| View previous topic :: View next topic |
| Better customization options |
| FreeBSD |
|
20% |
[ 2 ] |
| Gentoo |
|
80% |
[ 8 ] |
|
| Total Votes : 10 |
|
| Author |
Message |
Lifeonfull
n00b
Joined: 22 Dec 2014
Posts: 4
|
 Posted: Mon Dec 22, 2014 1:51 am Post subject: FreeBSD vs Gentoo tinkering and customization capabilities |
 |
|
Please also mention how?
Also, try to be as least biased as you can on a gentoo forum... |
|
| Back to top |
|
 |
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Mon Dec 22, 2014 5:09 am Post subject: |
|
|
This is a dumb poll, as you pointed out yourself.
Dubious given that you've just opened an account specifically to post this, too. |
|
| Back to top |
|
|
Ant P.
Watchman
Joined: 18 Apr 2009
Posts: 6920
|
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54237
Location: 56N 3W
|
| Posted: Mon Dec 22, 2014 6:20 pm Post subject: |
|
|
Lifeonfull,
Welcome to Gentoo.
You need to evaluate both and make your own decision
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
Lifeonfull
n00b
Joined: 22 Dec 2014
Posts: 4
|
| Posted: Mon Dec 22, 2014 6:59 pm Post subject: |
|
|
| steveL wrote: | This is a dumb poll, as you pointed out yourself.
Dubious given that you've just opened an account specifically to post this, too. |
As is every first post. |
|
| Back to top |
|
|
Lifeonfull
n00b
Joined: 22 Dec 2014
Posts: 4
|
| Posted: Mon Dec 22, 2014 7:03 pm Post subject: |
|
|
I posted that to get answers from both sides, honestly. |
|
| Back to top |
|
|
Wallsandfences
Guru
Joined: 29 Mar 2010
Posts: 378
|
| Posted: Mon Dec 22, 2014 11:18 pm Post subject: |
|
|
I don't get the hostile attiude by some towards a new user.
Welcome, Liefonfull, and thanks for the interesting question. Maybe both OS are not comparable, but in the end, one has to make a decision. Given that both OS have their limits, both have their specific user scenarios where they make sense.
For me, the outstanding feature of gentoo is, that you get the knowledge in the process as well as are given the tools to customize it to your needs to a very fine degree. There is a barrier, though, and that is the barrier of my/your own knowledge and skill. Shift it and you can do more... (than me...)
Rüdiger |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Tue Dec 23, 2014 11:09 am Post subject: |
|
|
| Lifeonfull wrote: | | As is every first post. |
Most people don't attach a poll to their first post.
Hence "dubious" since it looks like trolling. |
|
| Back to top |
|
|
khayyam
Watchman
Joined: 07 Jun 2012
Posts: 6227
Location: Room 101
|
| Posted: Tue Dec 23, 2014 5:23 pm Post subject: |
|
|
hehe ... I'm not sure I'd see "[n]ot knowing Gentoo at all [...] I'd still like to give an opinion" as "thorough", some of the criticism/opinion cited seems to be mostly due to lack of knowledge of the subject. Anyhow, such can probably be expected as the question is somewhat vague, and asking such an ill framed question to each respective camp is not likely to provide anything but "bias". That said, if we had some idea of what kind of "tinkering and customization" was needed, what the use case is, and if the level of T&C trumps other considerations, then we might be able to answer with something more substantial than our particular "bias".
Generally well framed questions illicit better responses, and asking "and how" means that for the OP's meagre input responders should provide gold ... which is a lot to expect.
So, here is some old iron for old paper ... not necessarily focused on "T&C".
Zirias on the FreeBSD forums stated "[...] FreeBSD is a complete OS in one project", this is correct, and because of this there is generally a greater level of continuity, so if there is a tool for configuring the network (ifconfig) then this tool provides a consistent means to do this. In linux such consistency is often lacking (some tools are not maintained to have full functionality, but replaced and/or spread across different tools ... such as we see with sys-apps/net-tools, sys-apps/iproute2, net-wireless/iw, net-misc/bridge-utils, etc). FreeBSD (and other BSD's) wins hands down in this regard, its frustrating that basic tools are not improved, updated, etc, with the "OS", and/or consistent with the idea of a "networking tool", rather than separate, duplicated, replaced by, etc, other tools. So, there is a tighter focus in FreeBSD on this consistency, though this doesn't necessarily go any further than the OS.
In the case of linux, the "linux", or "OS", part is just a kernel, the rest is made up of tools that may, or may not, have some consistency with this "OS" (ie, in the case of "ifconfig" this was ported in the early days from BSD's networking tool of the same name) and the "OS" is more the child of the "distribution" (more on that later). This can be seen as both positive and negative, so for example, FreeBSD "OS" has libc, whereas a linux distribution (or, a metadistribution like gentoo) might use glibc, musl, uClibc, or some other c library, one can replace the other ... if that's what your idea of "tinkering and customization" is all about.
WRT gentoo as a {meta}distribution, the decisions made by "the distribution" as to what exactly constitutes the "OS" is mostly left to the user. Obviously certain things need to happen in order for the "OS" to function, or be useful, but exactly how the "OS" is put together is left to the ingenuity of the user. So, for example, its quite possible to build a minimal "OS" for a specific purpose and a Raspberry Pi (ARM) target, or what-have-you, its also possible to build a Gentoo/FreeBSD.
Then there is a the simple question of the "kernel" itself, linux tends to have better hardware and/or feature support, so for example though FreeBSD has "geli" (hard disk encryption) its not possible to suspend/hibernate and resume from such a disk (at least the last I looked). In fact, FreeBSD's support for hibernation (suspend-to-disk) was somewhat problematic (again, not sure this is still the case).
So, I wouldn't argue that gentoo has it right, or that FreeBSD doesn't have a lot in it favour, in fact I'm the first to say that gentoo is broken in many regards (and yet, here I am). A lot depends on what your needs and expectations are so you are the only one that can evaluate which of the two fits your needs. I'll end by plagiarising the mutt (mail client) slogan "all OSes suck, this one just sucks less".
best ... khay |
|
| Back to top |
|
|
hasufell
Retired Dev
Joined: 29 Oct 2011
Posts: 429
|
| Posted: Tue Dec 23, 2014 6:09 pm Post subject: Re: FreeBSD vs Gentoo tinkering and customization capabiliti |
|
|
| Lifeonfull wrote: | Please also mention how?
Also, try to be as least biased as you can on a gentoo forum... |
Very good question.
I see two main differences:
First, as quoted from the freebsd thread, the following is totally true
| Quote: | | It [Gentoo] allows for (and the community, to an extent, seems to encourage) a lot of tedious, unnecessary tweaking of things that make no real difference in the long run. |
Not every developer agrees that this is a good idea in all circumstances and this attitude has bitten as more than once, introducing features which were not fully thought through.
Second: gentoo tries to be non-interactive. This isn't particularly a quality difference, more a style difference. But it's a big difference.
However it doesn't always work out as well as we thought it would. It's theoretically non-interactive in the sense that you don't have to configure packages in-between since everything must be in place before the package manager will start actually compiling. However, before that happens you often have to do a lot of work, resolving missing USE flag depends and whatnot. These sort of problems are sometimes easier to solve with an interactive approach. In addition, the interactive approach can be extended to do non-interactive, reproducible installs, but I'm not sure how well that works on FreeBSD (check out portconf).
All in all I found FreeBSD to be less consistent when it comes to package management tools. You can theoretically almost do anything you can on gentoo, but have to use more custom scripts, different tools etc. to mimic some functionality that's long integrated in the main package managers of gentoo. |
|
| Back to top |
|
|
khayyam
Watchman
Joined: 07 Jun 2012
Posts: 6227
Location: Room 101
|
| Posted: Tue Dec 23, 2014 7:06 pm Post subject: Re: FreeBSD vs Gentoo tinkering and customization capabiliti |
|
|
| hasufell wrote: | First, as quoted from the freebsd thread, the following is totally true
| Quote: | | It [Gentoo] allows for (and the community, to an extent, seems to encourage) a lot of tedious, unnecessary tweaking of things that make no real difference in the long run. |
Not every developer agrees that this is a good idea in all circumstances and this attitude has bitten as more than once, introducing features which were not fully thought through. |
hasufell ... well, without qualification of what this "tedious, unnecessary tweaking" involves I'm not sure I can agree on its "truth". That just reads to me as example of the "anecdotal fallacy". Sure, "introducing features which were not fully thought through" constitutes a problem, but this is, so to speak, a problem in itself, a problem that I'm not sure how to link to "a lot of tedious, unnecessary tweaking of things". Had they said the use of package.env, virtuals, eselect, or what-have-you, "encourages [things] that make no real difference in the long run" there might be something to argue against, or agree with, but as it stands its just anecdotal ... and vague ... (with a generous portion of everyone's favourite black bile, yawwwn, the "tedious" and "unnecessary").
best ... khay |
|
| Back to top |
|
|
hasufell
Retired Dev
Joined: 29 Oct 2011
Posts: 429
|
| Posted: Tue Dec 23, 2014 7:39 pm Post subject: Re: FreeBSD vs Gentoo tinkering and customization capabiliti |
|
|
| khayyam wrote: | | hasufell wrote: | First, as quoted from the freebsd thread, the following is totally true
| Quote: | | It [Gentoo] allows for (and the community, to an extent, seems to encourage) a lot of tedious, unnecessary tweaking of things that make no real difference in the long run. |
Not every developer agrees that this is a good idea in all circumstances and this attitude has bitten as more than once, introducing features which were not fully thought through. |
hasufell ... well, without qualification of what this "tedious, unnecessary tweaking" involves I'm not sure I can agree on its "truth". That just reads to me as example of the "anecdotal fallacy". Sure, "introducing features which were not fully thought through" constitutes a problem, but this is, so to speak, a problem in itself, a problem that I'm not sure how to link to "a lot of tedious, unnecessary tweaking of things". Had they said the use of package.env, virtuals, eselect, or what-have-you, "encourages [things] that make no real difference in the long run" there might be something to argue against, or agree with, but as it stands its just anecdotal ... and vague ... (with a generous portion of everyone's favourite black bile, yawwwn, the "tedious" and "unnecessary").
best ... khay |
It's lack of consensus about what degree of configurability (e.g. USE flags) justifies non-trivial work, bugs, abstractions and features (e.g. eclasses or even PMS additions).
One example is libsdl. Most of it's USE flags don't add any real value, because the USE cases for disabling them are SO rare that people who really need to turn all of it off can and probably will compile it locally anyway (e.g. for some embedded development use case).
Another example is games.eclass which allows you to manually configure all games directories. This is an utmost pain to support, because you have to patch almost every game in order to make it compatible with that eclass.
Another example is the overcomplicated abstraction of python eclasses with the main purpose of allowing you to install one library for all 3.x interpreters. Sure, everyone thinks it's cool, but it's a lot of magic going on under the hood and caused a lot of work, bugs etc. and may also cause very complicated dependency resolver errors which are not easy to understand because it relies on a lot of USE flag constraints.
Other distros are doing just fine without it and just support py2 and py3 without point-releases. Last time I spoke with some python team members, it seems they regret some of these decisions as well.
In terms of tedious, we now have REQUIRED_USE which was ONLY meant as a method to not break reverse deps of libraries. People are now using it all over the place to generally express dependencies between build system options (in non-libraries), requiring the user to micro-manage USE flags and again make the dependency resolver error output more complicated.
In terms of simply wrong we have stuff like app-admin/eselect-lua which is supposed to increase configurability. It can and will break stuff, because the PM doesn't really know what's going on there (the opengl eselect module already caused a lot of bugs). A similar mess is app-admin/eselect-cblas.
Generally, we are supporting a lot of extreme corner cases without knowing if anyone even makes use of them and if so, how many.
Since gentoo doesn't really have a review-workflow and there is barely consensus about what the spirit of ebuilds should be... you have all sorts of things in the tree, from clean strict ebuilds with reasonable amount of options to heavily hacked ebuilds with tons of random options, "because we can" (and keep in mind that this also heavily increases ebuild writing complexity... e.g. most of the time libsdl dependencies are wrong in new ebuilds, because it's not easy to figure out which USE flags are actually needed).
And no... not every distro is that asynchronous. |
|
| Back to top |
|
|
krinn
Watchman
Joined: 02 May 2003
Posts: 7470
|
| Posted: Tue Dec 23, 2014 8:24 pm Post subject: |
|
|
I share the sentiment this poll is stupid:
1/ of course freebsd users should tell you freebsd is better AND the same for gentoo users speaking about gentoo, no poll need for that.
2/ the only valid answer you could get would be from one user that use them BOTH : and you don't even need to read his answer, just on what forum he post his answer to guess his preference.
You can ask weakness or value of gentoo, but asking all gentoo users to speak about gentoo VS a distro they only know by name is dumb.
hasfufell: you're lost 
If anyone wants judge a distro, generally speaking, it have nothing to do with dev work. It "might" be relatvie to dev work result, but not dev work.
So if anyone ask if gentoo is a good or bad distro, he doesn't care or make any sense to know any dev trouble with this or that ; as long as the result of the dev make the distro fine.
Nobody cares if it's harder to build a ferrari or a porsche, when someone ask what is the best of a ferrari vs porsche he doesn't expect to get answer that screws are easier to set on ferrari but porsche is better because parts are easier to reach.
Just like your use example ; the value of the distro doesn't depend on how much time you need to create an ebuild with 5 use flag ; but the value of the distro could be seen out of that result (as your ebuild offer 5 switch to user, making a worthier value than an ebuild with only 1 use flag and less flexibility).
And if you even speak about it, yes in gentoo for a dev it's harder sometimes as you must add plenty choices (or not) to user, but it is certainly not any easier for a binary distro dev that MUST take the choice for the user ; and face the case he mistake and all users where expecting the package to have this or that use flag enable or disable. Harder to create the package because of more choice to user, but far easier as the result should get far less unhappy users. |
|
| Back to top |
|
|
hasufell
Retired Dev
Joined: 29 Oct 2011
Posts: 429
|
| Posted: Tue Dec 23, 2014 9:08 pm Post subject: |
|
|
| krinn wrote: | If anyone wants judge a distro, generally speaking, it have nothing to do with dev work. It "might" be relatvie to dev work result, but not dev work.
So if anyone ask if gentoo is a good or bad distro, he doesn't care or make any sense to know any dev trouble with this or that ; as long as the result of the dev make the distro fine.
Nobody cares if it's harder to build a ferrari or a porsche, when someone ask what is the best of a ferrari vs porsche he doesn't expect to get answer that screws are easier to set on ferrari but porsche is better because parts are easier to reach.
|
I cannot follow you or your analogy. Gentoo and FreeBSD have quite a lot in common actually, so comparing both involves looking at details.
| krinn wrote: | | Just like your use example ; the value of the distro doesn't depend on how much time you need to create an ebuild with 5 use flag ; but the value of the distro could be seen out of that result (as your ebuild offer 5 switch to user, making a worthier value than an ebuild with only 1 use flag and less flexibility). |
You missed the point. If the 5 additional switches increase the maintenance work of the user (e.g. because the PM is confused or you are *required* to micro-manage USE flags), cause more build failures and bugs... then that's a detail you cannot simply ignore. The main point was not that it makes ebuild writing more complicated. It has a lot of side effects for the user as well.
| krinn wrote: | | And if you even speak about it, yes in gentoo for a dev it's harder sometimes as you must add plenty choices (or not) to user, but it is certainly not any easier for a binary distro dev that MUST take the choice for the user ; and face the case he mistake and all users where expecting the package to have this or that use flag enable or disable. Harder to create the package because of more choice to user, but far easier as the result should get far less unhappy users. |
I've never used FreeBSD as a binary distro, so I cannot comment on that. And my previous post wasn't related to those parts.
edit: or are you suggesting that I answer the poll question without explaining the side effects? That would be kind of dull, especially because
| Lifeonfull wrote: | | Please also mention how? |
|
|
| Back to top |
|
|
krinn
Watchman
Joined: 02 May 2003
Posts: 7470
|
| Posted: Wed Dec 24, 2014 11:54 am Post subject: |
|
|
| hasufell wrote: | | I cannot follow you or your analogy. Gentoo and FreeBSD have quite a lot in common actually, so comparing both involves looking at details. |
Assuming you are not understanding it with good faith...
The point about the analogy to compare the porsche and the ferrari is that people expect answer from user not mechanical.
So answer should be about comfort, performance (speed or power of engine), or some less obvious effect of the car "for a user" like "what car help you the most to get babes".
But nobody cares or expect answer about the technician difficulty to handle one or the other engine or car parts ; but people may not stay blind to something like "the engine was customize by <plop> (and <plop> is a famous engine motor tweaker)". So none care if <plop> has any trouble to do it, but it doesn't mean nobody cares about <plop> result (as you expect <plop> to gave better result than a stock engine). Hence none cares about gentoo dev point of view about gentoo dev work, while still get interest about the result of that work.
| hasufell wrote: | | You missed the point. If the 5 additional switches increase the maintenance work of the user (e.g. because the PM is confused or you are *required* to micro-manage USE flags), cause more build failures and bugs... then that's a detail you cannot simply ignore. |
It's a real difference between the two distro handling of package ; why not something like "gentoo give more choice to user but that freedom prize is more complexity for user ; while distro using binary package gives simplicity to user at the prize of loosing choice".
While it's just true that it cause more complexity to handle USE flags for a user, the increase in bugs and failure is false, if devs are doing their work, the failures and bugs count should be set to 0.
But it has nothing to do for user, while a package may have 0 bugs or failures to build and gentoo dev has done their work ; still the choice gave to user can lead to "problem" when package manager face an impossible choice to take by automation. Like when a user try to install package X that must use ffmpeg while user (or package manager) has decide previously to install libav to use it with package Z.
Except because of that problem it doesn't create more failure or bugs, package X or package Z are working as fine in gentoo as in other distro. And i'm sure distro using binary package have also bugs not related to the package itself that is working fine, but just because of the choice set by the dev when he build it for the user.
So if only package Z or package Y could be install in a system, gentoo gave user ability to use one or the other and of course it increase complexity to user because he must do the choice himself, sometimes for package Z/Y he doesn't even know anything about.
But it is also not good that the simplicity of binary distro lead to only having Z or Y in its repo because the dev distro pickup the choice to use Z for the user ; as all users cannot use Y then. |
|
| Back to top |
|
|
hasufell
Retired Dev
Joined: 29 Oct 2011
Posts: 429
|
| Posted: Wed Dec 24, 2014 1:48 pm Post subject: |
|
|
| krinn wrote: | | The point about the analogy to compare the porsche and the ferrari is that people expect answer from user not mechanical. |
I don't think the thread starter expected an incomplete answer. In fact I initially gave one... but then a user in this thread asked for more details.
I think you are confused. |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Wed Dec 24, 2014 4:45 pm Post subject: Re: FreeBSD vs Gentoo tinkering and customization capabiliti |
|
|
| hasufell wrote: | | It's lack of consensus about what degree of configurability (e.g. USE flags) justifies non-trivial work, bugs, abstractions and features (e.g. eclasses or even PMS additions). |
That's one of the things about working in an ecosystem; everyone else is allowed to play too. Sure you could go the route of letting someone else do the thinking for you, as seems to be attractive to so many nowadays, but that just leads to cults of personality and having to kowtow to some moron who refuses to admit a mistake.
| Quote: | | One example is libsdl. Most of it's USE flags don't add any real value, because the USE cases for disabling them are SO rare that people who really need to turn all of it off can and probably will compile it locally anyway (e.g. for some embedded development use case).. most of the time libsdl dependencies are wrong in new ebuilds, because it's not easy to figure out which USE flags are actually needed |
I see, so this is all too hard because none of you cba to clean up the base ebuild, as you clearly have an idea about. On top of which you guys simply can't "figure out" dependencies for your own packages. Oh my.
| Quote: | | Another example is games.eclass which allows you to manually configure all games directories. This is an utmost pain to support, because you have to patch almost every game in order to make it compatible with that eclass. |
Yeah the horror. Poor distro devs having to actually patch things to make them play nicely on their end-users' machines; what kind of weird mess is this? /s
| Quote: | | Another example is the overcomplicated abstraction of python eclasses with the main purpose of allowing you to install one library for all 3.x interpreters. Sure, everyone thinks it's cool, but it's a lot of magic going on under the hood and caused a lot of work, bugs etc. and may also cause very complicated dependency resolver errors which are not easy to understand because it relies on a lot of USE flag constraints. |
Well I'll give you that one. python and multilib (which increases the nonsense scripting) are awful; not everyone thinks they're cool. Though again a problem of your own making (where "you" == the developer collective.)
| Quote: | | In terms of tedious, we now have REQUIRED_USE which was ONLY meant as a method to not break reverse deps of libraries. People are now using it all over the place to generally express dependencies between build system options (in non-libraries), requiring the user to micro-manage USE flags and again make the dependency resolver error output more complicated. |
I see, so again, you and your fellow developers are making a boo-boo. Begs the question why you're bleating on the forums about it.
| Quote: | | Since gentoo doesn't really have a review-workflow and there is barely consensus about what the spirit of ebuilds should be.. |
Ah yes, let's bring ciara in instead; at least the trains will run on time (or something.)
| Quote: | | And no... not every distro is that asynchronous. |
Funny, exherbites were making a big deal of how you can just clone and build, and gentoo moves too slow, yadda yadda, and here we are now with a plea for moving slower. Perhaps you should talk to your fellow exherbo-devs about it. |
|
| Back to top |
|
|
hasufell
Retired Dev
Joined: 29 Oct 2011
Posts: 429
|
| Posted: Wed Dec 24, 2014 4:56 pm Post subject: Re: FreeBSD vs Gentoo tinkering and customization capabiliti |
|
|
| steveL wrote: | | hasufell wrote: | | It's lack of consensus about what degree of configurability (e.g. USE flags) justifies non-trivial work, bugs, abstractions and features (e.g. eclasses or even PMS additions). |
That's one of the things about working in an ecosystem; everyone else is allowed to play too. Sure you could go the route of letting someone else do the thinking for you, as seems to be attractive to so many nowadays, but that just leads to cults of personality and having to kowtow to some moron who refuses to admit a mistake.
|
uh, lack of consensus is not a result of working in an ecosystem
| steveL wrote: | | Quote: | | One example is libsdl. Most of it's USE flags don't add any real value, because the USE cases for disabling them are SO rare that people who really need to turn all of it off can and probably will compile it locally anyway (e.g. for some embedded development use case).. most of the time libsdl dependencies are wrong in new ebuilds, because it's not easy to figure out which USE flags are actually needed |
I see, so this is all too hard because none of you cba to clean up the base ebuild, as you clearly have an idea about. |
If you want to know why libsdl is not cleaned up, ask the games team... oh wait. They won't answer. duh
You are just a bit uninformed, but that's ok.
| steveL wrote: | | On top of which you guys simply can't "figure out" dependencies for your own packages. Oh my. |
I probably forgot that I was mainly referring to user submitted ebuilds or those devs not regularly dealing with games ebuilds and not caring to get a review.
For both, there are solutions and I have presented them often enough.
| steveL wrote: | | Quote: | | Another example is games.eclass which allows you to manually configure all games directories. This is an utmost pain to support, because you have to patch almost every game in order to make it compatible with that eclass. |
Yeah the horror. Poor distro devs having to actually patch things to make them play nicely on their end-users' machines; what kind of weird mess is this? /s |
You mean like games.eclass causing security bugs, because it's misdesigned? You are uninformed again.
| steveL wrote: | | Quote: | | Another example is the overcomplicated abstraction of python eclasses with the main purpose of allowing you to install one library for all 3.x interpreters. Sure, everyone thinks it's cool, but it's a lot of magic going on under the hood and caused a lot of work, bugs etc. and may also cause very complicated dependency resolver errors which are not easy to understand because it relies on a lot of USE flag constraints. |
Well I'll give you that one. python and multilib (which increases the nonsense scripting) are awful; not everyone thinks they're cool. Though again a problem of your own making (where "you" == the developer collective.) |
You are mistaken. There is no developer collective. I've explained that often enough too.
| steveL wrote: | | Quote: | | In terms of tedious, we now have REQUIRED_USE which was ONLY meant as a method to not break reverse deps of libraries. People are now using it all over the place to generally express dependencies between build system options (in non-libraries), requiring the user to micro-manage USE flags and again make the dependency resolver error output more complicated. |
I see, so again, you and your fellow developers are making a boo-boo. Begs the question why you're bleating on the forums about it. |
You probably don't read mailing lists. That's not my problem. Why should I care?
| steveL wrote: | | Quote: | | Since gentoo doesn't really have a review-workflow and there is barely consensus about what the spirit of ebuilds should be.. |
Ah yes, let's bring ciara in instead; at least the trains will run on time (or something.) |
Cool, so you are in troll mood again. |
|
| Back to top |
|
|
khayyam
Watchman
Joined: 07 Jun 2012
Posts: 6227
Location: Room 101
|
| Posted: Fri Dec 26, 2014 2:46 pm Post subject: Re: FreeBSD vs Gentoo tinkering and customization capabiliti |
|
|
| hasufell wrote: | | Generally, we are supporting a lot of extreme corner cases without knowing if anyone even makes use of them and if so, how many. |
hasufell ... but as krinn pointed out these examples don't really relate specifically to the users ability to "tinker and customize", they are much more reflective of a burden on developers and the development process. Additionally, the poster on the FreeBSD forums point was that "gentoo allows for, [and] seems to encourage, a lot of tedious, unnecessary tweaking", which, besides being unqualified, is a blanket statement about the entire spectrum of things "tediously" and "unnecessarily" open for "tinkering" ... or in short, what the user has the ability to effect. So, it may be the case that libsdl has useflags that no one might need to disable, or that REQUIRED_USE is misused, or that eselect-lua "breaks stuff", but this is an altogether different subject, or "gentoo", to the one in reference.
best ... khay |
|
| Back to top |
|
|
hasufell
Retired Dev
Joined: 29 Oct 2011
Posts: 429
|
| Posted: Fri Dec 26, 2014 3:10 pm Post subject: Re: FreeBSD vs Gentoo tinkering and customization capabiliti |
|
|
| khayyam wrote: | | hasufell wrote: | | Generally, we are supporting a lot of extreme corner cases without knowing if anyone even makes use of them and if so, how many. |
hasufell ... but as krinn pointed out these examples don't really relate specifically to the users ability to "tinker and customize", they are much more reflective of a burden on developers and the development process. Additionally, the poster on the FreeBSD forums point was that "gentoo allows for, [and] seems to encourage, a lot of tedious, unnecessary tweaking", which, besides being unqualified, is a blanket statement about the entire spectrum of things "tediously" and "unnecessarily" open for "tinkering" ... or in short, what the user has the ability to effect. So, it may be the case that libsdl has useflags that no one might need to disable, or that REQUIRED_USE is misused, or that eselect-lua "breaks stuff", but this is an altogether different subject, or "gentoo", to the one in reference. |
No, it describes the side effects of those customization capabilities, which are not always user friendly, but quite the opposite. And that answers the question of the topic and your own comment where you questioned that those side effects actually cause trouble (so in fact you feeded this sub-topic).
Also, I don't think you need to explain the topic to me. The thread starter can tell me if he didn't want to hear details about gentoo (which I highly doubt). |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Fri Dec 26, 2014 3:57 pm Post subject: |
|
|
For anyone who's new to Gentoo, I'd just like to apologise about hasufell. It appears he's having one of his semi-annual "moments".
For some reason he likes to pick on me whenever someone else points out his spamming, even though I've got nothing to do with the discussion.
Like I told you before hasufell: I'm not interested. Good luck with your obsessive crusade to make everyone else think the way you do.
Somehow I don't think you'd actually like the results if you got your way, but then you'd be too stubborn to admit a mistake, ime. |
|
| Back to top |
|
|
khayyam
Watchman
Joined: 07 Jun 2012
Posts: 6227
Location: Room 101
|
| Posted: Fri Dec 26, 2014 7:22 pm Post subject: Re: FreeBSD vs Gentoo tinkering and customization capabiliti |
|
|
| hasufell wrote: | | No, it describes the side effects of those customization capabilities, which are not always user friendly, but quite the opposite. And that answers the question of the topic and your own comment where you questioned that those side effects actually cause trouble (so in fact you feeded this sub-topic). |
hasufell ... you are placing the cart before the horse, or more precisely the cart before the parrot. Whatever is logically prior, or whatever the causal relationship, the user still has "capabilities" at their disposal, and it is these capabilities which constitute the capacity for "tinkering and customization" (and to what degree). You're bringing in some other dimension that is essentially unrelated to that "capacity" and so conflating two separate domains.
I challenged the FreeBSD'er "tedious [and] unnecessary" because it was unqualified, you now wish to make it seem that your "tedious [and] unnecessary" are the same as theirs (and/or redefine "tinkering and customization" to match your particular view of things) having shown that some useflags are (in practice) unused, or that REQUIRED_USE is abused by developers, etc, none of which has anything to do with the capacity to effect the outcome of the install process ... or, in short, "tinkering and customization".
If you want to circumlocute logical domains then the only recourse you leave me is to accuse you of being red ... as it makes as much sense logically.
best ... khay |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Gentoo Chat
