Handling argc==0 in the kernel [LWN.net]

Handling argc==0 in the kernel

Posted Jan 28, 2022 16:57 UTC (Fri) by corbet (editor, #1)
In reply to: Handling argc==0 in the kernel by larkey
Parent article: Handling argc==0 in the kernel

Allowing argv[0]==NULL would indeed allow the "old as Unix" pattern to work. But it also works fine now in the argv==NULL case. The pattern, perhaps equally old, that fails is:

    for (arg = argv + 1; *arg; arg++)
        /* ... */

...and that pattern will still fail badly with argv[0]==NULL.

to post comments

Handling argc==0 in the kernel

Posted Jan 28, 2022 17:00 UTC (Fri) by larkey (guest, #104463) [Link] (3 responses)

Arguably a bad pattern :-)

But I think that Ariadne's patch should be applied nontheless. It's just that there's much confusion about these two different issues (although related).

Handling argc==0 in the kernel

Posted Jan 29, 2022 18:56 UTC (Sat) by adobriyan (guest, #30858) [Link] (2 responses)

The best, UB free, pattern is

for (int i = 1; i < argc; ++i) { const char *arg = argv[i]; ... }

Handling argc==0 in the kernel

Posted Jan 30, 2022 23:32 UTC (Sun) by rschroev (subscriber, #4164) [Link] (1 responses)

That's simple, readable, clearly states the intent. Why do people even try to come up with other ways to do things?

Handling argc==0 in the kernel

Posted Jan 31, 2022 1:01 UTC (Mon) by jreiser (subscriber, #11027) [Link]

>Why do people even try to come up with other ways to do things?

Because that does not accommodate state changes: deleting, re-arranging, or re-writing pieces of argv (and thus argc); because the GNU getopt routine has not always existed, has not always offered all its current features, and early suffered from bugs; and probably other reasons, too.