Re: [PATCH] INSTALL: clarify that libserf is auto-detected by pkg-config

Nathan Hartman <[email protected]> Tue, 4 Aug 2026 14:44:25 -0400
Newsgroups gmane.comp.version-control.subversion.devel
Message-ID <CAJT2EHonyJGXH1XBeJ0voQM7BgVRK8Md=Ob26FvfNfORwKQ2pg@mail.gmail.com>
On Tue, Aug 4, 2026 at 2:04=E2=80=AFPM Nathan Hartman <hartman.nathan@gmail=
.com> wrote:
>
> On Tue, Aug 4, 2026 at 12:37=E2=80=AFPM Pavel Lyalyakin via dev <dev@subv=
ersion.apache.org> wrote:
>>
>> Hello,
>
> (snip)
>>
>> The thing is that the INSTALL document needs a major rework (I think tha=
t everyone can agree with this).
>
> (snip)
>
> I agree with this!
>
> It is something I have looked into in the past, but every time I did, I b=
ecame overwhelmed by how extensively INSTALL needed to change.
>
> One good thing is that after the CMake & vcpkg build becomes the "officia=
l" way to build on Windows in 1.16, a huge amount of cruft can be removed (=
all of section II-E). That will result in renumbering the later sections, s=
o maybe it's a good time to just go in there with a wrecking ball.
>
> Thanks for giving INSTALL some much needed love!!
>
> By the way, I don't agree that it's not glamorous: for many users, INSTAL=
L is likely one of the first pieces of documentation they encounter once th=
ey decide to try Subversion. It was for me when I installed Subversion for =
my $dayjob for the first time in 2007. INSTALL was very good, very clear, a=
nd it gave me confidence that I was dealing with a well made tool. There's =
something glamorous in that :-)
>
> Cheers,


I have nominated both changes (r1936876, r1936877) for backport to
1.15.x in r1936879.

Cheers,
Nathan