Fwd: [9fans] Re: request: native graphics on OS X

Fwd: [9fans] Re: request: native graphics on OS X

[David Leimbach]

Meant for this to go to 9fans.

---------- Forwarded message ----------
From: David Leimbach <leimy2k@gmail.com>
Date: Jan 10, 2007 11:51 AM
Subject: Re: [9fans] Re: request: native graphics on OS X
To: Skip Tavakkolian <skip@9netics.com>


On 1/10/07, David Leimbach <leimy2k@gmail.com> wrote:
> On 1/10/07, David Leimbach <leimy2k@gmail.com> wrote:
> > On 1/10/07, Skip Tavakkolian <skip@9netics.com> wrote:
> > > > ranlib: file: ../libmachdep.a(md5block.o) has no symbols
> > > > ranlib: file: ../libmachdep.a(sha1block.o) has no symbols
> > >
> > > for your machine it should have picked posix-386 as machinedep.
> > > was it able to build that library?
> > >
> > >
> >
> > It did pick the 386 directory, and produced no symbols.
> >
> > sha1block.s is an empty file, same with md5block.s
> >
> > gcc -E uses the .spp files to produce these and comes up with nothing
> > in the output.
> >
> > Dave
> >
> And the reason it has no output is because gcc on Apple thinks that
> .spp or .sp files are linker input files, which will be ignored due to
> the -E option, since linking isn't done.
>
> The following makefile change to the posix-386 directory fixes it:
>
> Index: Makefile
> ===================================================================
> RCS file: /cvs/drawterm/posix-386/Makefile,v
> retrieving revision 1.6
> diff -u -r1.6 Makefile
> --- Makefile    9 Jan 2007 22:17:21 -0000       1.6
> +++ Makefile    10 Jan 2007 19:40:37 -0000
> @@ -20,8 +20,8 @@
>         $(AS) -o $*.$O $*.s
>
>  md5block.s: md5block.spp
> -       gcc -E md5block.spp >md5block.s
> +       cat md5block.spp | gcc -E - >md5block.s
>
>  sha1block.s: sha1block.spp
> -       gcc -E sha1block.spp >sha1block.s
> +       cat sha1block.spp | gcc -E - >sha1block.s
>

Ok this patch seems to get me 100% compiled :-)  I had to replace
__Darwin__ with __APPLE__ because gcc doesn't set __Darwin__ anymore.

And even after that, it behaves just as Andrey's binary did, it just
gives me the "spinny beachball of doom" on Mac OS X that causes me to
force quit.

I suppose I'll have to break out the debug tools sometime.

>8------------------------------->8

Index: Makefile
===================================================================
RCS file: /cvs/drawterm/posix-386/Makefile,v
retrieving revision 1.6
diff -u -r1.6 Makefile
--- Makefile    9 Jan 2007 22:17:21 -0000       1.6
+++ Makefile    10 Jan 2007 19:48:58 -0000
@@ -20,8 +20,8 @@
        $(AS) -o $*.$O $*.s

 md5block.s: md5block.spp
-       gcc -E md5block.spp >md5block.s
+       cat md5block.spp | gcc -E - >md5block.s

 sha1block.s: sha1block.spp
-       gcc -E sha1block.spp >sha1block.s
+       cat sha1block.spp | gcc -E - >sha1block.s

Index: md5block.spp
===================================================================
RCS file: /cvs/drawterm/posix-386/md5block.spp,v
retrieving revision 1.2
diff -u -r1.2 md5block.spp
--- md5block.spp        9 Jan 2007 22:17:21 -0000       1.2
+++ md5block.spp        10 Jan 2007 19:48:58 -0000
@@ -116,7 +116,7 @@
        .text

        .p2align 2,0x90
-#ifdef __Darwin__
+#ifdef __APPLE__
        .globl __md5block
        __md5block:
 #else
Index: sha1block.spp
===================================================================
RCS file: /cvs/drawterm/posix-386/sha1block.spp,v
retrieving revision 1.2
diff -u -r1.2 sha1block.spp
--- sha1block.spp       9 Jan 2007 22:17:21 -0000       1.2
+++ sha1block.spp       10 Jan 2007 19:48:58 -0000
@@ -1,7 +1,7 @@
 .text

 .p2align 2,0x90
-#ifdef __Darwin__
+#ifdef __APPLE__
 .globl __sha1block
 __sha1block:
 #else

8<-----------------8<

Re: [9fans] Re: request: native graphics on OS X

[andrey mirtchovski]

new binaries on 9grid fix a problem with the cat-tongued ibm mouse we
all love and fixes del keys to send interrupts.

russ has the diff so cvs is probably updated already.

http://9grid.net/andrey/p9/drawterm-osx-intel.gz
http://9grid.net/andrey/p9/drawterm-osx-power.gz

russ, you're welcome to put them on swtch.com if you want.

Re: [9fans] Re: request: native graphics on OS X

[David Leimbach]

Are these from the same CVS?

On 1/11/07, andrey mirtchovski <mirtchovski@gmail.com> wrote:
> new binaries on 9grid fix a problem with the cat-tongued ibm mouse we
> all love and fixes del keys to send interrupts.
>
> russ has the diff so cvs is probably updated already.
>
> http://9grid.net/andrey/p9/drawterm-osx-intel.gz
> http://9grid.net/andrey/p9/drawterm-osx-power.gz
>
> russ, you're welcome to put them on swtch.com if you want.
>

Re: [9fans] Re: request: native graphics on OS X

[andrey mirtchovski]

yes, but the patch hasn't made it there yet (i just checked).
everything else is the same.

On 1/11/07, David Leimbach <leimy2k@gmail.com> wrote:
> Are these from the same CVS?
>

Re: [9fans] Re: request: native graphics on OS X

[David Leimbach]

I just found out my spinny beachball of doom goes away if I run as
root vs running as a normal user... I wonder if my filesystem
permissions are dorked up.  I ran a ktrace on this and saw a ton of
syscalls to sigprocmask...

Dave

On 1/11/07, David Leimbach <leimy2k@gmail.com> wrote:
> Are these from the same CVS?
>
> On 1/11/07, andrey mirtchovski <mirtchovski@gmail.com> wrote:
> > new binaries on 9grid fix a problem with the cat-tongued ibm mouse we
> > all love and fixes del keys to send interrupts.
> >
> > russ has the diff so cvs is probably updated already.
> >
> > http://9grid.net/andrey/p9/drawterm-osx-intel.gz
> > http://9grid.net/andrey/p9/drawterm-osx-power.gz
> >
> > russ, you're welcome to put them on swtch.com if you want.
> >
>