Re: final batch of tiny string fixes [patches]

[Jim Meyering]

Benno Schulenberg <address@hidden> wrote:

> Jim Meyering wrote:
>> Benno Schulenberg <address@hidden> wrote:
>> > -             ngettext ("WARNING: %" PRIuMAX " of %" PRIuMAX
>> > -                       " listed file could not be read",
>> > +             ngettext ("WARNING: %" PRIuMAX " file could not be read",
>> ...
>> Ok, but rather, how about rewording it to retain the total number
>> of files considered.  That can be important information.  Imagine
>> the user thinks she is verifying the checksums of 10 files when
>> in fact only 9 input lines are valid.  Omitting the "9" could be
>> fatally misleading.
>
> Hmm...  But in general it is unlikely that the user knows how many

"unlikely" isn't a strong enough argument for presenting
less information, when security might be at stake.

> files should be verified.  For small numbers, okay, but not for
> tens or hundreds.  More importantly, though, this summary line only
> gets printed when one or more checksums do not match, so the user
> will not be alerted to too few files being verified (if she knows
> the total) when all valid lines verify OK.  If she wants to be
> alerted to misformatted lines, she should use -w.  But when
> verifying lots of files, any such error message will probably
> scroll quickly off the screen, so instead of mentioning the total
> number of checked files, it would be more useful to add another
> summary line that mentions the total number of improperly formatted
> lines.  This line should probably be printed before the summary of
> checksum mismatches, and probably only when -w is used.

If you find a way to make it more easily translatable
without removing information, we can revisit this.

>> > -  printf (" %19s", "        ???");
>> > +  /* TRANSLATORS: Real name is unknown; at most 19 characters.
>> > */ +       printf (" %19s", _("        ???"));
>> ...
>> Why bother translating these in the first place?
>
> For symmetry, because in print_long_entry() it is gettextized too.
> And because some languages may have a special word or character or
> designation for "unknown name", "unknown person", "unknown thing".
> Or some translators may simply prefer to use a clearer message
> than "        ???".

Ok.  You've convinced me.  I applied that, too.

>> I would welcome a patch that unifies and (if possible) then
>> factors out some of the duplication displayed by this command:
>>
>>     grep -E 'MB? =?10..\*10' src/*.c
>
> Okay.  But I won't have time to make this: in a few days I will
> break camp.  So this will probably be my last message to this list,
> not counting the upcoming pings.

Thanks again for all your work.