[m-users.] Misunderstanding (was: Link-time optimization)

[Massimo Dentico]

On 14/06/2020 10:59, Richard O'Keefe wrote:
> [...]
> 
> All things considered, generating a single file is not
> obviously useful for Mercury, even for distribution.

There is a fundamental misunderstanding here and it's probably my fault,
because I see that even Zoltan Somogyi is referring to the size of the
Mercury compiler.

So let's let's clarify this point: I'm interested in compiling the
Mercury compiler with LTO because I think there isn't another easy way
to force the compiler to compile *my programs* with LTO. At least the
run-time and the libraries must be compiled with LTO to have an
effective reduction in size, but I think disentangling the build process
of these from that of the compiler is no easy. Someone who know more
should correct me if I'm wrong. This was question a).

My question b) is about single file deployment and is not necessary
related: I'm *not* interested in obtaining the Mercury compiler as a
single C file. I was speculating about deploying *my Mercury programs*,
that I anticipate of the size of at least an order of magnitude smaller
than the Mercury compiler (see my last reply to Zoltan), as single C
files.

--
Massimo A. Dentico