[9fans] OT: plan9port compilation failure

[9fans] OT: plan9port compilation failure

[ISHWAR RATTAN]

Did a fresh cvs checkout and tried to compile.

System is (via uname -a):
SunOS cps222 5.10 Generic_118833-03 sun4u sparc SUNW,Sun-Blade-100

In plan9 directory (via ./INSTALL):

...
>>> cd /home/user/rattan/plan9/src/libthread; mk all
9c  -I. pthread.c
sparc-ucontext.h:20: error: conflicting types for 'makecontext'
/usr/include/ucontext.h:43: error: previous declaration of 'makecontext' was here
sparc-ucontext.h:20: error: conflicting types for 'makecontext'
/usr/include/ucontext.h:43: error: previous declaration of 'makecontext' was here
sparc-ucontext.h:21: error: conflicting types for 'swapcontext'
/usr/include/ucontext.h:42: error: previous declaration of 'swapcontext' was here
sparc-ucontext.h:21: error: conflicting types for 'swapcontext'
/usr/include/ucontext.h:42: error: previous declaration of 'swapcontext' was here
mk: 9c  -I. pthread.c  : exit status=exit(1)
mk: for i in ...  : exit status=exit(1)
$
----

Any ideas?
-ishwar

Re: [9fans] OT: plan9port compilation failure

[Richard Bilson]

On 7/6/06, ISHWAR RATTAN <rattan@cps.cmich.edu> wrote:
>
> Did a fresh cvs checkout and tried to compile.
>
> System is (via uname -a):
> SunOS cps222 5.10 Generic_118833-03 sun4u sparc SUNW,Sun-Blade-100
>
> In plan9 directory (via ./INSTALL):
>
> ...
> >>> cd /home/user/rattan/plan9/src/libthread; mk all
> 9c  -I. pthread.c
> sparc-ucontext.h:20: error: conflicting types for 'makecontext'
> /usr/include/ucontext.h:43: error: previous declaration of 'makecontext' was here

There's a block of code that is #ifdef __sun__ in threadimpl.h that
breaks Solaris. I was just in the process of preparing a diff when I
was distracted by other things.

In particular, this block is unnecessary and harmful:

#if defined(__sun__)
#	define mcontext libthread_mcontext
#	define mcontext_t libthread_mcontext_t
#	define ucontext libthread_ucontext
#	define ucontext_t libthread_ucontext_t
#	include "sparc-ucontext.h"
#endif

Having eliminated that, however, it's working well for me.

Re: [9fans] OT: plan9port compilation failure

[ISHWAR RATTAN]

On Thu, 6 Jul 2006, Richard Bilson wrote:

> There's a block of code that is #ifdef __sun__ in threadimpl.h that
> breaks Solaris. I was just in the process of preparing a diff when I
> was distracted by other things.
>
> In particular, this block is unnecessary and harmful:
>
> #if defined(__sun__)
> #	define mcontext libthread_mcontext
> #	define mcontext_t libthread_mcontext_t
> #	define ucontext libthread_ucontext
> #	define ucontext_t libthread_ucontext_t
> #	include "sparc-ucontext.h"
> #endif
>

After commenting out this code in src/libthread/thread_impl.h, I was able
to a little further now the compilation stopped at:
...
9l -o o.readcons readcons.o
9l -o o.resample resample.o
ld: fatal: library -lmemlayer: not found
ld: fatal: library -lmemdraw: not found
ld: fatal: File processing errors. No output written to o.resample
collect2: ld returned 1 exit status
mk: 9l -o o.resample ...  : exit status=exit(1)
mk: for i in ...  : exit status=exit(1)
$

---
-ishwar

Re: [9fans] OT: plan9port compilation failure

[William Josephson]

On Thu, Jul 06, 2006 at 11:27:07PM -0400, ISHWAR RATTAN wrote:
> After commenting out this code in src/libthread/thread_impl.h, I was able
> to a little further now the compilation stopped at:
> ...
> 9l -o o.readcons readcons.o
> 9l -o o.resample resample.o
> ld: fatal: library -lmemlayer: not found
> ld: fatal: library -lmemdraw: not found
> ld: fatal: File processing errors. No output written to o.resample
> collect2: ld returned 1 exit status
> mk: 9l -o o.resample ...  : exit status=exit(1)
> mk: for i in ...  : exit status=exit(1)
> $

You might try debugging it yourself.
I'd try on one of our niagra boxes
here, but the fellow across the
office from me is still swearing and
trying to get a compiler installed
on it, so it'll have to wait.

Re: [9fans] OT: plan9port compilation failure

[Russ Cox]

If you want Solaris, you're on your own.
I know that libtask works on Solaris 5.(10?)
but I haven't tried libthread recently.

For me, Solaris is all frustration and little reward.
The test machines I have are very slow, and it's too
different and backwards.

If someone else sends me fixes I'll be glad to believe
you and apply them.  Or send them to wkj and maybe
he can actually test them.

I'm sorry for the lack of Solaris support, but there
is only so much time in the world, and Solaris doesn't
deserve mine.

Russ

Re: [9fans] OT: plan9port compilation failure

[Axel Belinfante]

just curious what solaris versions and/or compilers
people are (trying to) build/use(ing) plan9port on/with,
and which experiences they have.

For example, in the past we have noticed differences between
sunos 5.8 and (I think) 5.10 regarding thread stuff.

I run it on sunos 5.8 and build it with sun cc (Sun WorkShop 6 C 5.1).
For the last year and a half I've been using (daily, including 9term)
a plan9port that I built early january 2005.

I've just rebuild plan9port from cvs (sunos 5.8, sun cc).
I needed minor tweaks (like the threadimpl.h thing reported here, thx)
I've not used much of it yet, some things (e.g. acme) do work, some not.

Things I encountered so far (and will look into):
 - ps aborts with:    sort: note: sys: child
   (I've seen that before and then ignored the problem,
	maybe there is a deeper problem to solve)
 - 9term seems not to give me a shell (does not print a prompt)
 - somehow I broke plumbing images (via acme, or directly via plumb)
 - b3 down in 9term and then moving the mouse (b3 still pressed)
   shows the outline of rio's b3 menu; this continues when
   moving the mouse outside the 9term window; rio b3 menu appears
   only when releasing b3
 - I'm not sure all tests in libthread run successfully;
   what 'effects' are to be expected?

Axel.

Re: [9fans] OT: plan9port compilation failure

[Roman Shaposhnik]

On Mon, 2006-07-10 at 17:26 +0200, Axel Belinfante wrote:
> just curious what solaris versions and/or compilers
> people are (trying to) build/use(ing) plan9port on/with,
> and which experiences they have.

  Sun Studio 11/Solaris 9.

> For example, in the past we have noticed differences between
> sunos 5.8 and (I think) 5.10 regarding thread stuff.

  The differences are glaring NxM model vs. 1-1 for one.

> I run it on sunos 5.8 and build it with sun cc (Sun WorkShop 6 C 5.1).

  Now, that is a pretty old piece of software. Is there any reason
you don't want to upgrade ?

> For the last year and a half I've been using (daily, including 9term)
> a plan9port that I built early january 2005.
>
> I've just rebuild plan9port from cvs (sunos 5.8, sun cc).
> I needed minor tweaks (like the threadimpl.h thing reported here, thx)
> I've not used much of it yet, some things (e.g. acme) do work, some not.

  Hm. It kinda works for me. At least things I care about.

Thanks,
Roman.

Re: [9fans] OT: plan9port compilation failure

[Axel Belinfante]

> > I run it on sunos 5.8 and build it with sun cc (Sun WorkShop 6 C 5.1).
>
>   Now, that is a pretty old piece of software. Is there any reason
> you don't want to upgrade ?

the sunos or the sun cc?

I guess the sun cc was not upgraded because nobody really cared enough.
for the usual c stuff it is ok (apart from those mis(sing) features
of which you will see the consequences in the plan9port cvs logs),
and those really caring here have probably been using gcc anyway.

the number of suns here is pretty low now, I'm not sure upgrades
would be considered worthwhile.
most people switched to pc's with linux or windows, or occasional mac.
this also changed how interchangeble the machines are.
on the suns we only have the bare os locally, and all home dirs
and (non-default) applications come from the fileservers.
for the linux machines most applications are local too,
which (potentially) makes them all different. that's not all.
with the suns we had usually a single os version on all machines,
or at most two at times of change.  I think we have at least three
versions of linux with some consequential shared lib incompatibilities.

> > For the last year and a half I've been using (daily, including 9term)
> > a plan9port that I built early january 2005.
> >
> > I've just rebuild plan9port from cvs (sunos 5.8, sun cc).
> > I needed minor tweaks (like the threadimpl.h thing reported here, thx)
> > I've not used much of it yet, some things (e.g. acme) do work, some not.
>
>   Hm. It kinda works for me. At least things I care about.

same here now.

however, I do have the feeling that typing in the acme tags is now even
slower than it already was. could this be a consequence of devdraw,
or other changes? (my previous acme version was from early 2005).
Also (but this is not new, I think?) in acme a single click to scroll
sometimes results in multiple scrollings.
when scrolling with b2 pressed the scrolling cannot keep up with
the mouse moving and you see the mouse cursor being warped up or down
into the scroll 'elevator' and (because of that?) the elevator jumps
up and down and up while scrolling on its way up, or overshoots and
scrolls back and forth until it is finally stable.
this all on a sun blade 100 - maybe that is too slow a machine now?

anyway, these are just little nits.
I agree that on solaris it works pretty well,
definitely those things I care about, and that's pretty impressive
(especially impressive when considering what russ recently wrote).

Axel.