11:01 a.m.
Hi,
Apparently the linker on Ubuntu is fussy about the order of flags,
I've tested this fix on Ubuntu 18.04 and 19.04.
Makes me wonder if it's worth having CI to catch this kind of stuff?
Thanks,
Richard
Richard Ipsum (1):
Fix linker error on Ubuntu
Makefile | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
--
2.11.0
11:01 a.m.
New subject: [PATCH 1/1] Fix linker error on Ubuntu
---
Makefile | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/Makefile b/Makefile
index 5826b73..a4965d2 100644
--- a/Makefile
+++ b/Makefile
@@ -52,7 +52,7 @@ luxio-5.1.so: luxio-5.1.o
$(CC) $(LDFLAGS) -shared $(EXTRA_LIBS) -o luxio-5.1.so $^
luxio-5.1: lua-5.1.c luxio-5.1.o
- $(CC) -o luxio-5.1 $(CFLAGS) $(LUA51_INC) $(LDFLAGS) $(EXTRA_LIBS) lua-5.1.c luxio-5.1.o
$(LUA51_LIB)
+ $(CC) -o luxio-5.1 $(CFLAGS) $(LUA51_INC) lua-5.1.c luxio-5.1.o $(LUA51_LIB) $(LDFLAGS)
$(EXTRA_LIBS)
.PHONY: lua-5.2
lua-5.2: luxio-5.2.so luxio-5.2
@@ -73,7 +73,7 @@ luxio-5.2.so: luxio-5.2.o
$(CC) $(LDFLAGS) -shared $(EXTRA_LIBS) -o luxio-5.2.so $^
luxio-5.2: lua-5.2.c luxio-5.2.o
- $(CC) -o luxio-5.2 $(CFLAGS) $(LDFLAGS) $(EXTRA_LIBS) $(LUA52_INC) lua-5.2.c luxio-5.2.o
$(LUA52_LIB)
+ $(CC) -o luxio-5.2 $(CFLAGS) $(LUA52_INC) lua-5.2.c luxio-5.2.o $(LUA52_LIB) $(LDFLAGS)
$(EXTRA_LIBS)
###
--
2.11.0
12:50 a.m.
On Sat, Sep 21, 2019 at 16:01:24 +0100, Richard Ipsum wrote:
Apparently the linker on Ubuntu is fussy about the order of flags,
I've tested this fix on Ubuntu 18.04 and 19.04.
Cool, nothing about it looks bad. Merged.
Makes me wonder if it's worth having CI to catch this kind of
stuff?
I've been pondering moving luxio to gitlab or somesuch for that kind
of thing.
D.
--
Daniel Silverstone http://www.digital-scurf.org/
PGP mail accepted and encouraged. Key Id: 3CCE BABE 206C 3B69
5:15 p.m.
On Mon, Sep 23, 2019 at 05:50:14AM +0100, Daniel Silverstone wrote:
On Sat, Sep 21, 2019 at 16:01:24 +0100, Richard Ipsum wrote:
> Apparently the linker on Ubuntu is fussy about the order of flags,
> I've tested this fix on Ubuntu 18.04 and 19.04.
Cool, nothing about it looks bad. Merged.
> Makes me wonder if it's worth having CI to catch this kind of stuff?
I've been pondering moving luxio to gitlab or somesuch for that kind
of thing.
Would you be moving to a PR submission process?
Have you or would you consider something like sourcehut?
Would the code continue to be self-hosted in addition to hosting
elsewhere or are you intending to self-host with gitlab?
Or is there merit in continuing to host in gitano and adding ick-based CI?
Richard
2:41 a.m.
On Tue, Sep 24, 2019 at 22:15:50 +0100, Richard Ipsum wrote:
> > Makes me wonder if it's worth having CI to catch this
kind of stuff?
> I've been pondering moving luxio to gitlab or somesuch for that kind
> of thing.
Would you be moving to a PR submission process?
If I moved to gitlab, then (a) I'd probably expect an MR type process yes,
since that way MRs can have CI, and (b) I'd probably be looking to make someone
(you?) officially co-maintainer :D
Have you or would you consider something like sourcehut?
I've never even heard of them :(
Would the code continue to be self-hosted in addition to hosting
elsewhere or are you intending to self-host with gitlab?
I'd probably run a mirror on git.gitano.org.uk for those who were expecting it
to be available from there, but luxio's official hosting location would move.
Or is there merit in continuing to host in gitano and adding
ick-based CI?
There would be, except that Ick 1 is deprecated and Ick 2 is not ready for
use :(
D.
--
Daniel Silverstone http://www.digital-scurf.org/
PGP mail accepted and encouraged. Key Id: 3CCE BABE 206C 3B69
5:50 p.m.
On Wed, Sep 25, 2019 at 08:41:09AM +0200, Daniel Silverstone wrote:
On Tue, Sep 24, 2019 at 22:15:50 +0100, Richard Ipsum wrote:
> > > Makes me wonder if it's worth having CI to catch this kind of stuff?
> > I've been pondering moving luxio to gitlab or somesuch for that kind
> > of thing.
> Would you be moving to a PR submission process?
If I moved to gitlab, then (a) I'd probably expect an MR type process yes,
since that way MRs can have CI, and (b) I'd probably be looking to make someone
(you?) officially co-maintainer :D
Sounds cool, I'm in :)
> Have you or would you consider something like sourcehut?
I've never even heard of them :(
Sourcehut's interesting because it integrates an email-based
workflow. It's still quite rough around the edges though.
Thanks,
Richard
1:43 a.m.
On Wed, Sep 25, 2019 at 22:50:50 +0100, Richard Ipsum wrote:
> If I moved to gitlab, then (a) I'd probably expect an MR
type process yes,
> since that way MRs can have CI, and (b) I'd probably be looking to make someone
> (you?) officially co-maintainer :D
Sounds cool, I'm in :)
Excellent, handy to know. I'm currently away so don't want to make any
decisions before I'm back, which given work etc, will be in a little over a
week. At that point I'll have a think and maybe sort out some gitlab stuff.
> > Have you or would you consider something like sourcehut?
> I've never even heard of them :(
Sourcehut's interesting because it integrates an email-based
workflow. It's still quite rough around the edges though.
That sounds very interesting, I shall spend some time reviewing Sourcehut
when I can. (Sadly likely to be quite a while away)
Ta,
D.
--
Daniel Silverstone http://www.digital-scurf.org/
PGP mail accepted and encouraged. Key Id: 3CCE BABE 206C 3B69
5:15 p.m.
On Thu, Sep 26, 2019 at 07:43:43AM +0200, Daniel Silverstone wrote:
On Wed, Sep 25, 2019 at 22:50:50 +0100, Richard Ipsum wrote:
> > If I moved to gitlab, then (a) I'd probably expect an MR type process yes,
> > since that way MRs can have CI, and (b) I'd probably be looking to make
someone
> > (you?) officially co-maintainer :D
>
> Sounds cool, I'm in :)
Excellent, handy to know. I'm currently away so don't want to make any
decisions before I'm back, which given work etc, will be in a little over a
week. At that point I'll have a think and maybe sort out some gitlab stuff.
Given the recent developments with Gitlab's TOS[1] I'd be opposed to switching
to Gitlab. Developers need to seriously wake up and consider why these services
are being offered for free[2].
For what it's worth I was able to setup basic CI for Luxio on sourcehut
running in both OpenBSD and FreeBSD in about 5 or 10 minutes.
We should consider other CI solutions, such as the one Lars is building
and if they're not ready yet, let's help make them ready instead of relying
on some corporations we likely cannot trust to have our best interests at
heart.
Thanks,
Richard
[1]: https://gitlab.com/gitlab-org/gitaly/issues/2113
[2]: https://www.devever.net/~hl/cloudflare
4 a.m.
I apologise for how long this reply has taken to come out, my week has been hectic
and I've had to learn to stem my F/LOSS work when I'm running out of energy.
Since
Gitano is fairly static at the moment, except for your excellent contributions to
luxio, it tends to be at the bottom of the list for energy :(
On Thu, Oct 24, 2019 at 22:15:26 +0100, Richard Ipsum wrote:
Given the recent developments with Gitlab's TOS[1] I'd be
opposed to switching
to Gitlab. Developers need to seriously wake up and consider why these services
are being offered for free[2].
I understand your reticence, there's been a lot of that with the recent TOS
changes. Someone's gotta pay eventually and while many people are comfortable
being the product, a lot (such as yourself at a guess, and myself to a lesser
extent) are not.
For what it's worth I was able to setup basic CI for Luxio on
sourcehut
running in both OpenBSD and FreeBSD in about 5 or 10 minutes.
That is pretty interesting. Sourcehut looks a little "boutique" but at least
they're upfront about their pricing structure which makes a little more sense.
We should consider other CI solutions, such as the one Lars is
building
and if they're not ready yet, let's help make them ready instead of relying
on some corporations we likely cannot trust to have our best interests at
heart.
My feeling is that if we wanted to move to something in the short-term which
provides CI etc, then soucehut is a good option. I think Lars' work on Ick
has been paused until he gets a lot of underlying services written better,
so we shouldn't rely on that.
On that basis I'm prepared to consider sourcehut. Would every contributor
(admittedly just you and me right now, and primarily you) need to pay for a
sourcehut account; or would it be the case that the project would pay for an
account and anyone else would then be able to contribute? If the latter then
I'm prepared to sponsor an account and consider moving luxio to it so we can
see how that feels.
If you agree that it's worthwhile, please start a fresh thread where you show
what you managed to do for the BSDs and we'll go from there.
D.
--
Daniel Silverstone http://www.digital-scurf.org/
PGP mail accepted and encouraged. Key Id: 3CCE BABE 206C 3B69
1321
days inactive
1356
days old
8 comments
2 participants
participants (2)
-
Daniel Silverstone -
Richard Ipsum