| View previous topic :: View next topic |
| Author |
Message |
gentoo_lan l33t


Joined: 08 Sep 2004 Posts: 887 Location: Osan AB, South Korea
|
Posted: Mon Jan 23, 2006 4:02 am Post subject: |
|
|
I'm glad I found this script again. I had an older version I hadn't used in a while and the new version works great. _________________ Registered Linux user# 375038. |
|
| Back to top |
|
 |
teh-steve n00b

Joined: 28 Jan 2005 Posts: 27 Location: Worcester
|
Posted: Fri Feb 10, 2006 4:14 am Post subject: |
|
|
| Still waiting for that update So Tired. How's it coming along? |
|
| Back to top |
|
 |
gentoo_lan l33t


Joined: 08 Sep 2004 Posts: 887 Location: Osan AB, South Korea
|
Posted: Sun Feb 12, 2006 10:32 pm Post subject: |
|
|
This script also needs improvements to include checks for package.provided. I may try to alter the script myself to see if I can improve it. _________________ Registered Linux user# 375038. |
|
| Back to top |
|
 |
red-wolf76 l33t


Joined: 13 Apr 2005 Posts: 713 Location: Rhein-Main Area
|
Posted: Mon Apr 17, 2006 11:01 pm Post subject: |
|
|
First of all, this script rocks for a lark!
I did find a bit of a cockup with some of the fonts from the modular Xorg list, though. The relevant part of portage.keywords reads: | Code: | media-fonts/font-adobe-100dpi
media-fonts/font-adobe-75dpi
media-fonts/font-adobe-utopia-100dpi
media-fonts/font-adobe-utopia-75dpi
media-fonts/font-adobe-utopia-type1
media-fonts/font-alias
media-fonts/font-arabic-misc
media-fonts/font-bh-100dpi
media-fonts/font-bh-75dpi
media-fonts/font-bh-lucidatypewriter-100dpi
media-fonts/font-bh-lucidatypewriter-75dpi
media-fonts/font-bh-ttf
media-fonts/font-bh-type1
media-fonts/font-bitstream-100dpi
media-fonts/font-bitstream-75dpi | The script will incorrectly mark duplicates for
media-fonts/font-adobe
media-fonts/font-adobe-utopia
media-fonts/font-bh
media-fonts/font-bh-lucidatypewriter and
media-fonts/font-bitstream,
all of which have numbers directly following them. So I guess the script thinks they're versioned...
Other than that, great! _________________ 0mFg, G3nt00 r0X0r$ T3h B1g!1111
Use sane CFLAGS! If for no other reason, do it for the lulz! |
|
| Back to top |
|
 |
slycordinator Advocate


Joined: 31 Jan 2004 Posts: 3050 Location: Houston, TX
|
Posted: Tue Apr 18, 2006 12:38 am Post subject: |
|
|
| red-wolf76 wrote: | First of all, this script rocks for a lark!
I did find a bit of a cockup with some of the fonts from the modular Xorg list, though. The relevant part of portage.keywords reads: | Code: | media-fonts/font-adobe-100dpi
media-fonts/font-adobe-75dpi
media-fonts/font-adobe-utopia-100dpi
media-fonts/font-adobe-utopia-75dpi
media-fonts/font-adobe-utopia-type1
media-fonts/font-alias
media-fonts/font-arabic-misc
media-fonts/font-bh-100dpi
media-fonts/font-bh-75dpi
media-fonts/font-bh-lucidatypewriter-100dpi
media-fonts/font-bh-lucidatypewriter-75dpi
media-fonts/font-bh-ttf
media-fonts/font-bh-type1
media-fonts/font-bitstream-100dpi
media-fonts/font-bitstream-75dpi | The script will incorrectly mark duplicates for
media-fonts/font-adobe
media-fonts/font-adobe-utopia
media-fonts/font-bh
media-fonts/font-bh-lucidatypewriter and
media-fonts/font-bitstream,
all of which have numbers directly following them. So I guess the script thinks they're versioned...
Other than that, great! |
Actually, what I'm seeing happen is this:
There is no such package as media-fonts/font-adobe. But there is a package called media-fonts/font-adobe-100dpi. The script is assuming the "100dpi" part is part of the version number when it is not.
For instance, I have media-fonts/font-adobe-100dpi-1.0.0 installed. 1.0.0 is the version. Everything before it is a package name. |
|
| Back to top |
|
 |
nesl247 Veteran

Joined: 15 Jun 2004 Posts: 1614 Location: Florida
|
Posted: Tue Apr 18, 2006 3:23 am Post subject: |
|
|
| Sly, that is exactly what he said.. |
|
| Back to top |
|
 |
slycordinator Advocate


Joined: 31 Jan 2004 Posts: 3050 Location: Houston, TX
|
Posted: Tue Apr 18, 2006 5:32 am Post subject: |
|
|
| nesl247 wrote: | | Sly, that is exactly what he said.. |
I assumed when he said "incorrectly mark duplicates" it was meaning that those entries would incorrectly be marked as duplicate entries, rather than being marked as entries for items not installed. |
|
| Back to top |
|
 |
nesl247 Veteran

Joined: 15 Jun 2004 Posts: 1614 Location: Florida
|
Posted: Tue Apr 18, 2006 5:34 am Post subject: |
|
|
| I was referring to your explanation, and when he said "all of which have numbers directly following them. So I guess the script thinks they're versioned...". Didn't catch that other part. |
|
| Back to top |
|
 |
red-wolf76 l33t


Joined: 13 Apr 2005 Posts: 713 Location: Rhein-Main Area
|
Posted: Tue Apr 18, 2006 9:02 am Post subject: |
|
|
| slycordinator wrote: | Actually, what I'm seeing happen is this:
There is no such package as media-fonts/font-adobe. But there is a package called media-fonts/font-adobe-100dpi. The script is assuming the "100dpi" part is part of the version number when it is not.
For instance, I have media-fonts/font-adobe-100dpi-1.0.0 installed. 1.0.0 is the version. Everything before it is a package name. | Absolutely correct. So while media-fonts/adobe-100dpi and media-fonts/adobe-75dpi are installed - regardless of their true version number - the script (erroneously) treats them as versions 100 and 75 of a fictitious package called "media-fonts/adobe".
Given that the RegEx is such a pain to edit, I'm guessing there would be a way to get it to recognize font dpis, but it would be somewhat hard and stony.
I just posted that to ensure that people don't go off deleting "duplicates" and thrashing their modular Xorg installation's fonts... _________________ 0mFg, G3nt00 r0X0r$ T3h B1g!1111
Use sane CFLAGS! If for no other reason, do it for the lulz! |
|
| Back to top |
|
 |
slycordinator Advocate


Joined: 31 Jan 2004 Posts: 3050 Location: Houston, TX
|
Posted: Tue Apr 18, 2006 8:26 pm Post subject: |
|
|
| red-wolf76 wrote: | | I just posted that to ensure that people don't go off deleting "duplicates" and thrashing their modular Xorg installation's fonts... |
Since the script only provides removing entries in an "all-or-none" fashion all you'd have to do is:
1) Run the script, remove all entries it says are bad.
2) Run "emerge world --update --deep --newuse -pv"
3) Whenever a package comes up as being needed but masked, unmask it again
4) Repeat steps 2 and 3 until no errors occur. |
|
| Back to top |
|
 |
Onip Advocate


Joined: 02 Sep 2004 Posts: 2735 Location: Parma (Italy)
|
Posted: Thu Jul 13, 2006 4:41 pm Post subject: |
|
|
I've noticed a little problem
| Code: | Hal9000 ~ # ./etcportclean
Check package.keywords for unnecessary keywords? (yes/no)
y
Check package.mask for needlessly masked packages? (yes/no)
y
Check package.use for unnecessary use flags? (yes/no)
y
Check package.cflags for repeat cflags? (yes/no)
n
Check package.cxxflags for repeat cxxflags? (yes/no)
n
./etcportclean: line 738: [: too many arguments <-- PROBLEM
Checking for unnecessary entries in package.keywords...
|
Here's what I have at that line
| Code: | | if [ ${ynarray[0]} -eq 1 -o ${ynarray[1]} -eq 1 -o ${ynarray[2]} -eq 1 ] ; then checkworld ; fi |
Thanks _________________ Linux Registered User n. 373835
Titus Lucretius Carus, De Rerum Natura - Tanto religio potuit suadere malorum |
|
| Back to top |
|
 |
aguettinger Tux's lil' helper

Joined: 26 Sep 2003 Posts: 83
|
Posted: Sat Jul 15, 2006 3:05 pm Post subject: |
|
|
This script is very useful.
I have a request, though. As I want maximum control myself, I would love to see a flag so the script just comments (#) the obsolete entries in packages.keywords instead to actually remove them. I usually have a lot of commented lines in package.keywords.
Furthermore I support the request for a --pretend flag.
Hope to see a rewrite of this soon. _________________ regards
- andy |
|
| Back to top |
|
 |
basiaf n00b


Joined: 25 Feb 2005 Posts: 38 Location: Dortmund
|
Posted: Sun Aug 27, 2006 8:18 pm Post subject: |
|
|
| Just found this script again (my last version was 1.13). I like this script very much, but it would be great if you could add support for the portage 2.1 config-files-as-directories feature ( e.g. package.keywords/xgl, package.keywords/foo ) |
|
| Back to top |
|
 |
mv Advocate


Joined: 20 Apr 2005 Posts: 2589
|
Posted: Thu Aug 31, 2006 7:48 am Post subject: |
|
|
| basiaf wrote: | | Just found this script again (my last version was 1.13). I like this script very much, but it would be great if you could add support for the portage 2.1 config-files-as-directories feature ( e.g. package.keywords/xgl, package.keywords/foo ) |
As it still seems to be rather unknown, I would like to point out that >=eix-0.6.3 provides with eix -tT a similar functionality as this script, however without actual removal of the entries; details of the mentioned functionality are highly cusomizable - see man eix. eix supports this config-files-as-directories feature.
Concerning the other bug with the falsely recognized version numbers and the fonts: It seems rather obvious to me that some regular expression in the script fails to be greedy, as the version number is the last (not the first) item of the name starting with -[0-9]. However, I did not check through the script where this happens, since for my purposes eix is more appropriate anyway... |
|
| Back to top |
|
 |
slycordinator Advocate


Joined: 31 Jan 2004 Posts: 3050 Location: Houston, TX
|
Posted: Thu Aug 31, 2006 8:25 am Post subject: |
|
|
| Gatak wrote: | | What do you mean with versioned keywords? The example ~dev-php/jpgraph-1.17 ~x86 is wrong. It should be =dev-php/jpgraph-1.17 ~x86 or >=dev-php/jpgraph-1.17 ~x86 or <=dev-php/jpgraph-1.17 ~x86 |
That's not true.
~dev-php/jpgraph-1.17 ~x86
is perfectly valid.
What it does is it unmasks any version which has 1.17 in it. So if I had version 1.17-r1 then it would get unmasked as well. So it essentially unmasks any revision of 1.17.
edit:
None of your suggestions do what this one does.
1) =dev-php/jpgraph-1.17 just unmasks that version and no others
2) >=dev-php/jpgraph-1.17 unmasks EVERY version greater than or equal to it.
3) <=dev-php/jpgraph-1.17 unmasks EVERY version less than or equal to it
The usefulness of this is if I think a newer branch is too buggy I can force my unmask to just work on the 1.17 branch. |
|
| Back to top |
|
 |
lmcogs Guru

Joined: 03 Apr 2005 Posts: 340
|
Posted: Fri Nov 17, 2006 7:51 pm Post subject: |
|
|
Hi I tried this and seems to work fine for me. Pretty handy it seems but have not tried extensively only etcportclean -c
2.6.17-gentoo-r4 #1 SMP Mon Aug 21 01:27:10 GMT 2006 x86_64 AMD Athlon(tm) 64 X2 Dual Core Processor 4400+ AuthenticAMD GNU/Linux |
|
| Back to top |
|
 |
cpicton n00b

Joined: 12 Oct 2004 Posts: 7
|
Posted: Tue Jan 01, 2008 5:56 am Post subject: |
|
|
Hi
Script has problems with directories:
I have /etc/portage/package.keywords/*, so I can keep track of keyworded packages more easily. The script fails saying that package.keywords must exist. |
|
| Back to top |
|
 |
red-wolf76 l33t


Joined: 13 Apr 2005 Posts: 713 Location: Rhein-Main Area
|
Posted: Tue Jan 01, 2008 12:48 pm Post subject: |
|
|
| slycordinator wrote: | | Gatak wrote: | | What do you mean with versioned keywords? The example ~dev-php/jpgraph-1.17 ~x86 is wrong. It should be =dev-php/jpgraph-1.17 ~x86 or >=dev-php/jpgraph-1.17 ~x86 or <=dev-php/jpgraph-1.17 ~x86 |
That's not true.
~dev-php/jpgraph-1.17 ~x86
is perfectly valid.
What it does is it unmasks any version which has 1.17 in it. So if I had version 1.17-r1 then it would get unmasked as well. So it essentially unmasks any revision of 1.17.
edit:
None of your suggestions do what this one does.
1) =dev-php/jpgraph-1.17 just unmasks that version and no others
2) >=dev-php/jpgraph-1.17 unmasks EVERY version greater than or equal to it.
3) <=dev-php/jpgraph-1.17 unmasks EVERY version less than or equal to it
The usefulness of this is if I think a newer branch is too buggy I can force my unmask to just work on the 1.17 branch. |
I concur. I use the | Code: | | ~package-foo/bar-0.0.1 | syntax almost exclusively in my package.keywords file. It would be nice if it were more commonly recognized. _________________ 0mFg, G3nt00 r0X0r$ T3h B1g!1111
Use sane CFLAGS! If for no other reason, do it for the lulz! |
|
| Back to top |
|
 |
dfelicia Apprentice


Joined: 11 May 2005 Posts: 281 Location: Southwestern Connecticut
|
Posted: Fri Jul 25, 2008 1:49 pm Post subject: |
|
|
Script doesn't work with latest sed (sys-apps/sed-4.1.5-r1):
| Code: |
sed: -e expression #1, char 41: Invalid range end
|
|
|
| Back to top |
|
 |
loftwyr l33t


Joined: 29 Dec 2004 Posts: 970 Location: 43°38'23.62"N 79°27'8.60"W
|
Posted: Fri Jul 25, 2008 4:29 pm Post subject: |
|
|
This script was abandoned two years ago.
We need someone to step up and revamp it. It would be nice if it also handled paludis and other package managers. _________________ My emerge --info
Have you run revdep-rebuild lately? It's in gentoolkit and it's worth a shot if things don't work well.
Celebrating 5 years of Gentoo-ing. |
|
| Back to top |
|
 |
forkboy Apprentice


Joined: 24 Nov 2004 Posts: 200 Location: Blackpool, UK
|
Posted: Fri Jul 25, 2008 4:37 pm Post subject: |
|
|
| I use app-portage/portpeek from portage, does it do the same thing? |
|
| Back to top |
|
 |
|