Let me reframe the question, then I'll go join the devel list...
So, I will take the suggestion to go participate on the devel list...
And I understand the reluctance to just keep stuff that isn't tested...
And, it even makes sense to see what people scream about, after all, there is an old saying, if you don't wear it for a year, throw it away...
But, I think it is fair to re-frame the question, in a _rescue_ context.
For rescue, one often finds oneself suddenly with _only_ one's rescue media. Having the latest kernel isn't so important, but being able to quickly and easily do something-or-other *at all* is important.
So the question is, we've been talking about what 11.12 removes. For _rescue_, what does it _add_ over 11.5? EFI boot support is a candidate for something that would matter, but cleanup, mDNS, auto-startups, build stuff for grml, zsh config, terminal fonts, etc . - doesn't really matter.
For me personally, I use grml for rescue and reformats and recoveries. Often I boot virtual machines with it, also, to do something-or-other to a copy of a running machine. I'm not kidding that the most common thing I do boils down to chains of netcat/tar/cpio/lzop/gzip/buffer/nc to move around disk images or virtual machines, and having both buffer and lzop even considered for removal gives me pause that maybe this tool is going in a direction different than what my uses and needs are.
So, for me, the question is really simple - 2011.05 is a pefect fit for my current needs - should I just take that as the gold-standard final- release and not look back? 2011.05 is working _fine_ for me - even a couple of hours to learn how to build a -full myself, never mind the hours to become active, is more driven by idle curiosity and a desire to support the tool than it is by common sense. Common sense for me would be, 2011.05 ain't broke and I'm unlikely to have hardware in the next 5 years that cares for newer kernels. Common sense would be to slide the write protect tab on a few USB sticks and figure you've painted your masterpiece and it is done.
The question is, what reasons are there to upgrade? Aside from it being smaller and advantages arising directly from it being smaller?
-Tom
* Tom {Tomcat} Oehser tom@toms.net [111228 14:29]:
[...] So, for me, the question is really simple - 2011.05 is a pefect fit for my current needs - should I just take that as the gold-standard final- release and not look back? 2011.05 is working _fine_ for me - even a couple of hours to learn how to build a -full myself, never mind the hours to become active, is more driven by idle curiosity and a desire to support the tool than it is by common sense. Common sense for me would be, 2011.05 ain't broke and I'm unlikely to have hardware in the next 5 years that cares for newer kernels. Common sense would be to slide the write protect tab on a few USB sticks and figure you've painted your masterpiece and it is done.
This is exactly the problem. By your common sense, we should not have released 2011.12 at all, and probably just stopped releasing completely.
You say - rightly so - that becoming active would cost you lots of hours. Staying active costs _us_ lots of hours. Sure, I can apply common sense here, too: stop the loss of hours, cut the expenses. Hope that someone else will take over.
Might just do that.
-ch
On Wed, Dec 28, 2011 at 04:32:46PM +0100, Christian Hofstaedtler wrote:
This is exactly the problem. By your common sense, we should not have released 2011.12 at all, and probably just stopped releasing completely.
Or, slow drastically. Just release kernel upgrades with new hardware. Occaisional enhancements. Stop following busy upstreams. Freeze a lot.
You say - rightly so - that becoming active would cost you lots of hours. Staying active costs _us_ lots of hours. Sure, I can apply common sense here, too: stop the loss of hours, cut the expenses. Hope that someone else will take over.
Might just do that.
There is active, and then there is active.
-Tom
Grml mailing list - Grml@ml.grml.org http://ml.grml.org/mailman/listinfo/grml join #grml on irc.freenode.org grml-devel-blog: http://blog.grml.org/
* Tom {Tomcat} Oehser tom@toms.net [111228 20:21]:
On Wed, Dec 28, 2011 at 04:32:46PM +0100, Christian Hofstaedtler wrote:
This is exactly the problem. By your common sense, we should not have released 2011.12 at all, and probably just stopped releasing completely.
Or, slow drastically. Just release kernel upgrades with new hardware. Occaisional enhancements. Stop following busy upstreams. Freeze a lot.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
That is exactly what we've done.
(Note that we can't just keep the packages at an old version as the release is always based on top of Debian testing/unstable.)
-ch
* Tom {Tomcat} Oehser wrote [28.12.11 14:28]:
But, I think it is fair to re-frame the question, in a _rescue_ context.
So the question is, we've been talking about what 11.12 removes. For _rescue_, what does it _add_ over 11.5? EFI boot support is a candidate for something that would matter, but cleanup, mDNS, auto-startups, build stuff for grml, zsh config, terminal fonts, etc . - doesn't really matter.
First of all it is not only about rescue but also installation. Meaning mDNS can be quite useful for remote access the machine itself.
The other things are mainly to make the environment viable and enhance it. Sure we could have removed iceweasel completly (which we did btw. in one version) but we re-added it.
For me personally, I use grml for rescue and reformats and recoveries. Often I boot virtual machines with it, also, to do something-or-other to a copy of a running machine. I'm not kidding that the most common thing I do boils down to chains of netcat/tar/cpio/lzop/gzip/buffer/nc to move around disk images or virtual machines, and having both buffer and lzop even considered for removal gives me pause that maybe this tool is going in a direction different than what my uses and needs are.
I personally never used lzop. And you can work without buffer as you said before. And I immediatly added lzop after you complained. I don't see that your requirements are totally different than our direction.
to become active, is more driven by idle curiosity and a desire to support the tool than it is by common sense. Common sense for me would be, 2011.05 ain't broke and I'm unlikely to have hardware in the next 5 years that cares for newer kernels. Common sense would be to slide the write protect tab on a few USB sticks and figure you've painted your masterpiece and it is done.
That is always a valid option.
The question is, what reasons are there to upgrade? Aside from it being smaller and advantages arising directly from it being smaller?
Never change a running system -> None. If you don't need anything/want any changes from the current Grml version none. Though if you stubmle upon a bug we may need you to upgrade to the latest version ;0
Ulrich
Teilnehmer (3)
-
Christian Hofstaedtler -
Tom {Tomcat} Oehser -
Ulrich Dangel