Fink 0.5.0a (Final) released
 

Yes, good news! And includes over 700 binary packages for OS X 10.2 as well as over 1800 source packages of all kinds.

And remember, this release is for Mac OS X 10.2 only.


Cheers...

Install issues: selfupdate attempts to install 0.4.1a
 

I was waiting all day to come home to "selfupdate" ;)

Anyway, I ran into the following issue on my TiBook:

[marley:~] % fink -version
Package manager version: 0.9.1
Distribution version: 0.3.0

The Fink Upgrade Matrix (http://fink.sourceforge.net/download/upgrade.php) tells me to:
"Fink official binary distribution, version 0.3.x or later:
Update using "fink selfupdate". If you have an install from 10.1, you may need to do so twice to be fully updated."

So, I run:
[marley:~] % sudo fink selfupdate
Password:
rm -f LATEST-FINK
curl -L -s -S -o LATEST-FINK http://fink.sourceforge.net/LATEST-FINK

A new Fink distribution release is available. I will now download the package
descriptions for Fink 0.4.1a and update the core packages. After that, you
should update the other packages using commands like 'fink update-all'.

gzip -dc packages-0.4.1a.tar.gz | tar -xf -

gzip: packages-0.4.1a.tar.gz: not in gzip format
./inject.pl /sw
Can't exec "./inject.pl": No such file or directory at /sw/lib/perl5/Fink/Services.pm line 180.
### ./inject.pl failed, exit code -1
Failed: injecting the new package definitions from packages-0.4.1a.tar.gz failed

So, the "selfupdate" command is trying to download and install release 0.4.1a!! Anyone else see similar issues?

That's a very old version. So probably it's allright if first it upgrades to 0.4.1a , as the message said: "you may need to do so twice to be fully updated"

What's your 'sw_vers'? Remember that this release is for 10.2 only.

It seems it fails to inject the package,because of a missing file.

Maybe you post the error message at the Fink-users list.


Cheers...

[marley:~] cschorr% sw_vers
ProductName: Mac OS X
ProductVersion: 10.2.2
BuildVersion: 6F21

I was assuming that the Distribution Version (0.3.0) was the version that is referenced in Fink's upgrade matrix. They don't differentiate between distro version and "package manager version"

LizardKing,

This is what I get *before* the upgrade.

Do you have by chance 'tetex' installed ?


Cheers...

no, I never installed tetex (quickly verified by running 'fink list' and checking installed packages).

The thing that is interesting to me is the link in this command (which is run during selfupdate):
curl -L -s -S -o LATEST-FINK http://fink.sourceforge.net/LATEST-FINK

If you click on the above link it will return a page with the text '0.4.1a' (at least it is as of 11:38PM EST).

curious....

LizardKing,

Well, it might be wrong or in your case it might be normal that it first installs 0.4.1a, as your former version is quite old.

But, I don't know why it fails to inject the package, as this last upgrade matrix method is new to me.

As I said, it would be interesting that you post the error messages you got at the fink-users list, together with your system information. And please, let us know what the fink developers say about your problem.


Cheers...

I am also have the problem where I am getting the message

You already have the package descriptions from the latest Fink point release.
(installed:0.4.1a available:0.4.1a)

This is definitely not true.... or we wouldnt have a thread titled Fink .0.50a relased now would we?

FWIW,

I'm doing the usual 'fink selfupdate-cvs' followed by 'fink update-all' and it's churning away on the update-all at the moment. Seems all is working well so I believe I will have the correct installation when it's done.

....

Okay, this is what I've got now:

fink -V
Package manager version: 0.11.1
Distribution version: 0.5.0a.cvs

This must be the new number, I assume.

existing software
 

I previously installed mysql (from entropy.ch I think). Any way to "convince" fink that I already have mysql installed so that it will be automagically updated?

Thanks

For new users they can download the Fink 0.5.0a Binary Installer - (11 MB) from here:

http://fink.sourceforge.net/download/index.php

Remember, 10.1 users - they should use Fink 0.4.1


I finally upgraded, but ran 'fink selfupdate-cvs', as I prefer to get packages from the cvs repository, so the distribution version now is:

Cheers...

ThAnX
 

Sao
Please let me thank you at this time for all the help you gave me via your
superbe posts concerning Fink before the 0.5.0a release. I have certainly
learned more about the labyrinth which is Unix and about compiling than I have
during all my years as a Debian and BSD user. I was very happy to note that when
I typed "fink --version" in a terminal, I was presented with a "0.5.0a" as
reply,
Last login: Tue Dec 10 07:18:51 on console
Welcome to Darwin!
[Kaos:~] kaos% fink --version
Package manager version: 0.11.1
Distribution version: 0.5.0a.cvs

so for me the release is a moot question. I am glad I did it the hard way
before, not only because of what I have learned through trial and error and on
these boards, but because I now have a system that compares favoribly with the
the most well configured Linux systems I have used.

PS: I must have spent days and days reading your posts in an historical, if not
hysterical, fashion. You could compie a book from this marvelous chronology of
help that you have shared with others!

Cheers and Thanks again!:D

Glanz,

Thanks for your kind words. :)

I enjoy posting about Fink very much but all the praise should be going to the Fink developers for the tremendous work done. As I read in a post at the fink-beginners list by Daniel Lord:

<<"It is more than the Source Forge Product of the Month:
for I cannot describe in one message the vastness of body of software fink opens to the OS X user.
In sum, it makes the Mac simply the most versatile, powerful desktop Unix on the market, bar none.

Thank you.">>

And I agree 100% with that.


Cheers...

Maybe somebody knows how to rebuild the package list? I updated to 0.50a.cvs, but when I try to install some packages (i.e. r-base) i get this message:

Failed: no package found for specification 'r-base'!

even though the package is in stable por 0.5.0a. My guess is that during the update process my package list was somehow corrupted..Anybody know what to do about this?

ateles

Arggggh...will it never end (insert hair-pulling)? Took me a while to finally get my XDarwin to start displaying the cursors properly again (e.g. with the proper masks so that you can see a black cursor on a black window or background). Never sure what I did to fix it either. Now after updating to 0.50a last night:

fink selfupdate-cvs

I am back to having the cursor screwed - no mask.

Any help appreciated! Thx.

INFO
------
Package manager version: 0.11.1
Distribution version: 0.5.0a.cvs

% fink install blackbox
sudo /sw/bin/fink install blackbox
Password:
Information about 792 packages read in 2 seconds.

pkg blackbox version ###
Failed: no version info available for blackbox

I really wish I knew what I was doing... IceWM installed just fine and is now set as my default (for now.)

% fink --version
Package manager version: 0.11.1
Distribution version: 0.5.0a.cvs

% sw_vers
ProductName: Mac OS X
ProductVersion: 10.2.2
BuildVersion: 6F21

ateles,

Check if you have it at:

/sw/fink/10.2/stable/main/finkinfo/sci/r-base-1.6.1-2.info

and can you show me the result of:

ls -l /sw/fink/dists


Maybe try running:

'fink fetch-misssing'

and then

'fink scanpackages'


Cheers...

A masterpiece of compilation!!!
 

I also agree! One more thing. After twisting and turning a few applications so that they would compile on Darwin 6.2, such as Mozilla before I even knew that Fink had already done it, I can appreciate the amount of work that went into this. That isn't the first time I have re-invented the wheel, nor will it be the last.

Fink is a masterpiece. It even works better than most Linux distros, except for Gentoo because there is no optimization lag. I plan to help the project in any way I can, mostly by finding glitches and their exact causes. I do not feel I am competent enough to do much more than that. The competence of all those who brought this together surprises and encourages me.

I'll test anything! After all, I wouldn't want to use something more stable than I am myself. So it's the unstable tree for meee. I started testing Debian when it first came out, and I'm still testing it;). I don't believe I have ever used a "stable" version of it either.

justinp, what does 'fink info blackbox' give you? Here's what it gives me:

fink info blackbox
Information about 1860 packages read in 2 seconds.

pkg blackbox version ###
pkg blackbox version 0.65.0-1

blackbox-0.65.0-1: Alternate Window Manager, basic version
Blackbox is yet another addition to the list of window
managers for the Open Group's X Window System, Version 11 Release 6
and above. Blackbox is built with C++, sharing no common code with
any other window manager (even though the graphics implementation
is similar to that of Window Maker).
.
Web site: http://blackboxwm.sourceforge.net/
.
Maintainer: Christian Swinehart <cswinehart@users.sourceforge.net>

Justnip,

Same as ateles, could you show me the result of:

'ls -l /sw/fink/dists'

and also of:

'fink list --section=x11-wm'

Check if you have it at:

/sw/fink/10.2/stable/main/finkinfo/x11-wm/...


Cheers...

Bluehz,

I remember the black cursor problem, the fix then, was to update to XFree86- 4.2.1.1

I don't know why you got it again.

Are you using OroborOSX ?

Maybe you try:

'fink rebuild xfree86-base'
'fink rebuild xfree86-rootless'


Cheers...

sao:

% ls -l /sw/fink/dists
lrwxr-xr-x 1 justinp admin 13 Dec 9 17:42 /sw/fink/dists -> /sw/fink/10.2

% fink list --section=x11-wm
Information about 792 packages read in 1 seconds.

blackbox-rootle 0.62.1-11 Alternate Window Manager, works in XDarwi...
enlightenment 0.16.5-8 Themeable, featureful window manager
fvwm 1.24r-5 X11 window manager with large virtual des...
fvwm-common 2.4.11-1 Common files for versions 1 and 2.xx of fvwm
fvwm-common-ewm 0.7-4 Common files for versions 1 and 2.xx of fvwm
fvwm-common-no- 2.4.11-1 Common files for versions 1 and 2.xx of fvwm
fvwm-ewmh 0.7-4 X11 win. man. w/ lrge virtual desktop, ad...
fvwm-icons 2.4.0-5 Icons for fvwm
fvwm2 2.4.11-1 X11 window manager with large virtual des...
fvwm2-no-gnome 2.4.11-1 X11 wind. man., large virtual desktop, v2...
i icewm 1.2.2-1 X11 window manager - with GNOME support.
icewm-basic 1.2.2-1 X11 window manager - no GNOME support.
sawfish 1.1-2 Themeable and scriptable window manager

And no, it isn't in /sw/fink/10.2/stable/main/finkinfo/x11-wm/... (the rootless version is though.)

thatch:

% fink info blackbox
Information about 792 packages read in 1 seconds.

pkg blackbox version ###
Failed: no version info available for blackbox

IIRC, the last time someone had so few packages in their list, it turned out they didn't have their trees configured correctly in /sw/etc/fink.conf

Mine is exactly the same except for the Trees: and the FetchAltDir

Trees: local/main stable/main stable/crypto local/bootstrap
Distribution: 10.2
FetchAltDir: /swbackup/src

Does this mean I need to edit that file so that the Trees: line reads the same as yours?

Well, sao is really the fink expert around here and so I defer to him. But I think having the unstable tree is a good thing and it wouldn't hurt to add it and see what happens. The unstable tree is for the cutting edge versions that are to be tested before they are made into stable.

As far as the FetchAltDir: /swbackup/src ...
I don't have that directory on my rig or the listing in my fink.conf. So, I don't know about that.

I wonder, did you do a clean install of fink after you upgraded to 10.2? Or did you upgrade fink after an archive and install? It might be good to read up on the sourceforge site about these ways.

I did a clean install of 10.2 and then waited for Fink 0.5.0a to come out before installing Fink again.

I installed Fink with the installer at the Fink sourceforge site, then installed xfree86-base and -rootless with Fink, and installed and updated the XDarwin app from the XonX sourceforge site.

Anyway, I guess I'll wait for sao's reply before I change anything.

Looks like sao might have signed off for the night as it's quite late in Singapore.

I don't remember ever having to update XDarwin.app separately from Xfree86. Anyway, my XDarwin.app version is at 1.1.1.1. What do you get when you:?

% fink list --section=x11-system
Information about 792 packages read in 1 seconds.

system-xfree86 4.2-1 Placeholder package for manually installe...
i xfree86-base 4.2.1.1-3 XFree86 libraries, utilities, clients and...
i xfree86-base-sh 4.2.1.1-3 XFree86 libraries, utilities, clients and...
i xfree86-rootles 4.2.1.1-3 XFree86 libraries, utilities, clients and...
i xfree86-rootles 4.2.1.1-3 XFree86 libraries, utilities, clients and...

justinp, I've been searching the forum for the place I recently saw something like what you're experiencing, not many packages coming up. Here is a link to that thread. You may have to scroll a bit to find the relevant 'off' topic on page 2, I believe.

If you have a good backup system whereby you can test anything you want and fairly quickly get back to your previous cloned backup, then if I were you, I would add those unstable trees and do a 'fink selfupdate-cvs' followed by a 'fink update-all'. It will take some time but I think it will fix you up.

Unstable
 

It is indeed a good thing! One way to have access to ALL the trees, after having entered them in the config file, is to use Fink Commander. The new version lets the user click on the "tree" from which he wishes to download. For example, one will not find Mozilla in the stable, so if you want a browser that works within Fink, one will have to go with unstable because Good ol' Mozilla is continuously in the "unstable" mode, albeit by name. Someone once said that "Beta" is a state of mind. Well...., one could say that "Unstable" is also a state on mind... L:p L...... especially mine.:mad:

sao,

Here is a summary of what I learned from the fink-users list regarding my issue (as described above) and what I eventually had to do to fix fink on my system. After posting a similar description of my problem, I received the following suggestions:

"I'm not 100% sure the Upgrade Matrix is appropriate here, since there's
an OS version change, too (fink 0.9.1 was for 10.1). Try following the
Jaguar update instructions as per

http://fink.sourceforge.net/news/jaguar.php"

"I may be wrong, but for a long time already the "Upgrade Matrix" has
been basically wrong and useless: You cannot upgrade automatically from
anything earlier than 0.4.1a. I would suspect that this is still the
case.

[]
> gzip -dc packages-0.4.1a.tar.gz | tar -xf -
> > gzip: packages-0.4.1a.tar.gz: not in gzip format

This is because your old fink version tries to download from a
non-existing URL. The file downloaded under the name
"packages-0.4.1a.tar.gz" is probably just an error message in html
format.

You will be better off by doing a complete fresh installation from
scratch."


I tried the first suggestion by following the instructions on http://fink.sourceforge.net/news/jaguar.php. This didn't work properly - the upgrade kept failing. At this point, I was frustrated with trying to upgrade my old version, so I removed it and installed 0.5.0a from scratch (off of the binaries).

You may have noticed that since my original post in this discussion thread, the upgrade matrice's instructions have changed for folks with version 0.3.X.

Thanks for helping out!

OK, blackbox is up and running. Adding the unstable lines to the file and then updating did the trick (shouldn't I have initially been given an option to use the unstable trees?)

OK, now how about XMMS...

I installed it yesterday and it (sortof) worked. It loaded up, but it would not play any music (I got a warning about the DiskWriter plug-in IIRC.) I switched to the only other available output (some eSound something or other) and when I hit play the application basically quit responding. Ever since then I get this warning when I try to load XMMS with "xmms &"

Gdk-WARNING **: locale not supported by C library

Also, I rebuilt the application with the hope that that would fix it, but it didn't (XMMS worked when I was using XWindows with 10.1)

justnip,

I checked in Fink database and there is no blackbox available in the "stable" branch for 10.2:

Package blackbox
In current-unstable:------Version 0.65.0-1
In current-stable:--------not present
In 0.5.0a-stable:---------not present
In 0.4.1-stable:-----------Version 0.61.1-3 (10.1 only)

http://fink.sourceforge.net/pdb/package.php/blackbox

thatch is correct. If you want just the "blackbox" package, you will have to use the unstable branch, and for that you should add "unstable/main unstable/crypto" to the Trees: line in your fink.config file in /sw/etc. And after ran 'fink selfupdate' or 'fink selfupdate-cvs'.

Or, you could install "blackbox-rootless", which you "have" in the stable branch now. It's the same package plus works in rootless-mode.

'fink install blackbox-rootless"

Package blackbox-rootless
In current-unstable: ___ Version 0.65.0-1
In current-stable: ___ Version 0.62.1-11
In 0.5.0a-stable: ___ Version 0.62.1-11
In 0.4.1-stable: ___ Version 0.62.1-1 (10.1 only)


Cheers...

PS: Justnip, I posted too late, and I just saw your post above...

justnip,

Check the forums for "xmms and no sound" and also read the Fink FAQ here:

http://fink.sourceforge.net/faq/usag...php#xmms-quiet


Gdk-WARNING **: locale not supported by C library:
it just means that the program will use the default English messages, date formats, etc. The program will function normally otherwise.

Check here for more info:

http://fink.sourceforge.net/doc/x11/trouble.php#locale


Cheers...

LizardKing,

Thanks a lot for the info. I guess I was quite sleepy last night, I should have realized that you had made also an OS version change in the meantime (fink 0.9.1 was for 10.1).

You made the right decision by removing and installing Fink- 0.5.0a from scratch.


Cheers...

justnip,

You wrote in a post above:
This is very weird, because when you install xfree86-rootless with Fink, it installs also the XDarwin.app. Why would you want to update it with the one at the Xonx project site?

Cheers...

thatch,

Good work at keeping the troops alive. :)

Cheers...

Ah, very good, thank you (all.)

Interestingly enough, while esd was running as a normal user as it is supposed to, killing it and then removing its temporary file fixed the problem anyway.

:)

EDIT: Missed your post. I installed the app from the XonX site because it hadn't been installed (wasn't in my Apps folder.)

It can't be, somethings must have gone wrong with your installation of the xfree86-rootless package with Fink. I would definitely ran:

'fink reinstall xfree86-rootless'

Cheers...

justinp, as I had noted earlier, I don't remember ever having to install XDarwin separately from xfree86 and that's because it's integrated into the same install. So, a reinstall is the correct call for it as sao said. Otherwise, glad to hear you added the unstable tree and it's worked out well for you.

sao, I'm glad to help out where I can as you and so many of the other great folks here have done for me. That's part of what makes this forum the best! We sure have some good people here and I appreciate it very much.

Best,

thatch

Hi everybody :)
I've got a very new installation of jaguar, and i wanted to install fink 0.5 . The problem is that i can't figure out how to make it start the way it should.

So, after the package installation, i follow the steps and add "source /sw/bin/init.csh" to my .cshrc file, quit and relaunch Terminal. But, after that, no fink instructions are recognized by the shell. "fink list" for example, etc...

when i open a new terminal window and type first "source /sw/bin/init.csh", no problem, i can play with this fantastic fink in this terminal window....

Obviously, my .cshrc file is not read and interpreted correctly at the shell startup. I have made just one change to the default configuration, i've followed the hint to recover the tcsh functionnalities in jaguar that were in puma (this hint )

So, the problem is i've different instructions in both .cshrc and .tcshrc. I tried to put everything in .cshrc, so that it looks like :But it doesn't work either.... (i get rid of .tcshrc in the same movement)

So, i'm a bit confused, thanks for your help :)

the call to rc is stomping the path after init.csh sets it.

reverse the order of the statements:

source /sw/bin/init.csh
source /usr/share/tcsh/examples/rc

should become:


source /usr/share/tcsh/examples/rc
source /sw/bin/init.csh

maousse,

Or, try by writing in your ~/.tcshrc file:

source ~/.cshrc

It worked fine with me.


Cheers...

Ok, thanks to both of you for your help :)

i've tried both solutions, and none works...
I've cheked up a little closer to the hint page, and finally found a solution in comments (here )
Here's the quote from the original post :
And it works ! I'm asking the same question, but it works... :cool: Thank you again

it may work, but it is logically wrong.

...just wrong on so many levels...
 

This subject is definitely resurfacing every now and again and can be illogical and somewhat unexplainable at times. There are so many end user variables and factors involved and everyone can't be posting their dirty laundry, the reference that Paul makes to the contents of ~/.*rc log files and such in the old resident favorite thread, Best way to set environment?, which is still alive and kicking and has transformed through various OS's and shells and is still probably the best place to continue the discussion if need be.

And why would I be so long windily pointing the focus of this in that direction you might ask? Well, other than my compulsive nature to keep things organized, it is somewhat self serving only because I recently revived that old thread with some of my own chunder, not that I expect any ground breaking revelations to come of it all.

Okay, I'll stop babbling and ease my *ss out the door now... ;)

here's a thread with some proof of why ~/.login is where to augment the path variable...

http://forums.macosxhints.com/showth...&threadid=7721

Yes, I saw that post yesterday and couldn't agree more. And that's where I augment my path variable too. Still, it's a mystery to me why xterm in windowmaker has redundancies when terminal.app is perfect and that xterm doesn't honor all the variables set in ~/.login or ~/.tcshrc.

i saw a good explanation of this some while back somewheres, but for the life of me, i couldn't google it.

does this help? ...

http://supportweb.cs.bham.ac.uk/howto/login/


i think the critter of it is that each new terminal.app window is a login shell window, but only the first xterm is the login window and some jiggery-pokery needs to occur to determine the config for subsequent shells.

i keep looking.

mT, thanks! That's an excellent article and I've not seen it before. It confirms something I had been wondering about with regard to this subject, that there is a such thing as an ~/.xlogin file.

I had asked that question not too long ago, here, but I thought it was wishful thinking on my part since I didn't hear anything to the contrary.

I've got some tinkering to do now to see if I can straighten out the path and dircolors stuff with xterm. Hopefully this will be the missing link that will do the trick. Looking back at this, I should have just tried it since I had thought it was necessary.

sao - I did the follwoing:

'fink rebuild xfree86-base'
'fink rebuild xfree86-rootless'

and still have maskless curors...

Any other ideas?

bluehz:

http://sourceforge.net/project/shown...ease_id=118483

?

justinp - thx - I already updated long ago when it first came out.

I just discovered - if I start X from CLI:

startx -- -rootless -quartz

my cursors are fine.

If I start the XDarwin app - which is my normal MO....that when I get the maskless cursors. So it appears the problem may be within XDarwin app.

FYI - XDarwin v 1.1.1.1 (circa Nov 18, 2002)

Bluehz,

How does it comes out when you start from terminal.app with:

startx -- -fullscreen

or

startx -- -rootless

Prior to 4.2, -quartz was used for fullscreen mode. The -quartz option no longer selects fullscreen mode, but rather uses the default mode set in the preferences.

Read the following info:

http://fink.sourceforge.net/doc/x11/....php#macosx-41


Cheers...

Both of those work fine - nice masked cursors.

bluehz,

Great !

I don't know about the Xdarwin.app, I have the same version as you, and when I start XFree with it, the cursor shows fine.

The code in X since 4.2.1.1 has fixed the black cursor outlines on Mac OS X 10.2.2 as it should be.

Remember that the bug was actually in Quartz in 10.2.1 and earlier. Quartz was initializing automatically that it wasn't supposed to be, according to the API.

Then the problem appeared when they fixed the bug in Quartz, it broke X code that didn't properly initialize the cursor in the first place. So they released X 4.2.1.1 which solved this problem.

It looks like somehow your installation keeps using the old code.


Cheers...