lexik|m has quit [Quit: Connection closed for inactivity]
sparetire_ has quit [Quit: sparetire_]
Kabouik has quit [Read error: No route to host]
Kabouik has joined #neo900
b1101 has quit [Quit: b1101]
Pali has joined #neo900
paulk-collins has joined #neo900
paulk-collins has quit [Ping timeout: 245 seconds]
paulk-collins has joined #neo900
infobot has quit [Ping timeout: 250 seconds]
paulk-collins has quit [Ping timeout: 240 seconds]
b1101 has joined #neo900
arcean has joined #neo900
Pali has quit [Ping timeout: 272 seconds]
Pali has joined #neo900
paulk-collins has joined #neo900
phre4k has joined #neo900
lexik|m has joined #neo900
Openbot has joined #neo900
mvaenskae has joined #neo900
mvaenskae has quit [Quit: leaving]
mvaenskae has joined #neo900
phre4k has quit [Remote host closed the connection]
<Openbot>
~Ping
<Openbot>
no pong ?Something is borked on my side but what ?
mvaenskae has quit [Client Quit]
SylvieLorxu has joined #neo900
Openbot has quit [Quit: used jmIrc]
mvaenskae has joined #neo900
mvaenskae has quit [Quit: leaving]
mvaenskae has joined #neo900
illwieckz has quit [Ping timeout: 240 seconds]
paulk-collins has quit [Ping timeout: 264 seconds]
paulk-collins has joined #neo900
<wpwrak>
happy 2015, sydney !
<DocScrutinizer05>
\o/
* DocScrutinizer05
idly wonders if openbot at least understands the concept of bots at large. Them being basically normal users that can't act when not around
<DocScrutinizer05>
and yeah, Tim's DSLAM obviously again went south
<mvaenskae>
hm, has ccc ended already?
<DocScrutinizer05>
should?
<bencoh>
I think it ends/ended today
<DocScrutinizer05>
:nod:
<mvaenskae>
yay, talks to listen to while studying system programming \o/
<bencoh>
waitamin, how can you listen a talk and study something else ?
<DocScrutinizer05>
is there finally a binding/decent definition of what exactly means "system programming"? - as e.g. opposed to application programming and particular system architecture?
<mvaenskae>
bencoh: multitasking ;)
<bencoh>
impressive ;p
<mvaenskae>
DocScrutinizer05: in our course we were going low level with C and learning the basics on it, writing optimised code in theory, learning about hardware-oriented thinking and exploring the linker-debugger world a bit :)
<mvaenskae>
i assume it is better than the average joe's lecture as the professor is working on his own research OS called barrelfish
<DocScrutinizer05>
aaah ok. That's fine then. Though I thought those were basically elementary skills for *all* decent programming ;-)
<DocScrutinizer05>
prolly my difficulties to understand "system programming" classification are due to me starting programming in a time when there was no other way to program computers that that
<DocScrutinizer05>
;-)
<mvaenskae>
no problem :) it is good to ask what "system programming" really refers to as it is very vast :)
<DocScrutinizer05>
in times of C64 and before, everybody was a system programmer. Or a hobbyist
<mvaenskae>
i will definitly use the neo900 and use it to learn arm-based programming :)
<DocScrutinizer05>
particularly when you programmed minicomuters and medium scale IT
<bencoh>
what do you mean by "arm-based" programming ?
<bencoh>
arm-specific asm ? neon optim ? :)
<mvaenskae>
bencoh: learn more about the arm world :)
<bencoh>
oh
<DocScrutinizer05>
"minicomputer" was everthing that fitted into a normal living room
<mvaenskae>
DocScrutinizer05: true, back in the days computers were massive and analog to some degree, right?
<DocScrutinizer05>
those were my professional first steps on programming
<mvaenskae>
bencoh: i have only experience with normal PCs but having an arm device which i can explore will teach me about different setups (those not including PCI e.g.) and also it will make me want to understand how to code better than what i do now as it is of limited capabilities :)
<bencoh>
:)
<mvaenskae>
the only thing lacking is multicore but that is fine for me, first learn single-core, then go multicore and worry about performance there and bash your head on the desk where your racecondition is occuring :D
<mvaenskae>
or why everything deadlocks ;)
<DocScrutinizer05>
duh, not really, even before there was Schneider CPC64, CPM80, and proprietary video processing hardware
<bencoh>
there are not-so-expensive multicore arm-based devices all around :)
<mvaenskae>
bencoh: but none are as awesome as the neo900 ;)
<mvaenskae>
and even fewer have such a great team developing the device :)
<kerio>
could you even process that amount of data with 37.5 MHz
<DocScrutinizer05>
hardly :-)
<DocScrutinizer05>
but the beast had convenient "intelligent peripherals"
<DocScrutinizer05>
which basically had their own little CPU and did a lot of stuff autonomously
<DocScrutinizer05>
sending a file from a harddisk via a PROMEA or the H1 bus maybe not needed CPU at all, once it been initiated
<kerio>
neat
<DocScrutinizer05>
PROgrammable Multi (E)Input/(A)Output card
<mvaenskae>
so a rudimentary dma system was exisiting back then?
<DocScrutinizer05>
not only rudimentary
<mvaenskae>
Eingang/Ausgang... looool, never bring german terminology into computing plz D:
<bencoh>
:]
<DocScrutinizer05>
harddisl was PLSK Plattenspeicher Lesen Schreiben Kombiniert
<DocScrutinizer05>
;-P
<DocScrutinizer05>
Siemens
<mvaenskae>
DocScrutinizer05: oh god, that sounds... just wrong
<DocScrutinizer05>
hehehehe
<DocScrutinizer05>
how about Kellerspeicher?
<DocScrutinizer05>
(stack)
<DocScrutinizer05>
please don't force my poor brain what been their term for program counter
<DocScrutinizer05>
force to remember...*
<DocScrutinizer05>
well, I guess with nax 6 chars per filename and 4 per device you have few options for sane naming, at least iside BS-M OS
<DocScrutinizer05>
NB there been no directories at all
<DocScrutinizer05>
there however been "libraries" which were sort of virtual directories, like QAP.AA been the "inode" and QAP.AB .. QAP.ZZ the 625(?) files in that "dir"
<DocScrutinizer05>
those were max 8 chars per virtual file iirc
<DocScrutinizer05>
QAP was "our" Quelle allgemain Pascal, for example
<DocScrutinizer05>
we did a complete CIM of a whole Siemens factory with that stuff
<DocScrutinizer05>
GWE, Geraetewerk Erlangen
<DocScrutinizer05>
and other non-Siemens companies
<DocScrutinizer05>
a few years later GWE received an award for their CIM
<DocScrutinizer05>
I maybe shouldn't mention that Siemens forced us to code that stuff in COBOL X-P
<DocScrutinizer05>
but for our own tools we resorted to semi-sane Pascal instead
<DocScrutinizer05>
tools like data dictionary etc, needed to turn that whole COBOL mess into sth that somehow actually works to build a system
<DocScrutinizer05>
you know, COBOL itself comes with a rather limited preprocessor ;-)
<DocScrutinizer05>
actually so limited I think it didn't have any
<DocScrutinizer05>
so I wrote one. In Pascal
paulk-collins has quit [Quit: Quitte]
<DocScrutinizer05>
EWOS, Eierlegende WollmilchSau
<DocScrutinizer05>
;-)
<DocScrutinizer05>
(yes we wrote in all CAPITALS backl when, all the time)
<kerio>
well, at least you had letters!
<DocScrutinizer05>
yeah, been freshly invented
<DocScrutinizer05>
actually we even alrady used real screens :-)
<mvaenskae>
DocScrutinizer05: kellerspeicher was a very foreign concept until i translated it to stack and most of the operations began to make sense quite quickly
<DocScrutinizer05>
indeed. as with a lot of that stuff back when
<DocScrutinizer05>
poor frenchmen
paulk-collins has joined #neo900
<DocScrutinizer05>
they MUST use french words for everything
<kerio>
hold on, doesn't it *mean* stack?
<DocScrutinizer05>
another two that come to mind, of our "libraries" we used in that project: QAA and OAA. Quelle and Object Assembler Allgemein
<DocScrutinizer05>
kerio: well, the translation for sure is not matching
<DocScrutinizer05>
but yeah, Siemens (and others) meant stack when they used the term Kellerspeicher iirc
<DocScrutinizer05>
and worst of all: they never provided a "dictionary" to translate their weird stuff into proper international (english) terms. Up to you to understand what they meant when using such weird words
<DocScrutinizer05>
even "Byte" was sort of forbidden, iirc those were Oktets
<DocScrutinizer05>
the manuals were mere "fun" to read
<DocScrutinizer05>
gave you headache
<DocScrutinizer05>
but a good training. Since then no manual really can shock me anymore
<bencoh>
yet octet != byte :p
<DocScrutinizer05>
well, luckily this was a 8/16 bit machine
<DocScrutinizer05>
not a 9bit or 15 or 18 bit arch
<mvaenskae>
12bit == 1B
<mvaenskae>
or let's take 13 for the fun of it :D
<DocScrutinizer05>
yeah
<DocScrutinizer05>
there been quite a number of such abominations
<DocScrutinizer05>
gosh, that sentimental age old stuff. This been a human's lifespan back. Probably more than 50% of chan members not been born back when
<kerio>
how many years?
<DocScrutinizer05>
early 80s of last century
<kerio>
so... 35 years ago
<DocScrutinizer05>
:nod:
<kerio>
1.75 kerios
* bencoh
wraps DocScrutinizer05 in naphtalene and closes the box
<DocScrutinizer05>
hehe
<bencoh>
there there
<DocScrutinizer05>
;)
<DocScrutinizer05>
[2014-12-31 Wed 15:19:43] [Notice] -NickServ- Information on DocScrutinizer51 (account jOERG_rw):
<DocScrutinizer05>
the idea being "XY is for old style phase difference probing. Nowadays you do this by pressing a key that triggers a math function" - maybe related: you can block the complete "AUTO" parameter automatic setup, and such lock can _not_ get unlocked on scope, only per remote control. Sounds pretty much like educational feature
<kerio>
what would you use XY mode for?
<DocScrutinizer05>
lightshow? ;-D
<kerio>
rofl
<DocScrutinizer05>
music looked good with L/R->X/Y back in the 1970s
<kerio>
also can you do XYRGB ?
<kerio>
nvm it makes no sense
<kerio>
at least, for an analog oscilloscope
varu|zZz has quit [Quit: .]
<DocScrutinizer05>
XY on Rigol has exactly _two_ input channels, out of the 4 available ones
<DocScrutinizer05>
the only thing you can adjust is attenuation/sensitivity of those 2 channels, sampling rate, and general screen brightness (and possibly you may edit some lables shown on screen)
<kerio>
did they ever make a color oscilloscope?
<DocScrutinizer05>
all DSOs are color today
<DocScrutinizer05>
the true cathode ray scopes usually were green phosphor (though I *think* they buolt a few very exotic things, maybe even RGB scope CRTs)
<DocScrutinizer05>
you know stuff gets serios when on the first 4 pages RFC2119 fullquote shows up
<DocScrutinizer05>
wut? is "EXPERIMENTAL" an amendment to original RFC that Zimmerman and friends made up?
<DocScrutinizer05>
~rfc2119
<infobot>
from memory, rfc2119 is Key words for use in RFCs to Indicate Requirement Levels. S. Bradner. March 1997. (Format: TXT=4723 bytes) (Also BCP0014) (Status: BEST CURRENT PRACTICE)
<infobot>
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.
<DocScrutinizer05>
happy this doesn't need update yet
<bencoh>
because it keeps disconnecting/reconnecting
<Pali>
but why to #maemo channel? DocScrutinizer05 is not there...
<DocScrutinizer05>
that once triggered +V (voice) autoresponder in my client since we had no access to channel's access list to set +V on bot
<bencoh>
now the poor thing is orphaned and voiceless :p
<DocScrutinizer05>
*shrug*
<DocScrutinizer05>
ask estel, he will know what to do about it ;-P
<DocScrutinizer05>
when somebody wants changes to infobot's config, they are free to ping me (or Tim)
<Pali>
but that message is send by infobot
<DocScrutinizer05>
anything beyond that, I think the mood was like "how do we get DocScrutinizer05 out of his positions where he could do any administration re maemo at large" - so up to whomever it may concern now to do whatever they think is approprioate to #maemo channel
<DocScrutinizer05>
I'm absolutely sure win7mac will happily raise his hand when it comes to IRC chanops for #maemo
<DocScrutinizer05>
Pali: the bot sends the message, my client answered with /msg chanserv voice infobot
<DocScrutinizer05>
I had the power to voice the bot, but not to set bot's accesslist flags to include autovoice
<DocScrutinizer05>
Pali: since I stepped down from maemo admin at large, I won't change bot's behavior either, until somebody from #maemo channel asks me to do
<DocScrutinizer05>
it's not anymore on the list of things I care about
<drathir>
happy new year for all...
<DocScrutinizer05>
and to you too
<kerio>
liar
<kerio>
you're CET
<kerio>
we've still got 4 hours left of 2014 :v
<DocScrutinizer05>
happy 2012 to kerio!
<kerio>
<3
<bencoh>
which reminds me ... 2015 is delorean year
<mvaenskae>
2012...
<mvaenskae>
wow, that is...
<mvaenskae>
that happened a long long time ago
<DocScrutinizer05>
well, kerio seems to prefer best wishes for things that definitely are _not_ in the future
<kerio>
2015 is delorean year, indeed
<kerio>
hopefully it's also neo900 year
<mvaenskae>
i will take a neo900 over a delorean any day of 2015 ;)
<bencoh>
I hope it wont need as much energy :p
<DocScrutinizer05>
depends on model of flux capacitor
<bencoh>
:D
<kerio>
will we need to shove organic matter in the stylus slot to make it work?
<wpwrak>
that may be what urban english speakers think is the correct word :) for the rest of us, it's still "chica"
varu|zZz has joined #neo900
<DocScrutinizer05>
wpwrak: do you have "Slow Sweep" mode in your Rigol? It's described like it should be roll mode for left side of trigger point, while right side (after trigger) gets written once on trigger event
<DocScrutinizer05>
manual claims "Slow Sweep" kicks in automatically instead of standard sweep on <=200ms/div
<DocScrutinizer05>
do you see that mode on your scope? does it work as advertised?
<wpwrak>
naw, mine doesn't have such fancy stuff :(
<DocScrutinizer05>
it never rolls anything in slow sweep mode
<DocScrutinizer05>
it simply updates left "half" in an instant, on trigger event or every now and then on "auto" mode
<DocScrutinizer05>
left half pops up and it instantly continues writing right half
<DocScrutinizer05>
that's pretty much nonsensical
<wpwrak>
yes, mine has that, too. but that's just what you'd expect to happen
<wpwrak>
no, makes perfect sense
<DocScrutinizer05>
not according to above manual note, no?
<DocScrutinizer05>
how does that make sense?
<wpwrak>
left side is buffered, so it's displayed immediately. the right side gets drawn as some passes (which is visibly slow, since you're at a low speed)
<DocScrutinizer05>
why doesn't it update buffer 50 times per second?
<DocScrutinizer05>
s/update/display/
<wpwrak>
(slow sweep) reading you manual screenshot, this is in fact what i see. so yes, i have it :)
<DocScrutinizer05>
I don't see *any* roll mode or sth similar to it in slow sweep
<wpwrak>
roll and slow seem to be two different things
<DocScrutinizer05>
roll mode means the screen content shifts left
<wpwrak>
slow seep is just a "normal" sweep, but slow enough that you can see what happens
<DocScrutinizer05>
the manual clearly states "slow sweep is similar to roll mode"
<wpwrak>
(well, chances are that the implementation is different. but in terms of visual effect, it's just the general case of a sweep)
<wpwrak>
sure, fancy lines appear on the screen ;-)
<DocScrutinizer05>
no
<wpwrak>
think of it: before trigger: scope records data. after trigger: scope shows data as it arrives.
<wpwrak>
this is what you get in slow sweep. at a fast timebase, the "show data as it arrives" is just so fast that it looks instantaneous
<DocScrutinizer05>
well, that's neither an explanation nor exactly a contradiction to a proper slowsweep that looks like lefts side roll mode
<wpwrak>
(plus, you have a gap between slow speed and regular display where it's displayed all at once but with a noticeable delay. so that's how you can tell that the implementation is different)
<wpwrak>
rigol seem to think what they have is proper enough ;-) and i would agree. just don't pay too much attention to the manual :)
<DocScrutinizer05>
I would expect to see left side rolling out of screen while new data gets displayed at trigger line, until trigger event happens. Then stop rolling and instead move point of writing from "center" (trigger line) to right end of screen
<DocScrutinizer05>
and aiui the manual, that's what they say it should do
<DocScrutinizer05>
and it would make perfect sense
<wpwrak>
they probably saw the similarity in that the display changes during the sweep. as opposed to - at a fast timebase - capturing a buffer full and only then generating screen content
<wpwrak>
btw, in my scope the "slow sweep" threshold is 20 ms, not 200 ms. 200 ms sounds fairly long.
<DocScrutinizer05>
meh, they could do slow sweep at fast sweep rate and it wouldn't make a difference for fast sweep
<DocScrutinizer05>
200ms is prolly waht you get from 24MB memory
astr has quit [Ping timeout: 255 seconds]
<DocScrutinizer05>
not even in singleshot trigger it updates the screen before trigger even happens
<DocScrutinizer05>
which would be mad useful if it would do
<wpwrak>
well, that's what you have roll mode for :)
<DocScrutinizer05>
no, in roll mode no triggering
<wpwrak>
it's slow, so you can trigger by eye ;-)
<DocScrutinizer05>
that's evidently a flaw in their feature specs
<DocScrutinizer05>
LOL
<DocScrutinizer05>
pathetic, Rigol!
<wpwrak>
i think pretty much everyone does it that way
<DocScrutinizer05>
anyway thanks for confirming that yours doesn't work as well
<wpwrak>
what you seem to be looking for could perhaps be described as "auto-trigger that ends when a real trigger happens"
<wpwrak>
alas, that's not quite possible in the general case, given that scopes have a dead time between captures
<DocScrutinizer05>
err wut?
<wpwrak>
you want to see what's happening while you don't have a trigger yet, but when the trigger occurs, you want the scope to stop and show you the result, right ?
<DocScrutinizer05>
no
<DocScrutinizer05>
I would expect to see left side rolling out of screen while new data gets displayed at trigger line, until trigger event happens. Then stop rolling and instead move point of writing from "center" (trigger line) to right end of screen
<wpwrak>
(or like ...) the description is correct. that's exactly what happens. just ignore the first sentence.
<DocScrutinizer05>
why ignore the first sentence? because Rigol forgot to implement display update?
<wpwrak>
sort of a split roll mode. ah well, will be a nice feature when you make the joergscope :)
astr has joined #neo900
<DocScrutinizer05>
they already *do* such continuous display update: in roll mode
<wpwrak>
the manual doesn't say that slow sweep is rolling. it only says it's "similar to roll mode". which is true, for some definition of "similar"
<DocScrutinizer05>
please elaborate
<wpwrak>
roll mode: display gets updated before capture ends. slow sweep: idem. ergo similar.
<DocScrutinizer05>
roll mode: there is no end of capture
<wpwrak>
sure there is. eventually proton decay will end it :)
<DocScrutinizer05>
to me that's a) cheating the customer, and b) poor design not doing absolutely feasible, expectable, logical way to implement stuff
<wpwrak>
anyway, you should check the missile launch facilities / anti-missile shields, and the booze. you don't want the end of the year to catch you by surprise :)
<wpwrak>
;-)))
<DocScrutinizer05>
meh
<DocScrutinizer05>
my windows are closed, booze is at zilch (even no milk for coffe anymore), and the only rocket stuff would be to pull out my 9mm
* DocScrutinizer05
swtiches off Rigol again
NIN101 has quit [Quit: No Ping reply in 180 seconds.]
NIN101 has joined #neo900
<wpwrak>
someone seems to already dislike 2015 :) but i'll be great. after all, the cross sum of 2015 is 8, "wealth" in chinese :)
<DocScrutinizer05>
doing accounting, should get finished THIS year
<DocScrutinizer05>
poor Nik is waiting
<mvaenskae>
wpwrak: i will dislike 2014 if i can't get ieee 754 deciphered D:
<mvaenskae>
i am soo confused as to when what happens with the numbers; no wonder noone understands floats
<bencoh>
:D
<bencoh>
ieee754 isnt that complex
<bencoh>
(well, the details can get tricky)
NIN101 has quit [Remote host closed the connection]
NIN101 has joined #neo900
* DocScrutinizer05
shoukd finally prepare the april's fool hoax about integer fraction 32bit and 64bit format
<DocScrutinizer05>
s/fraction/ratio/
<DocScrutinizer05>
M/N with M, N small ints of variable bitcount, packed together into a dword
<DocScrutinizer05>
or word
<mvaenskae>
DocScrutinizer05: what? i don't understand D:
paulk-collins has quit [Quit: Quitte]
<DocScrutinizer05>
dddddSMMMMMMMMMMMNNNNNNNNNNNNNNN; dddd [0..32]: <number of N bits> -1; S: sign; M[*n]: n bits for dividend M, with n = 26-number of N bits; N[*n]: n bits for divisor, with n=ddddd+1
<DocScrutinizer05>
actually dddd [0..24]
<DocScrutinizer05>
>24 invalid
<DocScrutinizer05>
reserved for overflow and other out-of-range or NAN or whatever
<DocScrutinizer05>
year in bcd suggests it's at least 99
<DocScrutinizer05>
so we got a few more years
<DocScrutinizer05>
mvaenskae: (superblock) LOL
<DocScrutinizer05>
AmigaOS forward-adjusted system clock to whatever date it found in metainfo of whatever file ;-)
<mvaenskae>
DocScrutinizer05: ext4 does that
<DocScrutinizer05>
I wonder why I even recall that, a proper cmosclock was very first thing I added
<DocScrutinizer05>
next was a Radio-controlled clock, on joystick-II port
<DocScrutinizer05>
of course internal
<DocScrutinizer05>
and the decoder for the DCF77 signal was my first own program on Amiga
<DocScrutinizer05>
luckily that OS was multitasking, so all the timing wasn't any problem
<DocScrutinizer05>
mvaenskae: any raationale given for this behavior?
<DocScrutinizer05>
sounds quite silly
<DocScrutinizer05>
man fs doesn't give away much
<dos1>
DocScrutinizer05: well, almost - it's october 2015 as seen in back to the future 2 :)
<mvaenskae>
DocScrutinizer05: well, it refused to mount the system during bootup for me and caused errors
<mvaenskae>
that was on archlinux
<DocScrutinizer05>
you got some of them errors for me?
<DocScrutinizer05>
yay, you learn sth new every day! \o/
<DocScrutinizer05>
man mount:
<DocScrutinizer05>
-r, --read-only
<DocScrutinizer05>
Mount the filesystem read-only. A synonym is -o ro.
<DocScrutinizer05>
Note that, depending on the filesystem type, state and kernel behavior, the system may still write to the device. For example, Ext3 or ext4 will replay its journal if the filesystem is dirty. To prevent this kind of write access, you may want to mount ext3 or ext4 filesystem with "ro,noload" mount options or set the block device to read-only mode, see command blockdev(8).
* DocScrutinizer05
found a can of beer \o/
<DocScrutinizer05>
CHEERS!
<dos1>
cheers!
<DocScrutinizer05>
LOLOLOL!!
<DocScrutinizer05>
check=none or nocheck
<DocScrutinizer05>
No checking is done at mount time. This is the default. This is fast. It is wise to invoke e2fsck(8) every now and then, e.g. at boot time. The non-default behavior is
<DocScrutinizer05>
unsupported (check=normal and check=strict options have been removed).
<kerio>
no way that's true
<kerio>
hold on, which checking?
<DocScrutinizer05>
there's exactly one option which 'enables' default behavior which is the only available behavior
<DocScrutinizer05>
kerio: fsck
<DocScrutinizer05>
aiui
<DocScrutinizer05>
man mount
<bencoh>
hm, that noload (or rather the lack thereof) is evil
<DocScrutinizer05>
actually I shouldn't have truncated quote. the best part is the end I left out: Note that these mount options don't have to be supported if ext4 kernel driver is used for ext2 and
<DocScrutinizer05>
bencoh: indeed
<DocScrutinizer05>
I would've killed a HDD during recovery when I hadn't read this negilgible lil detail
<kerio>
ah yes, you have to specify an extra option to actually mount an extfs read only
<DocScrutinizer05>
couldn't those fools make -o ro exactly what it means: NO write at all. And have a ro,load option if you want to replay journal?
<kerio>
DocScrutinizer05: it's not always feasible, but you should probably do a 1:1 copy before doing shit
<kerio>
ro is not necessarily for recover
<DocScrutinizer05>
kerio: yes, sure
<bencoh>
kerio: lvm snapshots :)
<bencoh>
those who dont use lvm in a (non-embedded) linux env deserve pain :p
<bencoh>
seriously though, it might not be possible to mount the fs without replaying logs
<DocScrutinizer05>
o.O
<bencoh>
but I'd expect it to SCREAM and tell me it cant
<kerio>
DocScrutinizer05: to be completely fair
<kerio>
with data writeback you don't even know if the data was written before or after the log
<bencoh>
DocScrutinizer05: mounting it ro without replaying logs would actually mean "losing" data
<bencoh>
or even worse, being out of sync with a later proper mount
<DocScrutinizer05>
I'm aware of the implications of journalling :-)
<bencoh>
but then again, I'd expect it to warn me, not to do it in my back
<bencoh>
:)
<DocScrutinizer05>
mounting a device (true) ro means it won't change in any way. so you can do whatever needed during next mount, when you want to write to the device
<bencoh>
I guess that's were they diverge from us
<kerio>
how do you "warn" without destroying
<kerio>
either you error, or you destroy data
<DocScrutinizer05>
the fs driver should be able to read journal and provide correct data from "virtual replay" of journal. No need to clean/rollout the journal
<bencoh>
kerio: just warn and error
<kerio>
spitting out "warning: i'm mounting read only but i still wrote shit lololol"
<kerio>
DocScrutinizer05: you can use an external journal i think
<bencoh>
haha
<DocScrutinizer05>
kerio: yes, journal can live on a separate volume even