[Buildroot] [PATCH] Remove readline for which

[Buildroot] [PATCH] Remove readline for which

[Wade Berrier]

Hi,

I've patched 'which' to not require readline on the source level
instead of selecting it on the buildroot dependency level.

Patch attached.

Wade
-------------- next part --------------
A non-text attachment was scrubbed...
Name: buildroot-which_no_readline.patch
Type: text/x-diff
Size: 937 bytes
Desc: not available
Url : http://busybox.net/lists/buildroot/attachments/20080805/72ac59e8/attachment.patch 

[Buildroot] [PATCH] Remove readline for which

[Bernhard Fischer]

Hi Wade, *,

On Tue, Aug 05, 2008 at 01:44:50PM -0600, Wade Berrier wrote:
>Hi,
>
>I've patched 'which' to not require readline on the source level
>instead of selecting it on the buildroot dependency level.

I suggest you try to get that fixed upstream so they do allow for
./configure --disable-readline-and-yes-i-really-do-not-want-it

Short of that, why don't you use busybox in the first place?

I'm really curious how big your "fixed" which binary is?

Mine looks like this, perhaps your's is smaller?
$ size debianutils/which.o
   text	   data	    bss	    dec	    hex	filename
    211	      0	      0	    211	     d3	debianutils/which.o

Just to give you some background: buildroot is ment to serve as an
environment for uClibc. That's it's primary goal AFAIU. Second, it's
ment to serve as an easy way to get you a _small_ filesystem (thus using
busybox targeted at uClibc).

There is really no reason why .._HIDE_OTHER is visible to a casual user
since it's not ment to be actually used, but just to test uClibc, in my
POV. Please, just set BR2_PACKAGE_BUSYBOX_HIDE_OTHERS=y in your toplevel
.config.
>
>Patch attached.

As said. Re-read the patch. Send it to the maintainers of "debianutils"
and understand why they will most likely reject it in it's current form.
HTH,

[Buildroot] [PATCH] Remove readline for which

[Wade Berrier]

Hi,

On Tue, Aug 5, 2008 at 2:05 PM, Bernhard Fischer <rep.dot.nop@gmail.com> wrote:
> Hi Wade, *,
>
> On Tue, Aug 05, 2008 at 01:44:50PM -0600, Wade Berrier wrote:
>>Hi,
>>
>>I've patched 'which' to not require readline on the source level
>>instead of selecting it on the buildroot dependency level.
>
> I suggest you try to get that fixed upstream so they do allow for
> ./configure --disable-readline-and-yes-i-really-do-not-want-it

I saw some changes in their 'NEWS' implying not wanting to depend on
readline any longer, so I'm wondering if this slipped through on a
bash update.  (ie: which 2.16 doesn't depend on readline, whereas 2.19
does).

>
> Short of that, why don't you use busybox in the first place?
>
> I'm really curious how big your "fixed" which binary is?
>
> Mine looks like this, perhaps your's is smaller?
> $ size debianutils/which.o
>   text    data     bss     dec     hex filename
>    211       0       0     211      d3 debianutils/which.o
>
> Just to give you some background: buildroot is ment to serve as an
> environment for uClibc. That's it's primary goal AFAIU. Second, it's
> ment to serve as an easy way to get you a _small_ filesystem (thus using
> busybox targeted at uClibc).
>
> There is really no reason why .._HIDE_OTHER is visible to a casual user
> since it's not ment to be actually used, but just to test uClibc, in my
> POV. Please, just set BR2_PACKAGE_BUSYBOX_HIDE_OTHERS=y in your toplevel
> .config.

The system I'm working on isn't as concerned with filesystem size.
I'm upgrading to a recent buildroot (mainly for a toolchain and kernel
upgrade) and need to keep the package selection as close as possible
to the original configuration from a few years ago.

Nonetheless, I do appreciate the background information and explanations.

>>
>>Patch attached.
>
> As said. Re-read the patch. Send it to the maintainers of "debianutils"
> and understand why they will most likely reject it in it's current form.
> HTH,
>

As above, I think this reference to readline slipped through.  I'll go
see what they say.

Thanks!

Wade

[Buildroot] [PATCH] Remove readline for which

[Wade Berrier]

Hi,

On Wed, Aug 6, 2008 at 8:48 AM, Wade Berrier <wberrier@gmail.com> wrote:
> On Tue, Aug 5, 2008 at 2:05 PM, Bernhard Fischer <rep.dot.nop@gmail.com> wrote:
>> Hi Wade, *,
>>
>> As said. Re-read the patch. Send it to the maintainers of "debianutils"
>> and understand why they will most likely reject it in it's current form.
>> HTH,
>>
>
> As above, I think this reference to readline slipped through.  I'll go
> see what they say.

I mailed the GNU which maintainer and it turns out this patch was
already committed upstream.

Wade