Re: regexec(3): REG_STARTEND is not documented

[Alejandro Colomar <alx.manpages@gmail.com>]

[-- Attachment #1.1: Type: text/plain, Size: 2225 bytes --]

On 4/21/23 03:41, наб wrote:
> Hi!
> 
> On Fri, Apr 21, 2023 at 03:28:42AM +0200, Alejandro Colomar wrote:
>> On 4/21/23 03:15, наб wrote:
>>> On Fri, Apr 21, 2023 at 03:07:00AM +0200, Alejandro Colomar wrote:
>>>> Here's a related question:
>>>>     regmatch_t  pmatch = {
>>>>         .rm_so = string,
>>>>         .rm_eo = string + 42,  // Assume this offset is valid
>>>>     };
>>>>     regexec(preg, string, 0, pmatch, REG_STARTEND);
>>>> Should regexec(3) write to the 1st element in pmatch[] because it knows
>>>> it exists (otherwise the call would be UB because it needs to read it)?
>>> (Which would run counter to how POSIX defines the API.)
>>>
>>>> Or is passing 0 in nmatch effectively another way of performing
>>>> REG_NOSUB behavior without actually using the flag?
>>> Hilariously enough, quoth 4.4BSD-Lite regex(3) again,
>>> which phrases it exactly like you do:
>>>   If  REG_NOSUB  was specified in the compilation of the RE, or if nmatch
>>>   is 0, regexec ignores the pmatch argument (but see below for  the  case
>>>   where REG_STARTEND is specified).
>> Touche; it looks like your right.  That sentence is unambiguous.  BTW, is the
>> reference to some other text about REG_STARTEND the one quoted first
>> (above)?
> That's the first mention of [np]match afterward, yeah. The file's
> massive, so I've attached it here, you can read more brain-destroying
> minutiae there; it's modern-enough mdoc(7) that anything should render
> it sensibly.
> 
> I've also noticed (even later, did I mention they couldn't stop writing?):
>   If REG_STARTEND is specified, pmatch must point to at  least  one  reg‐
>   match_t  (even  if nmatch is 0 or REG_NOSUB was specified), to hold the
>   input offsets for REG_STARTEND.  Use for output is still entirely  con‐
>   trolled by nmatch; if nmatch is 0 or REG_NOSUB was specified, the value
>   of pmatch[0] will not be changed by a successful regexec.

Holly antelope!  At least we can't have doubts of the intended interface :)
It seems to confirm everything you suspected ;)

> 
> Best,
> наб

-- 
<http://www.alejandro-colomar.es/>
GPG key fingerprint: A9348594CE31283A826FBDD8D57633D441E25BB5

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]