Re: [Fink-devel] fort77 can't be compiled.

2003-10-01 Thread Jeff Whitaker
On Tue, 30 Sep 2003, Pascal J.Bourguignon wrote:


 
  3.3 is not officially supported in fink yet, there will be a 10.2-gcc3.3
  distribution shortly however.  Some packages need 2.95, but that is
  handled automatically by the package. There are no packages in fink now
  that need 3.3.  The clisp package works with 3.1.

 Definitely not.  That's the only reason  why I upgraded to  3.3, to be
 able to compile clisp 2.29 from fink.



clisp 2.29 compiles just fine (and even runs) with gcc 3.1.

-Jeff

-- 
Jeffrey S. Whitaker Phone : (303)497-6313
NOAA/OAR/CDC  R/CDC1FAX   : (303)497-6449
325 BroadwayWeb   : http://www.cdc.noaa.gov/~jsw
Boulder, CO, USA 80305-3328 Office: Skaggs Research Cntr 1D-124


---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


[Fink-devel] /sw/var/lib/dpkg/status damaged?

2003-10-01 Thread Pascal Bourguignon
Now it seems the /sw/var/lib/dpkg/status is damaged. How do I recover  
it?

dpkg -i  
/sw/fink/dists/stable/crypto/binary-darwin-powerpc/openssl097- 
dev_0.9.7b-1_darwin-powerpc.deb
dpkg: parse error, in file `/sw/var/lib/dpkg/status' near line 1807  
package `openssl097-dev':
 missing version
### execution of dpkg failed, exit code 2

--
__Pascal Bourguignon__
http://www.informatimago.com/


---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


Re: [Fink-devel] /sw/var/lib/dpkg/status damaged?

2003-10-01 Thread Alexander K. Hansen
This happens occasionally--I've never heard a good explanation for it,
though.

You may be able to fix this if /sw/var/lib/dpkg/status-old is available,
and not also damaged:

sudo cp /sw/var/lib/dpkg/status-old /sw/var/lib/dpkg/status



On Wed, 2003-10-01 at 11:42, Pascal Bourguignon wrote:
 Now it seems the /sw/var/lib/dpkg/status is damaged. How do I recover  
 it?
 
 dpkg -i  
 /sw/fink/dists/stable/crypto/binary-darwin-powerpc/openssl097- 
 dev_0.9.7b-1_darwin-powerpc.deb
 dpkg: parse error, in file `/sw/var/lib/dpkg/status' near line 1807  
 package `openssl097-dev':
   missing version
 ### execution of dpkg failed, exit code 2
 



---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


Re: [Fink-devel] /sw/var/lib/dpkg/status damaged?

2003-10-01 Thread Jean-Francois MERTENS
On Wednesday, Oct 1, 2003, at 17:42 Europe/Brussels, Pascal Bourguignon  
wrote:

Now it seems the /sw/var/lib/dpkg/status is damaged. How do I recover  
it?

dpkg -i  
/sw/fink/dists/stable/crypto/binary-darwin-powerpc/openssl097- 
dev_0.9.7b-1_darwin-powerpc.deb
dpkg: parse error, in file `/sw/var/lib/dpkg/status' near line 1807  
package `openssl097-dev':
 missing version
### execution of dpkg failed, exit code 2
When this happens, status-old is in general not damaged, so the first  
fix to try is

a) diff status status-old to be sure for what other packages they  
differ, besides openssl097-dev
b) cp -p status status-svd (to be sure) ; cp -p status-old status
c) use iteratively (at least once each, since the corresponding  
file-lists are often also corrupted in
that case) fink reinstall and fink purge on each of the suspicious  
packages to bring them to their
desired state.

d) If in the above you discover that status-old was also damaged _  
there is only one remedy :
to edit it...

JF Mertens



---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


[Fink-devel] gcc3.3 tree open for business

2003-10-01 Thread David R. Morrison
Dear fink-devel,

The code freeze for the 10.2/stable tree is hereby lifted.  However, I would
like to ask all fink developers to try and keep 10.2-gcc3.3/stable in sync
with (or ahead of) 10.2/stable.  I hope that the conversion to the new
trees will be complete within a few weeks.

All fink developers are now invited to help out with the 10.2-gcc3.3 trees.
Here's how:  In fink 0.13.8 which was released yesterday, GCC 3.1 is being
enforced for the first time.  To use the revised developer tools and gcc 3.3,
(as well as the new trees), you should install fink-0.13.9-beta.  To install
this, download the source file from the sourceforge release page
  http://sourceforge.net/project/showfiles.php?group_id=17203
unpack the tarball, and run ./inject.pl from within the fink-0.13.9-beta
directory.  (Also, be sure that you have up-to-date developer tools.)

The vast majority of stable packages have already been moved from 10.2/stable
to 10.2-gcc3.3/stable.  I'll discuss some of the exceptions in a followup
message to this one.  We need help from fink developers both in testing
the 10.2-gcc3.3/stable tree, and also in getting 10.2-gcc3.3/unstable into
good shape.

Changes which you made to 10.2/unstable during the month of September might
not have gotten propagated to 10.2-gcc3.3/unstable.  So please check your
packages which might have been revised recently.

Let me remind you of what happens in the new tree.  Any package which uses
g++ or c++ is potentially affected by the ABI change in gcc between 3.1 and
3.3.  For any such package, you should be sure that the new version of the
package is labeled GCC: 3.3 (and that the previous version is labeled
GCC: 3.1).  You should also check all of the dependencies, making 
versioned dependencies on any package which is also tagged GCC: 3.3.

When a new version is made in this way, the revision number should be
increased by 10.

We had a similar conversion a year ago, when many packages were tagged
GCC: 3.1.  In the stable tree, I simply assumed that all such tags were
necessary, and I created new versions for all of those packages.  However,
many of them were probably not necessary -- checking the log files which are
generated as the package compiles is one way to see whether g++ or c++
is really being invoked.  So in the unstable tree, please feel free to
remove a GCC tag when that is appropriate.  If you remove it in
10.2-gcc3.3, you should also remove it in 10.2.

Note that simply adding or removing a GCC tag does not (in principle)
affect the .deb file, so a new revision number is not needed.  However,
we want different revision numbers when the value of the GCC tag is
different. 

I hope all of the above is clear, and I thank you in advance for your help.

  -- Dave



---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


Re: [Fink-devel] rsync vs. cvs

2003-10-01 Thread Jerry Talkington
On Wed, Oct 01, 2003 at 01:57:14PM -0400, David R. Morrison wrote:
 Hello again folks.
 
 Rob Braun has written a patch which lets us do selfupdates via rsync rather
 than cvs, and the opendarwin server is currently hosting a mirror of our
 cvs content which can be rsync'd.

Whoohoo!
 
 The simplest thing to do is to simply add this as an alternative method.
 But we were just discussing on #fink a more drastic change: we could
 scrap anonymous cvs access altogether, and offer only two access methods,
 rsync or ext cvs (the latter for fink developers only).

I personally don't have this problem, but some people might be behind a
firewall that doesn't allow for rsync.  If there isn't a technical
reason for scrapping CVS (i.e. it's incompatible with rsync,) I'd say
leave it as an option.

-- 
GPG public key:
http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0x9D5B8762


---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


[Fink-devel] what's missing in 10.2-gcc3.3/stable

2003-10-01 Thread David R. Morrison
Here's a list of what's currently missing from 10.2-gcc3.3/stable, and a
discussion of some of the reasons.

First, KDE has not yet been moved to 10.2-gcc3.3, but no problems are
anticipated. 

Second, fort77 does not compile under gcc 3.3, probably due to problems
with f2c (which probably needs some fixing).  As a consequence, a large
number of packages in sci cannot be compiled.

Third, freetype and freetype-hinting do not compile under gcc 3.3.  There
are a number of other packages which depend on these (but not a huge
number -- most packages now use freetype2).  Since freetype has given us
problems in fink in the past, I recommend that we do NOT bring this package
forward to the new tree.  Hopefully, many of the dependent packages can be
modified to use freetype2 instead.

For a small handful of packages, I got them to work in the gcc3.3 tree by
forcing the compiler to be gcc 3.1 (typically CC=gcc3 is what you need).
I haven't tried that with all of the packages below; in some case, there
may be a more up-to-date version which works with gcc 3.3.

 -- Dave

Here's the list:

crypto/dcmtk-ssl
crypto/mozilla
crypto/KDE stuff

database/gramps
database/mysql

devel/ddd

editors/abiword

games/danican
games/gnuchess
games/gnuchess-largebook
games/tornado
games/xmame

gnome/bundle-gnome
gnome/gnome-applets
gnome/shermans-aquarium

graphics/aview
graphics/filmgimp
graphics/freetype
graphics/freetype-hinting
graphics/ivtools

kde/EVERYTHING

languages/fort77
languages/gwydion-dylan
languages/gwydion-dylan-bootstrap

libs/pilot-link9

libs/perlmods/dbd-mysql-pm560
libs/perlmods/pilot-link9-perl560

net/bitchx
net/bitchx-gtk
net/gnut
net/gtkhx
net/mutella
net/nget

sci/atlas
sci/clhep
sci/dcmtk
sci/freehelix
sci/gmt
sci/grace
sci/hdf
sci/hdf5
sci/kseg
sci/ncarg
sci/netcdf
sci/numeric-atlas
sci/octave
sci/r-base
sci/scilab
sci/scilab-atlas

text/hevea
text/lilypond
text/sagasu

utils/units

x11/siag
x11/ttfmkfontdir
x11/uae
x11/wmweather
x11/xfonts-intl
x11/xv

x11-wm/blackbox-rootless
x11-wm/enlightenment
x11-wm/fvwm-ewmh
x11-wm/icewm
x11-wm/icewm-basic
x11-wm/sawfish



---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


[Fink-devel] fink binary g77 and ddd collide over /sw/lib/libiberty.a

2003-10-01 Thread Blair Zajac
Hello,

I'm getting a collision on trying to install the binary g77 and
binary ddd on a single system:

# apt-get install ddd
Reading Package Lists... Done
Building Dependency Tree... Done
The following NEW packages will be installed:
  ddd
0 packages upgraded, 1 newly installed, 0 to remove and 0  not upgraded.
Need to get 0B/7296kB of archives. After unpacking 0B will be used.
Selecting previously deselected package ddd.
(Reading database ... 44204 files and directories currently installed.)
Unpacking ddd (from .../ddd_3.3.1-4_darwin-powerpc.deb) ...
dpkg: error processing /sw/var/cache/apt/archives/ddd_3.3.1-4_darwin-powerpc.deb 
(--unpack):
 trying to overwrite `/sw/lib/libiberty.a', which is also in package g77
dpkg-deb: subprocess paste killed by signal (Broken pipe)
Errors were encountered while processing:
 /sw/var/cache/apt/archives/ddd_3.3.1-4_darwin-powerpc.deb
E: Sub-process /sw/bin/dpkg returned an error code (1)

Should libiberty be put somehow into it's own package or removed
in the install from the builds?

I'm on OS X 10.2.7 using fink CVS source builds except for these
two packages.

Best,
Blair

-- 
Blair Zajac [EMAIL PROTECTED]
Plots of your system's performance - http://www.orcaware.com/orca/


---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


Re: [Fink-devel] gcc3.3 tree open for business

2003-10-01 Thread Koen van der Drift
On Wed, 1 Oct 2003 13:19:41 -0400, David R. Morrison wrote:

 For any such package, you should be sure that the new version of the
 package is labeled GCC: 3.3 (and that the previous version is labeled
 GCC: 3.1).  You should also check all of the dependencies, making 
 versioned dependencies on any package which is also tagged GCC: 3.3.
 
 When a new version is made in this way, the revision number should be
 increased by 10.


How do I submit an (updated) package to either 10.2 or 10.2-gcc3.3 or 
both? According to the current naming policy the .info file for both 
versions will be foo.info, so the different revision numbers are not 
obvious.

thanks,

- Koen.


---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


Re: [Fink-devel] fink binary g77 and ddd collide over /sw/lib/libiberty.a

2003-10-01 Thread Blair Zajac
Jeff Whitaker wrote:
 
 Blair:  This is fixed in the newer versions of g77 (but you have to
 install from source).
 
 -Jeff

Hi Jeff,

Thanks.  I guess there are no issues with a g77 3.3 or 3.4 and
Apple's gcc 3.1 on the same system, as noted on fink's home page?

Best,
Blair

-- 
Blair Zajac [EMAIL PROTECTED]
Plots of your system's performance - http://www.orcaware.com/orca/


---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


[Fink-devel] Re: rsync vs. cvs

2003-10-01 Thread Blair Zajac
From: David R. Morrison [EMAIL PROTECTED]
 
 Hello again folks.
 
 Rob Braun has written a patch which lets us do selfupdates via rsync rather
 than cvs, and the opendarwin server is currently hosting a mirror of our
 cvs content which can be rsync'd.
 
 The simplest thing to do is to simply add this as an alternative method.
 But we were just discussing on #fink a more drastic change: we could
 scrap anonymous cvs access altogether, and offer only two access methods,
 rsync or ext cvs (the latter for fink developers only).

How about using Subversion?  There's already a fink package for it.  It's
designed to be the successor to CVS:

http://subversion.tigris.org/

I would think that a Subversion server would be less resource intensive on
the server than an rsync server, especially if people use rsync -c.

I'm a svn committer, so I'm just a little biased here :)

Best,
Blair

-- 
Blair Zajac [EMAIL PROTECTED]
Plots of your system's performance - http://www.orcaware.com/orca/


---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel


Re: [Fink-devel] Re: rsync vs. cvs

2003-10-01 Thread Charles Lepple
On Thursday, October 2, 2003, at 12:15  AM, Blair Zajac wrote:

I would think that a Subversion server would be less resource 
intensive on
the server than an rsync server, especially if people use rsync -c.
It takes a certain amount of paranoia to use rsync -c on a regular 
basis :-)

In the most common case, a 'fink selfupdate' would not be very 
resource-intensive as compared to CVS-- when both sides build their 
file lists, and realize that the vast majority of files are the same, 
there won't be much time spent checksumming. I don't know the specs of 
opendarwin, but I would imagine that the processor is idle more often 
than the network link is.

As for Subversion, here's my two cents' worth. At work, I looked at 
migrating our collection of RCS and CVS repositories to Subversion. I 
encountered very little trouble in converting a CVS-based project to 
svn, but it was fairly small, and scheduling a few minutes of downtime 
to upgrade to the next major version of svn would not be a big thing.

One time, when I couldn't get to the CVS server, I snagged the 
repository tarball. To use scientific terms, it was freakin' huge. I 
hope that was just because of the old file naming scheme, but before 
anyone decides to use Subversion to host the Fink info file collection, 
they should consider the effort involved in handling svn dumps of that 
magnitude.

--
Charles Lepple [EMAIL PROTECTED]
http://www.ghz.cc/charles/


---
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel