Bruno Haible <[email protected]> writes:

> Eric Blake wrote:
>> the list is sorted alphabetically,
>> with no bearing on dependency chains other than that explicitly
>> requested modules appear with less indentation.
>
> Yes, that's how it's done.

Oops.  Thanks for explaining, I was reading too much into the output.

>> Maybe listing the dependency of each module would be nicer than the
>> current alphabetic list
>
> Listing the dependencies of each module would lead to a much bigger output
> (think of how often 'stdint' or 'unistd' would occur...); this is not what
> most users want. I think we should be content to list every module just once.
>
> This indentation feature was the outcome of this discussion:
>   <http://lists.gnu.org/archive/html/bug-gnulib/2010-06/msg00132.html>
> Do you have a better idea for presenting the specified and the dependent
> modules in a lightweight way?

Not immediately.  My question was raised because I didn't know why a
certain module was pulled in, and it wasn't immediately clear from the
modules I requested.  A different way to resolve this problem could be
with a 'gnulib-tool --why strdup' command that could print:

  uniconv/u8-strconv-from-locale
  uniconv/u8-strconv-from-enc
  uniconv/u8-conv-from-enc
  striconveha
  strdup

with the first line being something I manually requested.

Just an idea, it might be too much work to implement this logic in shell
script.  There is complexity because there may be multiple paths too,
although I think naming only one would be sufficient.

(And, btw, is the strdup dependency from striconveha really needed?)

/Simon

Reply via email to