[PATCH v9 0/13] implement dlmem() function

[stsp]

29.03.2023 23:13, Carlos O'Donell пишет:
> It is not that the workload you have is unimportant, since importance is relative,
> but the workload is not generic enough for glibc to add dlmem() into the library.
>
> You need to have a very strong argument for the question "Why should this go into
> the library?"
I'll write that down to the cover-letter,
but I'd like to also summarize it here to
get your opinion before posting the patch.

My view on that picture is that dlmem()
is a very generic API that allows to implement
eg dlopen_with_offset() in a dozen of lines.
And in that case - quite importantly -
no "advanced" features are even used, so
the last arg is just NULL.

Now the advanced features offered by the
last argument, are indeed quite specific to
my not-so-common use-case. So my idea
is exactly that: implement what other ppl
asked before, make it powerful and generic,
and then sneak in an optional pointer arg
for myself. :)
Is it a viable plan?
I also tried to "hide" the me-specific functionality
in an auditor - that plan was rejected as
non-viable. Optional pointer arg looks a
bit better to me. :)

So I am trying to put this not as "I want
to implement the function solely for myself",
but as "please let me have an optional
pointer arg in a very generic and powerful
function". :)