| View previous topic :: View next topic |
| Most Wanted Portage Feature? |
| emerge -u --security |
|
13% |
[ 54 ] |
| simultaneous package downloads |
|
9% |
[ 38 ] |
| overall progress bar |
|
20% |
[ 83 ] |
| reverse dependencies |
|
51% |
[ 211 ] |
| other |
|
5% |
[ 21 ] |
|
| Total Votes : 407 |
|
| Author |
Message |
wzzrd
Apprentice
Joined: 18 Jul 2002
Posts: 245
|
 Posted: Wed Oct 08, 2003 6:29 am Post subject: |
 |
|
| Genone wrote: |
well if you want to do a lot of searching, i refer you to be my site it uses an sql backend version of the portage tree, im in the process of adding functionality right now. any suggestions would be great. |
I, for one, would like a frontend of some sort, just from portage-browsing purposes. Reverse deps would be nice. So would a db-backend.
Genone, are you planning to create a PHP-frontend for portage and release it? Sound pretty nifty; I am very interested  |
|
| Back to top |
|
 |
zenlunatic
Guru
Joined: 09 Apr 2003
Posts: 312
|
| Posted: Wed Oct 08, 2003 6:31 am Post subject: |
|
|
| Sourcemage has md5 checks in their port system. Maybe we could use some of their code, since it's GPL'd, to implement simliar feature in portage. |
|
| Back to top |
|
|
Gruffi
Apprentice
Joined: 15 Aug 2003
Posts: 209
Location: Antwerpen - Flanders - Belgium
|
| Posted: Wed Oct 08, 2003 8:14 am Post subject: |
|
|
Would love to be able to type
| Code: | | "emerge --fetchonly --no-md5-sum-checking" |
or
| Code: | | "emerge --fetchonly --only-md5-sum-checking-when-downloaded" |
I have a portage-mirror-pc that prefetches all packages for all other pc's, it's slow to MD5-check packages that are allready downloaded and have been checked before.
_________________
... and we will show Microsoft, that they cannot take whatever they want. And that Free Software is our software! |
|
| Back to top |
|
|
thrasher6670
Apprentice
Joined: 04 Aug 2003
Posts: 269
Location: London, Ontario
|
| Posted: Wed Oct 08, 2003 11:59 am Post subject: |
|
|
**oops**
_________________
Gentoo-Portage.com - Where the portage meets the web
Gentoo Wiki
Last edited by thrasher6670 on Wed Oct 08, 2003 2:07 pm; edited 1 time in total |
|
| Back to top |
|
|
Genone
Retired Dev
Joined: 14 Mar 2003
Posts: 9538
Location: beyond the rim
|
| Posted: Wed Oct 08, 2003 2:05 pm Post subject: |
|
|
| wzzrd wrote: | | Genone wrote: |
well if you want to do a lot of searching, i refer you to be my site it uses an sql backend version of the portage tree, im in the process of adding functionality right now. any suggestions would be great. |
I, for one, would like a frontend of some sort, just from portage-browsing purposes. Reverse deps would be nice. So would a db-backend.
Genone, are you planning to create a PHP-frontend for portage and release it? Sound pretty nifty; I am very interested |
You quoted the wrong person  |
|
| Back to top |
|
|
wzzrd
Apprentice
Joined: 18 Jul 2002
Posts: 245
|
| Posted: Thu Oct 09, 2003 7:59 am Post subject: |
|
|
| Genone wrote: | | wzzrd wrote: | | Genone wrote: |
well if you want to do a lot of searching, i refer you to be my site it uses an sql backend version of the portage tree, im in the process of adding functionality right now. any suggestions would be great. |
I, for one, would like a frontend of some sort, just from portage-browsing purposes. Reverse deps would be nice. So would a db-backend.
Genone, are you planning to create a PHP-frontend for portage and release it? Sound pretty nifty; I am very interested |
You quoted the wrong person |
Darn! Got some self-slapping with a large trout tot do now! |
|
| Back to top |
|
|
Bowyakka
Tux's lil' helper
Joined: 04 Oct 2002
Posts: 142
Location: London
|
| Posted: Sun Oct 12, 2003 10:55 am Post subject: |
|
|
| Quote: | oh so much is wrong with Portage I don't know where to begin...
....
2. bye bye FS based portagetree - hello sqlite (or something like it) backend - |
might be nice | Quote: |
....
6. layered overlays |
Lovechild man what is a layered overlay ?? | Quote: |
....
7. threading of emerge (to allow for downloading while we emerge)
|
Yes yes yes yes yes yes | Quote: |
10. Sticky flags
11. gtk USE flag for gtk1 deps, gtk2 USE flag for gtk2 deps - it should be that simple, but it's not - this is just SILLY !
|
Also:
Fixing portages behaviour on things like kernel modules (I have two dev kernels 2.6-tes6 2.6-test7 now would it be nice if portage kept the nvidia modules in 2.6-test6 when I do emerge nvidia-kernel)
The security thing would be nice
| Quote: | | * download/compile parallelism: well, you all know what I mean. |
[/quote]
_________________
glows in the dark |
|
| Back to top |
|
|
fca
Guru
Joined: 22 Feb 2003
Posts: 346
Location: Netherlands
|
| Posted: Sun Oct 12, 2003 11:37 am Post subject: |
|
|
[quote="Bowyakka"] | Quote: |
...
Fixing portages behaviour on things like kernel modules (I have two dev kernels 2.6-tes6 2.6-test7 now would it be nice if portage kept the nvidia modules in 2.6-test6 when I do emerge nvidia-kernel)
...
|
See bug #1477
Apparenltly it has been bugging people since april last year, and there exists a sort of fix.
Also what would fix this is a more refined ${KV}, which takes the patch level of the kernel into account, but that would require an extension of slots which need to accept non-integers (maybe this is already possible, but I can't find much documentation on this) |
|
| Back to top |
|
|
Bowyakka
Tux's lil' helper
Joined: 04 Oct 2002
Posts: 142
Location: London
|
| Posted: Sun Oct 12, 2003 1:51 pm Post subject: |
|
|
Oh well at least that one is known about
_________________
glows in the dark |
|
| Back to top |
|
|
kamikaz3
Apprentice
Joined: 06 Feb 2003
Posts: 187
|
| Posted: Sun Oct 12, 2003 3:20 pm Post subject: |
|
|
1. Something better then etc-update, this thing is very time consuming when updating world
2. Something to clean out /usr/portage/distfiles, I would like it to keep the sources installed on my system but delete the old ones (my /usr/portage/distfiles is 3.2G big now but it's used by 4 different PC's)
_________________
- Gentoo Linux gives me power, speed and flexibility - |
|
| Back to top |
|
|
qanuta
n00b
Joined: 04 Apr 2003
Posts: 43
Location: NJ, USA
|
| Posted: Sun Oct 12, 2003 10:46 pm Post subject: Version change threshold! |
|
|
| The most important new feature is some sort of version threshold to be set in emerge. For example, when I do an --update world, I don't really want to update package-2.1.3-r2 simply because package-2.1.3-r3 is now available. For most packages, these infinitesimal changes don't make a substantial difference, and they take forever to re-emerge every week. Instead, I would like to be able to tell it not to bother unless the version change is > 0.1, or some other user-specified threshold. |
|
| Back to top |
|
|
Jyrinx
Tux's lil' helper
Joined: 03 May 2002
Posts: 92
Location: Carleton College - Northfield, MN
|
| Posted: Tue Oct 14, 2003 4:52 am Post subject: Re: Version change threshold! |
|
|
| qanuta wrote: | | The most important new feature is some sort of version threshold to be set in emerge. For example, when I do an --update world, I don't really want to update package-2.1.3-r2 simply because package-2.1.3-r3 is now available. For most packages, these infinitesimal changes don't make a substantial difference, and they take forever to re-emerge every week. Instead, I would like to be able to tell it not to bother unless the version change is > 0.1, or some other user-specified threshold. |
Well, sometimes that extra -r is critical - a security update, for instance. What does bother me is updates 1) strictly fixing build problems, which are irrelevant if the package is installed and hence able to be upgraded; and 2) for other platforms. A way to skip those updates (or everything but security updates) would be très spiffy. |
|
| Back to top |
|
|
3lue9
n00b
Joined: 12 Oct 2003
Posts: 11
|
| Posted: Tue Oct 14, 2003 5:26 am Post subject: If we can't make it faster, make it smarter |
|
|
Well some of the ideas are slightly ridiculous, like C++ instead of python; however, there are a lot of good ideas here. I have been using bash scripting to solve most of the problems people have faced. I suggest looking at www.tldp.org for coding via bash. A good example is -u security. In essence you know what you want to make secure (openssh, apache, ...) so make a bash script to handle this specific upgrade.
Well the three features I'd like to see are time constraints, time approximations, simultaneous download/compiling. These will really help speed the process up and make it more user friendly. What I mean by time constraints is emerge -u --time world 0630 which would install everything it could with a 6 and a half hour limit. Nothing bothers me more then -u world before I go to bed and then when i wake up a few hours later to have my laptop still compiling.
Time approximations are slightly overlapping, but needed general use. I think using the bash units is good so long they are adjusted for you. Thus when compiling bash, some script logs the start and end time to use for your system.
The only downside to portage is speed, but these features will allow for greater user flexibility thus minimizing the unavoidable downsides of portage.
_________________
i don't have a signature |
|
| Back to top |
|
|
Evangelion
Veteran
Joined: 31 May 2002
Posts: 1087
Location: Helsinki, Finland
|
| Posted: Tue Oct 14, 2003 7:06 am Post subject: |
|
|
| Lovechild wrote: | oh so much is wrong with Portage I don't know where to begin...
3. a proper gtk2 frontend |
Well, what's stopping you? Should Portage-devels create GTK2-frontend? Why couldn't some other people do it? There seems to be plenty of Qt/KDE-based frontends:
Kemerge
Kportage
There doesn't seem to be anything preventing a GTK2-frontend to be made. It just might be that people are not interested in one, and I don't think it's the duty of Portage-developers to write one.
In other words: you are barking at the wrong tree .
_________________
My tech-blog | My other blog |
|
| Back to top |
|
|
TGL
Bodhisattva
Joined: 02 Jun 2002
Posts: 1978
Location: Rennes, France
|
| Posted: Tue Oct 14, 2003 7:10 am Post subject: Re: If we can't make it faster, make it smarter |
|
|
| 3lue9 wrote: | | Well some of the ideas are slightly ridiculous, like C++ instead of python |
Agreed.
| 3lue9 wrote: | | I have been using bash scripting to solve most of the problems people have faced. |
What about sharing your work? Helpful scripts are always welcome in the "Doc, Tips and Tricks" forum, and also in this thread.
| 3lue9 wrote: | | A good example is -u security. In essence you know what you want to make secure (openssh, apache, ...) so make a bash script to handle this specific upgrade. |
The goal of a security update is to only update packages which fix security flows, not to update a well known list of packages. And this obviously requires portage additions. |
|
| Back to top |
|
|
Gushy
Apprentice
Joined: 03 Aug 2002
Posts: 284
Location: Reading, UK
|
| Posted: Tue Oct 14, 2003 12:57 pm Post subject: |
|
|
I want them all although I voted other. My most wanted feature would be a snapshot/rollback facility.
so with that in mind:
- snapshot/rollback
- reverse dependencies
- emerge -u --security
- simultaneous package downloads
- overall progress bar
_________________
--
Gushy |
|
| Back to top |
|
|
binary_runner
n00b
Joined: 26 Oct 2003
Posts: 64
|
| Posted: Sun Nov 02, 2003 11:40 am Post subject: wanted portage features |
|
|
Ability to download differences instead of whole packages would be very useful, on slower lines it could boost speed.
.. And let's think about kind of transactions support : "rollback, last N updates of package X" e.g.
I'm new to portage and gentoo so I do not know how it binary packages are accessible, maybe some resporitory should be usefull -- no need to waste sources (time, energy) to compile something the same way anothe rperson already did. |
|
| Back to top |
|
|
thrasher6670
Apprentice
Joined: 04 Aug 2003
Posts: 269
Location: London, Ontario
|
|
| Back to top |
|
|
romaninsh
n00b
Joined: 09 Jun 2003
Posts: 13
|
| Posted: Mon Nov 03, 2003 11:59 pm Post subject: |
|
|
other:
rewrite portage core in C using sqlite.
remove patchfiles from portage and cleanup it
think of some way to make optional portage categories. Only few need app-benchmarks, why can't it be optional.
im against reverse dependencies. revdep-clean/world is enough. |
|
| Back to top |
|
|
dbergst
Tux's lil' helper
Joined: 25 Mar 2003
Posts: 88
|
| Posted: Tue Nov 04, 2003 12:47 am Post subject: |
|
|
| gsfgf wrote: | | it not to take 20 mins to regnerate portag cache after rsyncing. |
Amen to that!  |
|
| Back to top |
|
|
Genone
Retired Dev
Joined: 14 Mar 2003
Posts: 9538
Location: beyond the rim
|
| Posted: Tue Nov 04, 2003 2:05 am Post subject: |
|
|
| romaninsh wrote: | | rewrite portage core in C using sqlite. |
I won't say anything about this, there are enough threads about it.
| Quote: | | remove patchfiles from portage and cleanup it |
Not sure I understand what you mean here
| Quote: | | think of some way to make optional portage categories. Only few need app-benchmarks, why can't it be optional. |
RSYNC_EXCLUDE_FROM in make.conf |
|
| Back to top |
|
|
zhenlin
Veteran
Joined: 09 Nov 2002
Posts: 1361
|
| Posted: Tue Nov 04, 2003 10:23 am Post subject: |
|
|
| Genone wrote: | | Quote: | | remove patchfiles from portage and cleanup it |
Not sure I understand what you mean here |
Probably wants the portage tree to be a purely ebuild-only tree.
Could be an interesting idea, with a corresponding ${P}-aux.tar.bz2 for each package... But it will probably consume more bandwidth than possibly imaginable. Also, it would also ruin the versioning policy, since changing the auxilliary files would result in a tarball with a different MD5sum... |
|
| Back to top |
|
|
RikBlankestijn
Guru
Joined: 01 May 2003
Posts: 315
Location: Arnhem the Netherlands
|
| Posted: Tue Nov 04, 2003 8:05 pm Post subject: |
|
|
| Lovechild wrote: | oh so much is wrong with Portage I don't know where to begin...
1. Trust, there's no trust in portage besides md5sum checks... a man in the middle attack would be easy to stage fx.
2. bye bye FS based portagetree - hello sqlite (or something like it) backend
3. a proper gtk2 frontend
4. per package USE flags instead of global flags (handled in the before mentioned GTK2 frontend)
5. games-roguelike -> games-rpg (silly silly silly)
6. layered overlays
7. threading of emerge (to allow for downloading while we emerge)
8. UUID based internal naming aliased to human readable names (fixes all move, renaming, etc. problems)
9. Gentoo menu item handling system (like Debian, following the freedesktop.org standard maybe?)
10. Sticky flags
11. gtk USE flag for gtk1 deps, gtk2 USE flag for gtk2 deps - it should be that simple, but it's not - this is just SILLY !
those are in no important order what I can think off of the top of my head - but portage could be improved much more than that I think. |
Start working on it 
btw. I think it's great already. I've tried a few other distro's, but since I'm using Gentoo I finally get to know Linux! That includes the portage...I now know what is being installed and how etc.. Couldn't say that about these "nice frontended package managers" that don't work anywayz. I'm happy with no gtk frontend or whatever just the reverse dependensies would make me even more happy. |
|
| Back to top |
|
|
guy
Apprentice
Joined: 31 Mar 2003
Posts: 286
Location: USA
|
| Posted: Tue Nov 04, 2003 8:29 pm Post subject: |
|
|
I would love to see an XML file associated with each merged package / binary package giving details about it, such as the architure and USE flags it was merged/compiled with. This would be a big step towards sharing binaries and a big step towards my live cd project 
_________________
www.manatee.servemp3.com
www.seas.upenn.edu/~bdodson/gina
www.goatse.cx [it will rise again!] |
|
| Back to top |
|
|
zhenlin
Veteran
Joined: 09 Nov 2002
Posts: 1361
|
| Posted: Wed Nov 05, 2003 9:47 am Post subject: |
|
|
| Lovechild wrote: | oh so much is wrong with Portage I don't know where to begin...
1. Trust, there's no trust in portage besides md5sum checks... a man in the middle attack would be easy to stage fx.
2. bye bye FS based portagetree - hello sqlite (or something like it) backend
6. layered overlays
7. threading of emerge (to allow for downloading while we emerge)
8. UUID based internal naming aliased to human readable names (fixes all move, renaming, etc. problems)
10. Sticky flags
|
#8 is not exactly widely accepted... And #10 has implementation issues... #2 complexity, and #1 just too much paranoia... Then again, I can't really trust the hundreds of rsync mirrors out there...
Besides, when you speak about Reiser4, you talk like it is the end-all of all filesystems with the efficiency of a database... Then again, maybe the performance gain is from the algorithms used to search real databases.
| Quote: | 3. a proper gtk2 frontend
4. per package USE flags instead of global flags (handled in the before mentioned GTK2 frontend) |
#4 doesn't make sense. +gtk2 turns on GTK2 support in all apps where GTK1 support would have been taken otherwise. I don't want to have to +mozilla-gtk2 +moz-firebird-gtk2 ... to turn on GTK2 support everywhere. Unless I'm hideously misinterpreting it, which is likely, since everyone is pushing for that.
| Quote: | | 11. gtk USE flag for gtk1 deps, gtk2 USE flag for gtk2 deps - it should be that simple, but it's not - this is just SILLY ! |
+gtk turns on any kind of GTK+ support.
+gtk2 turns on GTK+2 support only if GTK+1 would have been used instead.
The people who think -gtk +gtk2 will result in a GTK+1 free system need to RTFM more. Besides, even with a -gtk1 flag, XMMS will still try to dump GTK+1 onto your system... Masking is the only way!
| Code: | | echo '<x11-libs/gtk+-2' > /etc/portage/package.mask |
|
|
| Back to top |
|
|
|
Gentoo Chat
