Re: [PATCH 0/2] More uses of make_unique_xstrdup

[Guinevere Larsen <guinevere@redhat.com>]

On 1/15/26 3:01 PM, Simon Marchi wrote:
> On 1/15/26 12:39 PM, Guinevere Larsen wrote:
>> On 1/15/26 2:33 PM, Tom Tromey wrote:
>>> This replaces some uses of '.reset(xstrdup())' with
>>> '= make_unique_xstrdup()'.  I think explicit uses of reset
>>> should be avoided when possible.
>>>
>>> Regression tested on x86-64 Fedora 41.
>>>
>>> Signed-off-by: Tom Tromey <tromey@adacore.com>
>> This makes me wonder... is there any reason to prefer xmalloc/xstrdup instead of using std::string?
>>
>> My preference is to use stdlib stuff when possible, but maybe I'm misjudging the size of the refactor or there's a reason to prefer the current memory strategy.
> When practical to switch to std::string, it typically is an improvement.
> It makes strings manipulation safer and clearer.  There might be cases
> where it is not trivial to use std::string immediately though, so
> switching to make_unique_xstrdup is a good incremental change
> nonetheless.
>
> Simon
>
Right that makes sense. I was moreso double-checking my gut feeling, 
rather than asserting that the series should convert to std::string

-- 
Cheers,
Guinevere Larsen
It/she