[PATCH] libio: Add terminating NUL when the first character is EOF in getdelim [BZ #28038]

[Florian Weimer]

* Eric Blake:

> On Thu, Nov 13, 2025 at 01:23:04PM -0800, Collin Funk wrote:
>> > I don't have any strong opinions as to which course glibc should take;
>> > but I do think that gnulib should match whatever behavior glibc
>> > decides on, since we have now got a good list of multiple other libcs
>> > that are split on which behavior makes more sense.
>> 
>> Thanks! I'll work on a patch that reverts the previous change and
>> documents the behavior. The subsequent Gnulib patch should be simple as
>> well. It is just adjusting the tests that I added for this to expect the
>> opposite.
>
> One thing to consider: you could patch getdelim() to guarantee that if
> *lineptr is NULL on entry, then it writes *lineptr[0] to 0 after
> allocation and before reading from the buffer; and also guarantee that
> if it fails with errno set after a partial read that it writes a NUL
> at the point where read stopped.  That way, the only time *lineptr
> would NOT be NUL-terminated would be if the user passed in a
> pre-existing buffer that was not NUL-terminated; in all other cases,
> leaving the buffer unchanged on error or EOF means you still have tne
> NUL in place from entry.  That would at least get rid of the UB
> pointed out in the bug where the newly-malloc'd buffer is completely
> uninitialized.  That's different than the current patch under question
> which merely writes *lineptr[0] on encountering EOF; and my
> alternative patch idea will not protect users that pass in their own
> pre-malloc'd uninitialized buffer, but might be a better compromise
> position compared to a full reversion.

I like this proposal.

Thanks,
Florian