Re: [RP] select patch

[Jeff Abrahamson]

Thanks!  This looks good, up to my late-afternoon ability to review code.  I trust anyway that you'll eventually test it, too. ;-)

Do you ever allocate or deallocate curname?

Jeff Abrahamson
+33 6 24 40 01 57   <-- brièvement indisponible le 4 juillet

+44 7920 594 255    <-- will change 18 July

http://jeff.purple.com/
http://blog.purple.com/jeff/

On 19 June 2014 19:12, Johannes Altmanninger <address@hidden> wrote:

On 06/19/2014 03:47 PM, Bernhard R. Link wrote:

* Jeff Abrahamson <address@hidden> [140618 20:52]:

    - Using raw strcmp() isn't a very good idea -- if one of the strings is
    not null-terminated, you introduce a subtle bug.  Prefer instead strncmp()
    and strnlen().  This requires defining some reasonable upper limit on
    window names, of course.

I have to disagree quite vehemently here. Either things are
null-terminated or they are not.
If they are then introducing a limit where there is not can only
produce bugs (subtle and non-subtle).
If they are not then most of the time things starting with str* are
usually the wrong tools anyway.

        Bernhard R. Link

In this case strncmp() is practical because I do not have to
write two different blocks for full match and prefix match (atleast until I have figured out function pointers).
strncmp() seems to be safe with a reasonable limit (it stops comparing after n bytes as far as I know)
strnlen() seems unnecessary as the whole program uses strlen()