ASPN ActiveState Programmer Network
ActiveState
/ Home / Perl / PHP / Python / Tcl / XSLT /
/ Safari / My ASPN /
Cookbooks | Documentation | Mailing Lists | Modules | News Feeds | Products | User Groups


Recent Messages
List Archives
About the List
List Leaders
Subscription Options

View Subscriptions
Help

View by Topic
ActiveState
.NET Framework
Open Source
Perl
PHP
Python
Tcl
Web Services
XML & XSLT

View by Category
Database
General
SOAP
System Administration
Tools
User Interfaces
Web Programming
XML Programming


MyASPN >> Mail Archive >> tcl-mac
tcl-mac
Re: [MACTCL] "proper" arrangement for packages?
by Jon Guyer other posts by this author
Jul 9 2007 9:57AM messages near this date
Re: [MACTCL] "proper" arrangement for packages? | Re: [MACTCL] "proper" arrangement for packages?
On Jul 9, 2007, at 9:19 AM, Daniel A. Steffen wrote:

>  FWIW on my 10.4.10 intel box, the system TclAE binary is universal,
>  maybe the user in question has run thinning software? (probably to
>  "optimize" his system by getting rid of all that useless ppc code...)

That's my suspicion, too.

>  There really isn't very much you need to know to use it, it's just
>  standard configure & make & make install.
>  Building universal is purposefully exactly the same as for the tcl/tk
>  core, as documented e.g. in the tcl/macosx/README, c.f. below.

Right, but...

>  TclAE is the only extension that was ever built as a bundle I think,

Ahah.


>  Now that none of these issues are relevant any longer, I don't think
>  the bundle format makes any sense for tcl extensions going forward,
>  it unnecessary complicates the extension building process, and even
>  for OSX-only extensions like TclAE, it makes building them
>  unnecessarily different from the standard TEA process.

Fine by me.


>  The only reason Apple ships TclAE as a bundle is that I did not have
>  time to move it to a TEA buildsystem before I had to complete the
>  TclBI submission for Tiger...

And have you learned your lesson and stopped frittering your life  
away on that dissertation?

> 
> > As far as I can tell, TEA doesn't provide any help for generating
> > either, is that correct?
> 
>  No, TEA in fact does everything needed, I think you're just missing
>  the last step in the process: it is 'make install' that packages
>  everything up in the right arrangement at the install location

I did 'make install' and it scattered things all over hell and gone.  
Not very useful for a binary distro...

>  (or an
>  alternate location if a DESTDIR arg is passed to make).

OK. I don't know about that.



>  The following steps will build & install universal TclAE from the
>  repository (if universal tcl is installed):
>  	$ svn co http://tclae.svn.sourceforge.net/svnroot/tclae/trunk tclae
>  	$ mkdir -p build; cd build
>  	$ export CFLAGS='-Wno-deprecated-declarations -mmacosx-version-
>  min=10.2 -arch ppc -arch i386'
>  	$ ../tclae/configure --libdir=/Library/Tcl
>  	$ make -j 2
>  	$ sudo make install
>  or to install in an alternate directory, e.g. for packaging a
>  distribution
>  	$ make install DESTDIR=`pwd`/../install/

OK, I'll try this tonight. Thanks. Is this, like, you know, written  
down anywhere, or anything?


>  and to run the testsuite (no failures on i386 or ppc for me)
>  	$ make test

Good. I don't get any failures, either.


>  I have just committed an update to svn that fixes running the
>  testsuite from outside the source dir (as above), handles 64-bit arch
>  flags in CFLAGS and updates to the latest TEA files from CVS.

Ah, good. That was going to be my next question, as I had to remove  
the two 64-bit -arch flags before I could build.

>  It would probably be good to release this as 2.0.2 (or 2.0.3 ?) in
>  source and universal binary form?

I think it's still 2.0.2, but yes, I agree. Any idea if there's still  
any prospect of getting this into Leopard? What Apple presently ships  
is totally unusable on Intel machines.


>  or package
>  things up manually if you prefer:

I have absolutely no desire to package things up manually. It just  
looked to me like I had to.

>  FYI, running the tclAE testsuite with this binary against tcl HEAD
>  errors out immediately with 'alloc: invalid block: 0x4140f0: 0 0',
>  haven't looked into the cause for that (the same happens with a tclAE
>  binary built against tcl HEAD FWIW).

Hmmm... that's not good. I don't suppose that you're relishing the  
prospect of debugging this, by any chance? I would need to (fail to)  
check out HEAD and then (fail to) build it before I could even start  
to (fail to) debug the problem.

>  given Kevin's reply, I think this is a completely separate issue,

I agree. I had originally included that because I was getting  
symptoms very much more like what Kevin was seeing (for the record,  
my original message contained significantly more profanity), but by  
the time I posted, that was a red herring. Should have clipped it.


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Tcl-mac mailing list
tcl-mac@[...].net
https://lists.sourceforge.net/lists/listinfo/tcl-mac
Thread:
Kevin Walzer
Bernard Desgraupes
Kevin Walzer
Bernard Desgraupes
Kevin Walzer
Jon Guyer
Daniel A. Steffen
Jon Guyer
Kevin Walzer
Jon Guyer
Kevin Walzer

Privacy Policy | Email Opt-out | Feedback | Syndication
© 2004 ActiveState, a division of Sophos All rights reserved