of using `?' to see details.  She
is not locked into any modal dialog - that message can
be ignored if she wants.  If she doesn't notice the
message it's not a big deal - like any other `message'
call.  The message is in `*Messages*' and it can be
acted on anytime - or not.

Consider that `dired-do-*' works for 1 or 2 files but
also for 1000 files.  `dired-log' is important in
providing details, but it is not critical - there is
no reason to force the user to stop and pay attention
(e.g. requiring some acknowledgement that she's seen
the message).

An alternative could be to always pop up the log buffer
(unless it is empty).  I think the existing behavior is
preferable, but that alternative could be an option.

I really don't think it's a big deal.  The info in the
log is not critical.  If it were then the dialog would
be handled differently.

bug#35287: dired-do-copy English beef-up

[積丹尼 Dan Jacobson]

Hmmm, today I tried "carelessly copying some files to / as a regular
user without lifting my eyes up to see if they failed",

 SPC [dired-next-line]
 SPC [dired-next-line]
 C [dired-do-copy]
 / [self-insert-command]
 <return> [exit-minibuffer]
 SPC [dired-next-line]
 SPC [dired-next-line]
 SPC [dired-next-line]
 SPC [dired-next-line]
 C [dired-do-copy]
 / [self-insert-command]
 <return> [exit-minibuffer]
 SPC [dired-next-line]
 SPC [dired-next-line]...

Error messages like
Copy ‘/tmp/.X0-lock’ to ‘/.X0-lock’ failed:
(file-error Opening output file Permission denied /.X0-lock)
I guess were too short to trigger the "?" perhaps,
but anyway, by the time the user types the next SPC any hint of an error
is swept away. Just like I suppose the "type '?' to see more" note would
be.

One might say "well on the shell command line if you don't look up to see
the cp error messages, it is your own fault too."

But at least they will still be sitting on the screen when you finally
look up.

And won't disappear due to a single additional key press, whether if be
from you, keyboard static electricity, a cat / baby walking on it, etc....

bug#35287: dired-do-copy English beef-up

[Drew Adams]

> Error messages like
> Copy ‘/tmp/.X0-lock’ to ‘/.X0-lock’ failed:
> (file-error Opening output file Permission denied /.X0-lock)
> I guess were too short to trigger the "?" perhaps,
> but anyway, by the time the user types the nexest SPC any hint of an error
> is swept away. Just like I suppose the "type '?' to see more" note would
> be.

Hm.  I think maybe you're right.  I haven't dug into it well
(hope someone will), but it looks like `dired-do-create-files'
no longer calls `dired-log-summary' or `dired-log' for the
actual file creation (it does so for regexp-match problems and
file-date errors).  Maybe it does indirectly somewhere -
didn't spend a lot of time looking into it.

`dired-do-copy' used to do so.  Other `dired-do-*' commands
do so, and it too should, I think.  It looks like this
functionality was lost in Emacs 24, when Emacs added
`dired-do-create-files'.  If so, not good, IMO.

No call to dired-log*' means no echo of the possibility
of using `?' to see the Dired log.

bug#35287: dired-do-copy English beef-up

[Eli Zaretskii]

> From: 積丹尼 Dan Jacobson
> 	<jidanni@jidanni.org>
> Date: Fri, 19 Apr 2019 21:48:32 +0800
> Cc: 35287@debbugs.gnu.org, Juri Linkov <juri@linkov.net>
> 
> Error messages like
> Copy ‘/tmp/.X0-lock’ to ‘/.X0-lock’ failed:
> (file-error Opening output file Permission denied /.X0-lock)
> I guess were too short to trigger the "?" perhaps,
> but anyway, by the time the user types the next SPC any hint of an error
> is swept away. Just like I suppose the "type '?' to see more" note would
> be.
> 
> One might say "well on the shell command line if you don't look up to see
> the cp error messages, it is your own fault too."
> 
> But at least they will still be sitting on the screen when you finally
> look up.

They are in *Messages*, aren't they?  Clicking the mouse on the echo
area displays them.

bug#35287: dired-do-copy English beef-up

[Drew Adams]

> They are in *Messages*, aren't they?  Clicking the mouse on the echo
> area displays them.

Maybe, maybe not.  If `dired-log*' isn't used anymore
for such errors then there won't be such messages.

But even if there are equivalent or similar messages,
without seeing the `dired-log' message you likely
won't know that they are there to find.

It seems that we lost the feature of invoking
`dired-log' for most `dired-do-creat-files' errors (?).

bug#35287: dired-do-copy English beef-up

[積丹尼 Dan Jacobson]

>>>>> "EZ" == Eli Zaretskii <eliz@gnu.org> writes:

EZ> They are in *Messages*, aren't they?  Clicking the mouse on the echo
EZ> area displays them.

Yes but the user in my example wouldn't know anything went wrong hence
wouldn't go looking in there.

bug#35287: dired-do-copy English beef-up

[Juri Linkov]

[-- Attachment #1: Type: text/plain, Size: 433 bytes --]

>> Then a complete set of messages:
>>
>>   Copy: 3 files done
>>   Copy: 2 of 3 files skipped
>>   Copy: 2 of 3 files failed
>>   Copy: 2 files in 3 requests failed
>
> What is the difference between the last two?

The latter is used for recursive operations.

> What about the current existing messages when
> there are failures, telling you to use `?' to
> see the details (the `dired-log' output)?

This is now unified as well:


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: dired-aux.i18n.patch --]
[-- Type: text/x-diff, Size: 1903 bytes --]

diff --git a/lisp/dired-aux.el b/lisp/dired-aux.el
index b81c0d1a4f..e967f94e06 100644
--- a/lisp/dired-aux.el
+++ b/lisp/dired-aux.el
@@ -1798,32 +1798,36 @@ dired-create-files
                (progn
                  (push (dired-make-relative from)
                        failures)
-                 (dired-log "%s `%s' to `%s' failed:\n%s\n"
+                 (dired-log "%s: `%s' to `%s' failed:\n%s\n"
                             operation from to err))))))))
     (cond
      (dired-create-files-failures
       (setq failures (nconc failures dired-create-files-failures))
       (dired-log-summary
-       (format "%s failed for %d file%s in %d requests"
-		operation (length failures)
-		(dired-plural-s (length failures))
-		total)
+       (format (ngettext "%s failed for %d file in %d requests"
+			 "%s failed for %d files in %d requests"
+			 (length failures))
+	       operation (length failures) total)
        failures))
      (failures
       (dired-log-summary
-       (format "%s failed for %d of %d file%s"
-		operation (length failures)
-		total (dired-plural-s total))
+       (format (ngettext "%s: %d of %d file failed"
+			 "%s: %d of %d files failed"
+			 total)
+	       operation (length failures) total)
        failures))
      (skipped
       (dired-log-summary
-       (format "%s: %d of %d file%s skipped"
-		operation (length skipped) total
-		(dired-plural-s total))
+       (format (ngettext "%s: %d of %d file skipped"
+			 "%s: %d of %d files skipped"
+			 total)
+	       operation (length skipped) total)
        skipped))
      (t
-      (message "%s: %s file%s"
-	       operation success-count (dired-plural-s success-count)))))
+      (message (ngettext "%s: %d file done"
+			 "%s: %d files done"
+			 success-count)
+	       operation success-count))))
   (dired-move-to-filename))
 \f
 (defun dired-do-create-files (op-symbol file-creator operation arg

bug#35287: dired-do-copy English beef-up

[Juri Linkov]

>>> Then a complete set of messages:
>>>
>>>   Copy: 3 files done
>>>   Copy: 2 of 3 files skipped
>>>   Copy: 2 of 3 files failed
>>>   Copy: 2 files in 3 requests failed
>>
>> What is the difference between the last two?
>
> The latter is used for recursive operations.
>
>> What about the current existing messages when
>> there are failures, telling you to use `?' to
>> see the details (the `dired-log' output)?
>
> This is now unified as well:

Pushed to master and closed.