strtok behaviour when uninitialized

[Javier Serrano Polo]

El dg 11 de 02 de 2018 a les 19:02 -0500, Rich Felker va escriure:
> The issue is that a crash is the
> best way to inform the programmer that their code is incorrect.

This is your way to handle undefined behavior in specification and it is
fine, but you do define behavior by asserting a catastrophic failure.
However, your wiki says "correct programs performing better is more
important to us than buggy programs being easier to debug". This
counters catastrophic failure if it means an extra runtime cost.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3386 bytes
Desc: not available
URL: <http://sourceware.org/pipermail/libc-alpha/attachments/20180212/0b6bb90d/attachment.p7s>