Re: error_at_line, request to add

2015-03-10 Thread Mohammad Akhlaghi

On 03/10/2015 02:03 AM, Paul Eggert wrote:

By tradition, Gnulib never installs anything.  Perhaps that tradition
should change but it should be a conscious change, presumably one
supported by the infrastructure.


I understood, thank you for the explanation.

Fortunately I was able to compile and test GNU Astronomy Utilities on a 
Mac OS X thanks to Gnulib.


I sincerely want to thank you for this huge effort in helping the 
packages become so portable. Many more users can now easily install and 
use gnuastro thanks to your efforts.


Thanks again,
Mohammad



Re: syncing regcomp.c with glibc, or further diverging, or applying patches separately.

2015-03-10 Thread Paul Eggert

Konstantin Serebryany wrote:

But some patches are not applied to Gnulib:
e.g. in Glibc's commit 39a12f8db4656bec88ebd51e524778ef2e5468b9 or
8c0ab919f63dc03a420751172602a52d2bea59a8


That's only because we haven't synced from glibc to gnulib since August.


Do you have a suggestion regarding the patch discussed in
https://sourceware.org/ml/libc-alpha/2014-09/msg00663.html?
The patch does not lead to a functionality change, but affects the
sources significantly (560 line diff)

Is that ok to apply it only to Glibc?
Or would you like to have it for Gnulib too?


We'd apply it for Gnulib too, as we prefer to keep the sources in sync as much 
as possible.  That patch applies only to "#ifdef _LIBC" code, right?  So Gnulib 
users shouldn't care about the change anyway.




Re: syncing regcomp.c with glibc, or further diverging, or applying patches separately.

2015-03-10 Thread Konstantin Serebryany
On Tue, Mar 10, 2015 at 12:15 AM, Paul Eggert  wrote:
> Konstantin Serebryany wrote:
>>
>> But some patches are not applied to Gnulib:
>> e.g. in Glibc's commit 39a12f8db4656bec88ebd51e524778ef2e5468b9 or
>> 8c0ab919f63dc03a420751172602a52d2bea59a8
>
>
> That's only because we haven't synced from glibc to gnulib since August.
>
>> Do you have a suggestion regarding the patch discussed in
>> https://sourceware.org/ml/libc-alpha/2014-09/msg00663.html?
>> The patch does not lead to a functionality change, but affects the
>> sources significantly (560 line diff)
>>
>> Is that ok to apply it only to Glibc?
>> Or would you like to have it for Gnulib too?
>
>
> We'd apply it for Gnulib too, as we prefer to keep the sources in sync as
> much as possible.

Great!

>That patch applies only to "#ifdef _LIBC" code, right?

Most of it, at least.

Thanks,

--kcc

> So Gnulib users shouldn't care about the change anyway.