This is the mail archive of the newlib@sources.redhat.com mailing list for the newlib project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: isspace() & i18n


egor duda wrote:
> 
> Hi!
> 
> Thursday, 07 June, 2001 J. Johnston jjohnstn@cygnus.com wrote:
> 
> JJ> My proposal is to change ctype.h now to use a new pointer instead
> JJ> of _ctype_.  The pointer will get set in ctype_.c.  In ctype_.c,
> JJ> we can use your GNUC code to set up _ctype_ to point to within the
> JJ> expanded table.  Additionally, we set the new ctype pointer
> JJ> appropriately.
> 
> i guess this pointer has to be exported from library as we change
> ctype.h, hasn't it? at least for cygwin this means changes to export
> definitions file in cygwin sources. i'm not sure about other
> platforms, however. won't some of require similar changes?
> 
> but what worries me more is that new applications won't link with old
> dll (in case of cygwin). as long as almost every nontrivial program
> uses <ctype.h> this can be a problem -- "cygwin-1.3.3 or higher
> required" may be quite disappointing restriction.
> 
> so, i've changed ctype.c to define end export both pointer and array,
> but left ctype.h untouched for now. eventually, we can make use of the
> pointer, but i think we shouldn't do it right now. patch attached.
> 

Egor,

  Thanks for the updated patch.  I will be checking it in shortly with some minor tweaking.

  Could you explain what scenarios you envision where the header files end up newer than the Cygwin
dll?  I could see a user updating the dll without updating the header files, but I find it highly
unlikely the opposite would occur.  I also think a user would understand if they were to downgrade
their dll without also back-leveling the accompanying header files, all bets would be off. 

-- Jeff J.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]