View previous topic :: View next topic |
Author |
Message |
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Sat Apr 23, 2016 9:21 pm Post subject: |
|
|
NeddySeagoon wrote: | I agree with all of that. When I said "just happen" I meant overnight with no further thought or other changes. |
Neddy ... I see, I was reading it in terms of the political will to make it happen, resistance, etc. In the above I'm referring back to previous discussions, specifically the claim that "the gentoo project" is not subject to the charter, what powers are derived from where, etc, etc.
NeddySeagoon wrote: | Feel free to join the Foundation and become trustees. |
Look at it this way: the charter has empowered me (as part of "the community") to act on behalf of the community, and I'm perfectly ok with the scope of the powers currently granted me. I could put myself forward for some role, including that of trustee, but I'm not limited in some way by not doing so, as I'm fully empowered to act within the currently existing framework without having to assume the role of trustee, or whatever. That is how the charter, and the community it posits, should function.
NeddySeagoon wrote: | If you look at the history of Gentoos metastructure, its never been fixed until it was demonstrably broken. The top level project leads, which were left in charge and without a leader when drobbins left, eventually failed to resolve technical issues. The council emerged from that as a solution to an immediate problem. |
I'm a political realist, I don't expect politics (or socio-political arrangements) to be perfect, its an ongoing, and fallible, human project. The criticisms (or observations) I make are first and foremost about shaping opinion, because it is by being of the same mind about the problem, terms of reference, etc, etc, that allow us to reflect on where we've been, and to shape what we do next. Understanding the problem, the mechanisms involved, etc, is central to being able to do anything in relation to them ... and that is my motivation. I understand the history of the council, and the various forces/pressures that feature in its development, but I'm focused more on the more general questions involved, namely how such a community can function. These things have an epistemology, and can be understood.
best ... khay |
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54550 Location: 56N 3W
|
Posted: Sat Apr 23, 2016 10:01 pm Post subject: |
|
|
khayyam,
Very perceptive. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Fri Apr 29, 2016 6:31 pm Post subject: |
|
|
axl wrote: | i would put all X stuff in usr. FOR MY SYSTEM if i had that... gentoo flexibility. |
steveL wrote: | Please, show me which X apps you have which install to '/'. |
axl wrote: | i never said i have X apps in root. I think you heard that i would move these X apps that are in root to usr. what i meant is that i prefer usr to be separate to distiguish X apps. i said i dont want them in root. |
I see; so exactly what we currently have, then.
I don't see what you want changed to effect this, nor why you want it changed, given that you already have your apps in /usr.
It seems that the only thing that's stopping anyone use it as flexibly as one might like, is the neophyte decision to merge /usr with rootfs, 2 or 3 years ago, the flawed premise for the current bindir merge.
Quote: | i WOULD merge libs, but not apps. for simple user stupid reasons. |
I disagree; the way to deal with "user stupid", is to educate your users when they do "stupid" things.
Given that we're talking about software development (the context was long LDFLAGS), though, that argument totally falls down.
If anyone should pro-actively educate themselves when they clearly know nothing about build-systems, it is software "developers."
Seriously, that kind of attitude to an issue of vanilla ignorance, would get such a verbal slap around here.
Quote: | again, i think a simple portage feature would make gentoo the most flexible distro and make everyone happy. package.bin_paths and package.lib_paths or something of the sort. |
I kinda agree with you there, though you should simply use package.env or perhaps bashrc, for the vast majority of autotools-alike cases (S?BINDIR and LIBDIR come to mind.)
RPi-userland you'd probably need to do with a bashrc, or simply mod the ebuild to respect any user-set PREFIX in the env, especially if it is from binary (which may entail fun with rpaths.)
And you might want to use a different variable name; by which time you've implemented your feature, and can start working it into eutils.eclass to start, since autoconf will respect the variables already though you might as well use autotools.eclass as your first instance, then move on to cmake and python.
Though you should be aware of the Gentoo Prefix project as well, that is what we call OS_PREFIX when we have to deal with it make-side, and not the same as PREFIX. (It's about EROOT or ROOT/OS_PREFIX in split terms, and PREFIX=OS_PREFIX or OS_PREFIX/usr etc.) |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Fri Apr 29, 2016 6:36 pm Post subject: |
|
|
with respect to the politicking, what krinn said:
krinn wrote: | If the trustees protect the foundation, what are they protecting if they are not protecting the gentoo project? The foundation exist only because the project exists, if your task is to protect the foundation, you de-facto get also the protection of the project itself.
It's like if anyone is saying : "i have the task to protect the head of that person", and when the person is about to jump under a train : "we couldn't do anything, we're only protecting the head and not the corpse" ; hey can't you see you will have no head to protect once that body has been under a train???
So if nobody has the task to protect the body, and you are in charge of protecting the head, sorry, but you are also in charge of protecting the body just to do your task ; protecting the head.
So until council have a legal existence, any technicals or not, decisions made fall under trustees power.
It's not trustees that should be scared council have no legal existence, but council members that should be scared it could be closed without notice. |
++
I think that's what is meant by the other meaning of "happen overnight"; any decision can be taken unilaterally on the part of the Trustees, at any point.
Such would never be ultra vires, as they are the legal authority in Gentoo.
I for one am not positing "so make the Council a legal entity then" as a solution; that just sounds like the kind of thing apparatchiks would love, and real developers would hate, iow: totally counter-productive.
Council is deliberately constituted solely to deal with technical matters about the ebuild tree. If they don't want to be reconsituted by the community and Trustees, then they should simply stop acting outside their mandate, limited by their own choice.
See: no need to fork ;) You only end up maintaining an overlay, anyhow. :P |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Fri Apr 29, 2016 6:48 pm Post subject: |
|
|
Tony0945 wrote: | I started using UNIX when AT&T was THE Telephone Company. I remember clearly a seminar where they were explaining their new programming language called C. |
Lucky b*sterd. ;-)
Some of us suffered C++ at Uni and VB then Java at work (the horror encumbent in those years cannot be overemphasized) and had to find C for ourselves much later.
awk was always cool, though.
/me narrowly avoids rambling about other languages. (web-coding is the pits.) |
|
Back to top |
|
|
Tony0945 Watchman
Joined: 25 Jul 2006 Posts: 5127 Location: Illinois, USA
|
Posted: Sat Apr 30, 2016 11:50 pm Post subject: |
|
|
steveL wrote: | Some of us suffered C++ at Uni and VB then Java at work |
Gawd! That must have been terrible! Had to use VB on one embedded project. The actual embedded controller was a Motorola 68000 programmed in assembly language. A PC sent it commands. NT 3.5.1 They had to pay $300 per license as 3.5.1 was passed it's expiration date. Some hacker used internal knowledge of NT to "speed things up". The same hacks didn't work in 4.0 Eventually M$FT refused to sell them licenses at ANY price and they had to bite the bullet and recode it. Luckily, I was on a different project for a different client by then. |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Mon May 02, 2016 1:23 pm Post subject: |
|
|
steveL wrote: | Some of us suffered C++ at Uni and VB then Java at work |
Tony0945 wrote: | Gawd! That must have been terrible! Had to use VB on one embedded project. |
Heh, I actually have a soft spot for VB (Option Explicit ofc), though that may be because I hadn't had wider exposure to UNIX as yet.
Certainly, I see it as a terrible language overall, nowadays, but its auto-completion is a nice thing.
Mainly it was just so much nicer than VC++ in terms of actually getting a result.
VB was useful for UI against embedded, I recall as well.
Reviewing the code for MFC was an eye-opener of the worst sort; so I went down a level, and it was even worse.
C++ is a bit like Java in that regard for me: if most uses are in fact awful to read, modify and maintain, then there's something wrong with the language.
I realise it may sound odd to like VB; it's because my background was assembler, so as far as I was concerned, they were all high-level, and it only came down to which I could reasonably write cleanly in, and still provide a Windows UI (in the days before DHTML.)
That, and VBA became a really useful option for integration, ie: it paid.
I really wish someone had pointed K&R out to me in the 80s; but back in those days, in the UK at least, the internet was not a thing as it is now (acoustic modems were, and there was JANET, or its predecessor, but not a household thing.)
As I am sure you recall, information used to be hoarded and you had to apprentice yourself in order to learn the privy secrets of most trades.
Unfortunately there wasn't really anyone around to talk computers to, apart from a very seldom occasion (and I didn't want to talk about assembler any more.)
I think that's where the cargo-cult comes from; I see it a lot now on IRC, where people simply do not take the initiative and follow-up on books (they don't even follow up on web-pages often enough), instead preferring to be told information in bite-sized chunks (over and over, until finally you lose your temper with them.)
Leaving aside the "ADHD" aspect, I think the social aspect goes back to however many millennia of obfuscation as an art-form.
Over two years to get a supposedly-accomplished, and professional, C coder to learn the basics about conversions, because he refused to buy a book, recommended over 5 times in that period.. insane.
I can deal with the neophytes not yet knowing that you simply must buy books if you want to become a programmer, but not professionals, in work and drawing a salary for their attitude. Researching the problem domain is essential.
It's the waste of everyone's time, hours of it, because matey cba to spend less than 15 minutes' salary on an essential reference, that pisses me off.
I tend to presume shared background of whatever I've advised someone to read, or at least the essential content, if they read it elsewhere: when they later turn out to have NFC what you're on about, because it comes up again and again, which is why you took the time to make sure they knew about it, that starts to rankle.
It's got nothing to do with age or "modern times" afaic, since plenty of younger people have the right nouse.
Though perhaps "modern" times do tend toward chattering idiocy, that has been the case in lay-society for quite a while (it's easier to divide-and-conquer people whose brains are full of propaganda and poison, rather than measured thought, so off-balance first.)
Quote: | Luckily, I was on a different project for a different client by then. |
Lulz; that naughty-guilty-relieved feeling of escaping a dud employer is delicious, ain't it? If it's not just me, that is. ;-) |
|
Back to top |
|
|
1clue Advocate
Joined: 05 Feb 2006 Posts: 2569
|
Posted: Mon May 02, 2016 5:15 pm Post subject: |
|
|
Back when I was in school it was FORTRAN and Pascal and various forms of Assembly language(mc68000, 8086, and the CDC mainframe we used). C didn't have enough of a market share yet, so they weren't teaching it unless you took a specific course for it. That course wouldn't have existed except due to student demand.
We even made (as in soldered chips onto a board) an 8086 computer in one semester and hand-assembled code for it in the next, punched in with a hexadecimal keypad. Required class for all majors. Heh. Most thought that was a ridiculous waste of time, but they were actually pretty neat boards if you dug into them. EE majors needed these things all through the undergrad program and often used them in grad work.
The mc68000 was unavailable except to universities or corporations. Modems still had acoustic couplers. A few places still used punch cards for real work.
Thinking back on those days it's hard to believe how much computer science has changed. |
|
Back to top |
|
|
Ant P. Watchman
Joined: 18 Apr 2009 Posts: 6920
|
Posted: Mon May 02, 2016 5:29 pm Post subject: |
|
|
steveL wrote: | Lulz; that naughty-guilty-relieved feeling of escaping a dud employer is delicious, ain't it? If it's not just me, that is. ;-) |
Far from it... and the last coding job I had can only be described as a "death march". Endless, needless rewriting of the world when they were 90% done because the higher-ups found some new pointless shiny to rebase everything on; sound familiar? |
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6059 Location: Removed by Neddy
|
Posted: Mon May 02, 2016 5:29 pm Post subject: |
|
|
1clue wrote: |
We even made (as in soldered chips onto a board) an 8086 computer in one semester and hand-assembled code for it in the next, punched in with a hexadecimal keypad. Required class for all majors. Heh. Most thought that was a ridiculous waste of time, |
And those are the type of CS I steer clear of. We had to "hand build" a Z80 in the 1st year. The 2nd year this was then used to improve the control of an SR motor. The SR was span in pure hardware but the Z80 was then used to improve the performance.
You need a firm understanding of what a register is, round trip time etc... the more turtles these new CS graduates are cushioned with, the more resources are just eaten out of poor code...
Its not limited to JAVA or C++... if an individual does not have an overview of what is going on some VERY bad directions can occur. Likewise REQUIREMENTS... the overall system needs to be validated, flowed down, linked & validated BEFORE real coding can take place. Then verification is needed and not to just to requirements but boundary conditions _________________
Quote: | Removed by Chiitoo |
|
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54550 Location: 56N 3W
|
Posted: Mon May 02, 2016 6:27 pm Post subject: |
|
|
Heh, You bairns.
Neither the microprocessor nor computer science was invented when I was in further education. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
1clue Advocate
Joined: 05 Feb 2006 Posts: 2569
|
Posted: Mon May 02, 2016 6:37 pm Post subject: |
|
|
Naib wrote: | 1clue wrote: |
We even made (as in soldered chips onto a board) an 8086 computer in one semester and hand-assembled code for it in the next, punched in with a hexadecimal keypad. Required class for all majors. Heh. Most thought that was a ridiculous waste of time, |
And those are the type of CS I steer clear of. We had to "hand build" a Z80 in the 1st year. The 2nd year this was then used to improve the control of an SR motor. The SR was span in pure hardware but the Z80 was then used to improve the performance.
You need a firm understanding of what a register is, round trip time etc... the more turtles these new CS graduates are cushioned with, the more resources are just eaten out of poor code...
Its not limited to JAVA or C++... if an individual does not have an overview of what is going on some VERY bad directions can occur. Likewise REQUIREMENTS... the overall system needs to be validated, flowed down, linked & validated BEFORE real coding can take place. Then verification is needed and not to just to requirements but boundary conditions |
Yes and no. The college I went to was an engineering college. Back when I went, the classes I took were pertinent to the real world. As I said, this was in the latter days of the punch card.
We had an extremely good idea about registers and round-trip times. In fact one of my professors thought round trip time was critical and would dock points if your round trip time wasn't the best it could be. THAT was pointless, he insisted on things like using rotate to multiply or divide if you knew it was a power of two. We had hardware-assisted knowledge of boundary conditions because physical devices would break if we got it wrong. We had compilers as an undergrad course. We knew about optimal vs non-optimal and when it was best to use each.
Basically it was about things that are mostly taken care of by compilers today. About the only thing I actually actively use from that time was Information Theory. Ordering your code to test for most likely occurrences first is one application of that, and if it's simple code the compiler does that too. |
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6059 Location: Removed by Neddy
|
Posted: Mon May 02, 2016 7:01 pm Post subject: |
|
|
"handled by the compiler" That is fair enough if the individual has a sound understanding of what it is meant to do, if the user does not however...
Two recent "synthesis tool bugs" (the firmware equiv of a compiler) produced some interesting results
1) synthesis tool misinterpreted a valid VHDL with regards to a memory pointer to block ram and thus "synthesised" out a load of logic -> never incremented.
2) in declaring a moore state machine using if/else resulted in an explosion in gate utilisation due to the cascade nature in an A3P
Then there is those that think they know what they are doing and code god awful stuff...
Combinational logic who's propagation delay is longer than a clock period & they had only provided flipflops at the start and end stages...
leaving it upto the compiler ends up with DSP's being over-utilised and process steps pulled out of the sequencer flow simply because those that coded chose a convoluted multiplier over a simple shift+add OR the concept of a "signed number/unsigned" was so alien at BIT level that some completely unneeded block of code was instantiated to "pack" an unsigned number onto an I/O that was "signed" ... rather than just packing the bits _________________
Quote: | Removed by Chiitoo |
|
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54550 Location: 56N 3W
|
Posted: Mon May 02, 2016 7:18 pm Post subject: |
|
|
Naib,
That's people not bing close enough to the hardware implementation detail.
Then there are those who test for equality between two floating point numbers only to be surprised that it works 'sometimes'.
Its all beginners problems.
When I were a lad, floats were too slow to be useful. It was all fixed point. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
Tony0945 Watchman
Joined: 25 Jul 2006 Posts: 5127 Location: Illinois, USA
|
Posted: Mon May 02, 2016 10:54 pm Post subject: |
|
|
NeddySeagoon wrote: | Heh, You bairns.
Neither the microprocessor nor computer science was invented when I was in further education. |
B.Sc. Physics, June 1967 Illinois Institute of Technology.
Did I beat you in decrepitude, Neddy?
EDIT:
We programmed in FORTRAN IV on an IBM 360 (model 30?) |
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6059 Location: Removed by Neddy
|
Posted: Tue May 03, 2016 6:38 am Post subject: |
|
|
NeddySeagoon wrote: | Naib,
That's people not bing close enough to the hardware implementation detail.
Then there are those who test for equality between two floating point numbers only to be surprised that it works 'sometimes'.
Its all beginners problems.
When I were a lad, floats were too slow to be useful. It was all fixed point. |
Floating point is still too slow and too large on FPGA fabric.
You still need an appreciation of what the bits are doing, what your compiler will do.
As per not being close enough to HW... The app notes only ever mentioned signed 16bit IO registers so when we were asking for unsigned 12bit they were getting confused and went down some convoluted route. They had already pre IO restricted it to 12b but we're getting hung up on the apps said the IO was signed 16b... The hardware at that point doesn't care... _________________
Quote: | Removed by Chiitoo |
|
|
Back to top |
|
|
depontius Advocate
Joined: 05 May 2004 Posts: 3509
|
Posted: Tue May 03, 2016 3:19 pm Post subject: |
|
|
Tony0945 wrote: | We programmed in FORTRAN IV on an IBM 360 (model 30?) |
That certainly predates me. But long ago, we were having the Quarter Century Club celebration for my manager, and someone brought in and gave him an old CCROS unit. That was one of the projects he'd worked on earlier in his career, which was used in the 360/30. It was an absurdly big device - about two feet tall, and had air bags inside.
EDIT - I meant to include this link: http://www.computerhistory.org/revolution/mainframe-computers/7/164/578 _________________ .sigs waste space and bandwidth
Last edited by depontius on Tue May 03, 2016 4:39 pm; edited 1 time in total |
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54550 Location: 56N 3W
|
Posted: Tue May 03, 2016 4:27 pm Post subject: |
|
|
Tony0945,
Looks like you do :)
By about five years. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
truekaiser l33t
Joined: 05 Mar 2004 Posts: 801
|
Posted: Sun May 08, 2016 1:54 am Post subject: |
|
|
Do you mind if i ask for a clarification.
Right now I am in the middle of installing gentoo on a machine that will be my NAS system and router/hotspot playground.
I read the news bit that this topic is talking about but I am confused as to if it is saying that I, with only two partitions. One for efi stub booting and one for the rest of the system drive, will need to mess with the beast that is intramfs. I call it a beast imho because its way to complicated and never worked right for me.. |
|
Back to top |
|
|
1clue Advocate
Joined: 05 Feb 2006 Posts: 2569
|
Posted: Sun May 08, 2016 3:21 am Post subject: |
|
|
You do not need an initramfs if your entire runtime operating system is on the same partition. |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Sun May 08, 2016 5:19 am Post subject: |
|
|
... and in a similar vein, the great openrc-run migration ... another case in which the solution was worse than the problem (and again I'll offer the challenge to anyone to find one single bug in which net-dialup/minicom's runscript could be said to require such a solution, and further reflection on the ratio of effort required to provide said "solution" by renaming openrc's runscript to openrc-run).
best ... khay |
|
Back to top |
|
|
frostschutz Advocate
Joined: 22 Feb 2005 Posts: 2977 Location: Germany
|
Posted: Sun May 08, 2016 6:25 am Post subject: |
|
|
truekaiser wrote: | I am confused as to if it is saying that I, with only two partitions. One for efi stub booting and one for the rest of the system drive, will need to mess with the beast that is intramfs. |
That's fine... if all you have is a plain unencrypted root partition, you don't need initramfs.
truekaiser wrote: | I call it a beast imho because its way to complicated and never worked right for me. |
It's a lot less complicated than it's made out to be. Usually it comes down to a shell script. But anything is complicated if you don't understand how it works. |
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Sun May 08, 2016 10:36 am Post subject: |
|
|
krinn ... oiiiiyyy, if sheer exasperation is what you mean then, yes ;) Speaking of which, on my musl install I wasn't able to build sys-apps/openrc-0.12.4 and so ended up downgrading from 0.19.1 to 0.17 (because I remember Neddy, and others, mentioning issues) ... for some reason (I haven't even begun to investigate) 'shutdown -h' hangs the system, no services seem to be stopped, no filesystems unmounted, and no shutdown. I tried switching vt but getty is dead so have to force power off. OK, musl is "experimental" but as I seem to remember others reporting similar issues with shutdown its probably not musl related at all. It's incredible this level of crap makes it into stable, yet there they are all busy migrating hashbangs for all the initscripts in the tree because they made the incredibly dumb decision to rename runscript.
best ... khay |
|
Back to top |
|
|
truekaiser l33t
Joined: 05 Mar 2004 Posts: 801
|
Posted: Sun May 08, 2016 5:57 pm Post subject: |
|
|
frostschutz wrote: | truekaiser wrote: | I am confused as to if it is saying that I, with only two partitions. One for efi stub booting and one for the rest of the system drive, will need to mess with the beast that is intramfs. |
That's fine... if all you have is a plain unencrypted root partition, you don't need initramfs.
truekaiser wrote: | I call it a beast imho because its way to complicated and never worked right for me. |
It's a lot less complicated than it's made out to be. Usually it comes down to a shell script. But anything is complicated if you don't understand how it works. |
okay.
Yea I understand that, though the last time i tried was back when I was still using a bootloader. Something that i don't think is needed on a uefi system unless you are booting multiple os's. If i gave it a try now I could get it to work using the stub method as well.
Still thank you for clearing that up, They should make the news item a bit clearer as well. |
|
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
|
|