On Wed, Jul 31, 2013 at 8:09 AM, Lionel Cons <[email protected]> wrote:
> On 31 July 2013 02:18, Cedric Blancher <[email protected]> wrote:
>> I've a small RFE for the sake of writing shell signal traps which use
>> .sh.sig a bit easier.
>>
>> Currently ksh either use custom names ("stopped", "exited", "dumped")
>> or numbers in .sh.sig.code. This is frustrating from the viewpoint of
>> script programmer who actually have to look into the ksh sources or
>> try to trigger the matching event to figure out the string value ksh
>> uses for a given si_code.
>>
>> Could we just avoid this and just use the uppercase names defined by
>> Opengroup in <signal.h>, i.e.
>> http://pubs.opengroup.org/onlinepubs/009696699/basedefs/signal.h.html?
>> Otherwise ksh93(1) must be updated to list each signal and the names
>> or numbers (which aren't portable BTW) used.
>
> It would also provide safety in case you have si_code names
> ("${prefix}_${suffix}", like prefix="CHD", suffix="EXITED") which have
> a different prefix but the same suffix. The current model risks such
> clashes.
>
> +1 for the proposal

In theory the original naming scheme tried to be KISS (Keep It Simple,
Stupid). Unfortunately there are already naming clashes between prefix
and suffix, like SEGV_ADRERR and BUS_ADRERR, which makes the current
approach risky.

So, +1 for the proposal

Irek
_______________________________________________
ast-developers mailing list
[email protected]
http://lists.research.att.com/mailman/listinfo/ast-developers

Reply via email to