On 04/18/2016 12:18 PM, Pedro Alves wrote:
>>> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>> >>
>>> >> Symbols which may be used and are missing in new libreadline:
>>> >> rl_complete_with_tilde_expansion
>>> >> _rl_mark_modified_lines
>>> >>
>>> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>> >>
>>> >> I wasn't expecting to ever see a "rl_" symbol here?  What was the
>>> >> logic used to decide whether a symbol should be exported?
>> > 
>> > I only kept symbols which are present in public headers.
> ...
> 
>> > All three are declared in a file with descriptive name rlprivate.h.
> Ah.  I wonder why they're named that way instead of using
> the _rl_ prefix?  In any case, seems to me that whatever policy applied
> to _rl_ symbols used by applications should be applied to these,
> and there's at least one program using rl_complete_with_tilde_expansion.
> 

For the record, here's where the two references above come from, on Fedora 23:

$ readelf -sDW /usr/bin/nmcli | grep rl_complete_with_tilde_expansion
  706   1: 000000000028e6a0     4 OBJECT  GLOBAL DEFAULT  26 
rl_complete_with_tilde_expansion

$ readelf -sDW /usr/bin/lftp | grep _rl_mark_modified_lines
  272  43: 000000000020f598     4 OBJECT  GLOBAL DEFAULT  25 
_rl_mark_modified_lines


-- 
Thanks,
Pedro Alves

_______________________________________________
Bug-readline mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/bug-readline

Reply via email to