<dovsienko>
obviously, password-based authentication must work in the first place, which usually requires adding "PermitRootLogin yes" to /system/settings/ssh/sshd_config
<dovsienko>
after the key is in place, that line can be removed
mmu_man has quit [Ping timeout: 480 seconds]
mmu_man has joined #haiku
<Halian>
o/
<adrian_>
hihi
x10z has quit [Quit: My Unrecognized Mac has gone to sleep. ZZZzzz…]
AD_Haiku_onPC has quit [Ping timeout: 480 seconds]
HaikuUser has joined #haiku
HaikuUser has quit [Quit: Vision[]: i've been blurred!]
janking has quit [Quit: Vision[]: i've been blurred!]
TristanRoss[m] has joined #haiku
mmu_man has quit [Ping timeout: 480 seconds]
kescher has quit [Quit: Bye]
kescher has joined #haiku
HaikuUser has joined #haiku
MisthaLu has joined #haiku
HaikuUser has quit []
xet7 has joined #haiku
xet7 has quit [Remote host closed the connection]
xet7 has joined #haiku
JakeSays has quit [Read error: No route to host]
probono98 has quit [Quit: Ping timeout (120 seconds)]
probono98 has joined #haiku
janking has joined #haiku
<Cacafuego>
I know I'm still a relative newby but I'm still floored at how well things like PyCharm are working
<janking>
Good morning
<coolcoder613>
'morning
nephele_xmpp has left #haiku [Error from remote client]
JakeSays has joined #haiku
HaikuUser has joined #haiku
<coolcoder613>
Hi HaikuUser
HaikuUser has quit []
netyuui has joined #haiku
AD_Haiku_onPC has joined #haiku
Cacafuego has quit [Quit: Vision[]: i've been blurred!]
janking has quit [Quit: Vision[]: i've been blurred!]
cp- has quit [Ping timeout: 480 seconds]
cp- has joined #haiku
cp-- has joined #haiku
cp- has quit [Ping timeout: 480 seconds]
HaikuUser has joined #haiku
HaikuUser has quit []
netyuui has quit [Quit: Vision[]: i've been blurred!]
janking has joined #haiku
janking has quit []
nicoco has joined #haiku
dpirate has quit [Ping timeout: 480 seconds]
janking has joined #haiku
AlienSoldier has quit [Quit: Vision[]: i've been blurred!]
Cacafuego has joined #haiku
HaikuUser has joined #haiku
HaikuUser has quit []
jmairboeck has joined #haiku
<nipos>
Anyone from the developers awake?
<nipos>
I was playing around with https://review.haiku-os.org/c/haiku/+/8384 a lot.I have a solution that looks clean and makes sense and all,but it doesn't actually work and I don't get why.Can anyone please look at my experimental diff at https://bpa.st/MBRA ?
<nipos>
The code should at least printf "called......" and the value whenever FormatToolTip() is called,but it never prints anything,so maybe it's not called at all?But it should...
<PulkoMandy>
your printf are missing a \n so they may not be showing up due to line buffering
<nipos>
Normally they do at least after quitting the application,but I'll try again with the \n
<nipos>
I tried it,makes no difference unfortunately
<nipos>
No console output and the tooltip doesn't show up at all.Do I miss any obvious mistake in the code?
* janking
is idle: BRB
DKnoto has quit [Ping timeout: 480 seconds]
dodo75 has joined #haiku
<PulkoMandy>
nipos: I don't see an obvious problem, I would guess your code is not loaded and some other version is running
<nipos>
Then it would show the tooltip with the too high value instead of none at all
DKnoto has joined #haiku
<nipos>
I'm LD_PRELOADing my own libbe.so and libmedia.so,one of them should contain the changes
<PulkoMandy>
You should copy them in a lib/ directory next to the executable you're testing. Ld_preload gives unreliable results
diver has joined #haiku
<nipos>
Oof,I found the mistake m(
<nipos>
With correct file paths,LD_PRELOAD would have worked perfectly fine
<nipos>
It's kits/media/libmedia.so and not kits/libmedia.so
dodo75 has quit [Quit: Vision[]: i've been blurred!]
<nipos>
Nevermind,I wasted a whole day looking for different solutions,not noticing that I was just loading a wrong library m( Now it's working exactly as it should,and pushed to Gerrit
dodo75 has joined #haiku
mmu_man has joined #haiku
deneel has joined #haiku
FreeFull_ has joined #haiku
FreeFull has quit [Ping timeout: 480 seconds]
gouchi has joined #haiku
MisthaLu has quit [Ping timeout: 480 seconds]
gouchi has quit [Quit: Quitte]
FreeFull_ has quit []
FreeFull has joined #haiku
* janking
has returned
x10z has joined #haiku
diver has quit [Quit: Leaving.]
janking has quit [Quit: Vision[]: i've been blurred!]
MisthaLu has joined #haiku
deneel has quit [Quit: deneel]
x10z has quit [Quit: My Unrecognized Mac has gone to sleep. ZZZzzz…]
diver has joined #haiku
nicoco has quit [Ping timeout: 480 seconds]
nicoco has joined #haiku
diver has quit [Read error: Connection reset by peer]
diver has joined #haiku
x10z has joined #haiku
x10z has quit []
gouchi has joined #haiku
<Hanicef[m]>
i did some measurements with nanosleep and found that haiku is actually more accurate than linux, achieveing 5-30 microsecond accuracy compared to ~50 microseconds that linux is capable of
<Hanicef[m]>
interesting to see how we've surpassed linux in one regard
<phschafft>
define accurate? and also how did you messure?
gouchi has quit [Remote host closed the connection]
<waddlesplash>
Hanicef[m]: it's possible that Linux would be more accurate in some other scheduler mode than whatever the default is
<waddlesplash>
we prioritize such things by default for GUI responsiveness
<waddlesplash>
that's worse for server type workloads though
<Hanicef[m]>
yeah, was thinking that the default kernel is the culprit, too
<Hanicef[m]>
<phschafft> "define accurate? and also how..." <- i wrote a small program for it, let me send you it, one sec
<phschafft>
waddlesplash: maybe. I mean it it is really hard to find defaults that work good in many cases. way harder than people think it is.
<phschafft>
Hanicef[m]: the reason for my question is both interest in the statistical model and if you messured the clock againt itself or against an external clock.
<Hanicef[m]>
basically, it takes the time, sleeps for as short time as possible, and measure the time that passed
<Hanicef[m]>
phschafft: is haiku using the same clock for sleep?
<phschafft>
Hanicef[m]: I mean that is what you request in your code.
<phschafft>
for reference you might also want to print the result of clock_getres() and compare it to that.
<Hanicef[m]>
fair enough, i guess i can make it more accurate
<Hanicef[m]>
even then, i did some research into this and it seems like the linux kernel is not using the wall clock for sleep even if you request it, which is why it's around 50 us
<phschafft>
the wall clock is like the worst idea for sleeping in most cases anyway.
<phschafft>
Hanicef[m]: please don't get me wrong that I'm a bit critical about your results. your tests sounds like a good starting point. but there are so many factors that there is just so much ahead to explore.
<Hanicef[m]>
yeah, i don't blame you
<Hanicef[m]>
it was just a fun experiment anyway, nothing serious
<phschafft>
one of my problems with what you do is that there is no relation to real time passing. that is something you can only messure if you can swiftly toggle some state that you can messure from outside the box.
<phschafft>
e.g. a super fast GPIO.
<Hanicef[m]>
does haiku allow using the cpu clock for measurement?
<waddlesplash>
but I think CLOCK_MONOTONIC might also?
<Hanicef[m]>
i used CLOCK_MONOTONIC for that, so yeah
<waddlesplash>
yes it does
<waddlesplash>
just calls system_time
<waddlesplash>
so it's the same
<Hanicef[m]>
either way, this isn't really a serious measurement, just a fun thing i did
<waddlesplash>
appears nanosleep() has only microsecond accuracy on Haiku tho
<Hanicef[m]>
yeah, i noticed that, too
<dovsienko>
waddlesplash: do you know if TSC is specific to each CPU core or the entire package?
<Hanicef[m]>
but it landed on around 10 microseconds, although not sure how accurate that is in real time
<waddlesplash>
I don't remember the details here, I think it's specific to a CPU core but we synchronize it at boot
<dovsienko>
also in systems with more than one CPU socket obviously there would be more than one TSC register, and it will likely be different
<Hanicef[m]>
which, thinking of it, can potentially interfere with the results if the cpu core changes during scheduling
<dovsienko>
I remember somebody somewhere saying their TSC reading went backwards, which seemed impossible until I considered the process could be hopping between different CPUs
<waddlesplash>
well it shouldn't since the TSCs should be synchronized
<waddlesplash>
but it can sometimes happen on VMs in odd circumstances I think
<dovsienko>
CPUs are not step-locked to a single oscillator tick, are they? the frequency can go up and down, different cores can go offline and online, so there's space for different TSC registers to become offset
<dovsienko>
so the point I am trying to make, for the purposes of system_time() and related functions would it make sense to get TSC from one specific CPU core at all times?
<Hanicef[m]>
isn't that what the CLOCK_PROCESS_CPUTIME_ID on linux is for?
<dovsienko>
"This is a clock that measures CPU time consumed by this process..." -- looks like a statistics of the scheduler to me
Babaj has joined #haiku
<Hanicef[m]>
yeah, i don't know this well enough :/
<dovsienko>
I am not well versed in time matters, but one of the few related cases I remember had to do with routing protocol implementations
<waddlesplash>
dovsienko: modern x86 machines have invariant TSC
<waddlesplash>
older ones it changes with frequency but it doesn't anymore
<dovsienko>
waddlesplash: I didn't know
<dovsienko>
the thing with routing protocols (and network protocols in general) is that the protocol instance has to make certain time-delayed moves by certain deadlines, and when the code uses wall clock time to set the timers, the timers fire after a wrong delay if the wall clock changes
<dovsienko>
(e.g. the protocol starts before the host has NTP-synchronised its wall clock, or the clock is ticking faster/slower to transition between daylight saving and no saving etc)
<dovsienko>
switching to monotonic clock eliminates these effects from the problem space: if the protocol peer is expecting you to send a message in 5.0 seconds, your timer fires in 5.0 seconds no matter what is going on with the wall clock
B2IA has quit [Quit: Vision[]: i've been blurred!]
<dovsienko>
most times hosts do not know what the wall clock is at the remote host, so protocol time is measured from event A to event/timeout B
B2IA has joined #haiku
janking has joined #haiku
janking has quit []
bbjimmy_64 has joined #haiku
bbjimmy has quit [Read error: Connection reset by peer]
bbjimmy_64 has quit [Quit: Vision[]: i've been blurred!]
bbjimmy has joined #haiku
Babaj has left #haiku [#haiku]
<jmairboeck>
oh well, here we go again: I'm once again building perl itself because we didn't set the default installation path for man pages in the perl configuration. These could be set in the individual recipes too, but I think it would be better to set them "globally", so they are automatically correct for all perl packages
<phschafft>
one for the wired and one for the wireless interface.
<Anarchos>
phschafft the 0.0.0.0 ?
<phschafft>
it doesn't report any cost setting or something. so they should both be equaliy valid from the kernel's perspective.
<phschafft>
yes.
<Anarchos>
yes and the wired one should not be used
<phschafft>
then you need to remove that route.
<Anarchos>
for an ethernet packet, 'thou shall not pass 'through the wire :)
<phschafft>
haha.
<dovsienko>
Anarchos: on the wired interface, if you configured a default gateway manually, then that contradicts what you write the configuration should be
<Anarchos>
dovsienko yes but in Network i could not erase the gateway
<dovsienko>
if the wired interface has got the default gateway from a DHCP server, then perhaps the DHCP server isn't configured correctly, in that if it sets a default gateway, the default gateway should be usable
erysdren has joined #haiku
<Anarchos>
dovsienko that's normal : my dhcp server is not powered up .
<dovsienko>
my Haiku VM has two interfaces, one is configured statically for IPv4 only, and it has no default gateway
<dovsienko>
that is, the "Gateway:" input in network preferences for that interface is empty and editable
<dovsienko>
Anarchos: what gets in the way in your configuration?
<Anarchos>
dovsienko i put a gateway manually in Network, and when i erase it , it comes back automatically
<dovsienko>
well, then this is a Haiku bug
HaikuUser has joined #haiku
HaikuUser has quit []
erysdren has quit [Quit: Konversation terminated!]
diver has quit [Read error: Connection reset by peer]
diver has joined #haiku
DKnoto has quit [Ping timeout: 480 seconds]
<phschafft>
Anarchos: you can try puting in 0.0.0.0 for me real quick and see if that changes things?
DKnoto has joined #haiku
<Anarchos>
phschafft 0.0.0.0 seems to work to erase the route.
<phschafft>
is it then gone in the GUI as well?
<phschafft>
maybe the input just doesn't like being empty or something.
<phschafft>
(still a bug, but that might make it easier to hunt it down :)
HaikuUser has joined #haiku
HaikuUser has quit []
<Anarchos>
phschafft yes but not sure if it is the command line 'route delete' which solved the trouble or not
<phschafft>
hm.
<phschafft>
I would say just put it all in the ticket so whoever will have a look will have a good set of hints
zardshard has left #haiku [Disconnected: Replaced by new connection]
zardshard has joined #haiku
<Begasus[m]>
jmairboeck want me to test run a build for the PR?
<jmairboeck>
yes please, if you can
x10z has joined #haiku
<jmairboeck>
and then run perl -V:installvendorman1dir to see if it worked correctly
<Begasus>
on that note, only man1 and man3 are used?
<jmairboeck>
ah, sorry, could have pushed it upstream instead :)
<jmairboeck>
the output of that command should then be something like /packages/perl-5.40.0-5/.self/documentation/man/man1 and not empty like it is now
<Begasus>
doesn't seem to pull the latest changes /
<jmairboeck>
the branch is called perl in my fork
<Begasus>
rebooting to my default install
Begasus has quit [Quit: Vision[]: i've been blurred!]
Begasus has joined #haiku
<Begasus>
launching
<jmairboeck>
make sure you have a clean work dir. perl is one of those annoying ones that don't work correctly when the build is repeated.
<jmairboeck>
at least I had problems with that before
<Begasus>
failed build with crashes
<Begasus>
there was no work dir around
<jmairboeck>
I guess the same as my system :(
<jmairboeck>
do you run beta 5?
<Begasus>
yeah
<jmairboeck>
then what is it that broke this?
<jmairboeck>
it has clearly worked before
<Begasus>
not seeing anything that changed for a while in the recipe (external deps)
<Begasus>
biab
<jmairboeck>
the crashes I had occur in strcmp, which is in libroot, so it comes from Haiku itself. Unfortunately, there is no useful stack trace
<dovsienko>
jmairboeck: what type of crash is it? SIGSEGV?
<scanty>
0x0000002f6b5ca694: 0f1f4000 nopl %eax, (%rax) -- definitely not why you crashed, but seeing code like this make me think what is the compiler doing... maybe it's there for alignment?
<dovsienko>
jmairboeck: if the two arguments to strcpy() are identical and do not contain 0x00 before strcpy() runs off the end of the allocated page, that's the expected behaviour now matter which implementation of strcmp()
<dovsienko>
in other words, could be a miniperl bug
<scanty>
well, it would help if we know what was in rdi and rsi
<scanty>
and maybe stick with strncpy() instead where applicable
<jmairboeck>
unfortunately I have no idea where it crashes because there is no stack trace
<jmairboeck>
and it's not my own code, obviously
<dovsienko>
is it realiably reproducible?
<jmairboeck>
yes
<dovsienko>
could be one of the libraries miniperl links with
<jmairboeck>
the stack only contains miniperl besides the top frame (which is strcmp)
<Begasus>
autodoc.pl is mentioned there in the crash window
<jmairboeck>
make minitest in the haikuporter chroot crashes the same
<jmairboeck>
in t/lex.t in this case
<jmairboeck>
t/base/lex I mean
Begasus has quit [Quit: Vision[]: i've been blurred!]
<Anarchos>
Begasus i can't make my mind around this KDiskDevice so i will ask waddlesplas in the review comment.
Begasus has joined #haiku
mmu_man has quit [Ping timeout: 480 seconds]
<Anarchos>
Welcome Page in WEb+ is broken ?
mmu_man has joined #haiku
<dovsienko>
jmairboeck: it may be a good idea to recompile miniperl using CFLAGS=-g or an equivalent, which may translate frame addresses to file names and line numbers
<dovsienko>
in my experience, when Perl compiles as a pkgsrc dependency, that always includes miniperl and always works, so I never had to debug it
<Begasus>
afk, TV :)
<jmairboeck>
miniperl is used to bootstrap perl, if I understood it correctly
<Anarchos>
Begasus i replied on my PR to waddl. Wait and see :)
MsInput has quit [Quit: WeeChat 4.5.1]
andreaa71 has joined #haiku
<Begasus>
closing down here
<Begasus>
cu peeps
Begasus has quit [Quit: Vision[]: i've been blurred!]
<andreaa71>
beos collection is of 2020 ...
<andreaa71>
news ? ;)
andreaa71 has left #haiku [#haiku]
<Halian>
???
arjen_ has quit [Quit: Vision[]: i've been blurred!]
<dovsienko>
jmairboeck: yes
HaikuUser has joined #haiku
HaikuUser has quit []
<scanty>
you can tell he just tried vision for the first time.
<Anarchos>
scanty lol
<Anarchos>
scanty as we all did :)
dby has joined #haiku
janking has joined #haiku
<scanty>
yeah, and i'm guilty of doing the same thing :^)
<scanty>
vision setup is a little bit clunky
x10z has quit [Quit: My Unrecognized Mac has gone to sleep. ZZZzzz…]
janking has quit [Quit: Vision[]: i've been blurred!]
vezhlys has joined #haiku
Anarchos has quit [Quit: Vision[]: i've been blurred!]
Anarchos has joined #haiku
OrngBomb has joined #haiku
OrangeBomb has quit [Ping timeout: 480 seconds]
MisthaLu has quit [Quit: Leaving]
x10z has joined #haiku
<jmairboeck>
dovsienko: I found the problem (after running make minitest outside the chroot because otherwise the debugger would still show nothing)
<jmairboeck>
there were build warnings about a missing declaration for memrchr (which is guarded with _DEFAULT_SOURCE), I'm pretty sure this is relevant because the crash happens right after a call to that function. Without a proper declaration, you probably get garbage from it.
<dovsienko>
jmairboeck: and...?
<dovsienko>
hmm
<jmairboeck>
now I have to see if I get the build system to use -std=gnu99 instead of -std=c99
<dovsienko>
without the declaration the function still is in the library, so it does not seem related
<jmairboeck>
no, without the declaration the compiler assumes it returns int instead of a pointer and truncates the result
<jmairboeck>
and when you put that into strcmp then, it crashes
<dovsienko>
that's a good point actually
<dovsienko>
I do not remember seeing a case of this before though
<jmairboeck>
it seems that the configure check for memrchr is faulty, because it apparently sees the function in Configure, but it doesn't actually work correctly then.
<jmairboeck>
because there is a fallback if the function is not available
<dovsienko>
Autoconf function tests come in two kinds: header test (include and test for the prototype presence) and library test (try to link a binary that uses the function)
<jmairboeck>
this isn't based on autoconf, but a custom shell script
<jmairboeck>
called "Configure" (with a big C)
<dovsienko>
these are separate checks for reasons I do not remember, and quite Autoconf scripts test only for the header or only for the library
<dovsienko>
oooooh
<dovsienko>
this may need a bug report to the upstream then
<jmairboeck>
yeah, I should probably do that
<jmairboeck>
is this an ABI break when switching from c99 to gnu99? i.e. do we need to rebuild other perl packages now?
<dovsienko>
I do not know
<Anarchos>
i have some vim files (syntax/ftdetect) to install, where shoul i put them ?
<jmairboeck>
also I'm not sure if the old version (built on beta 4) used memrchr or the fallback
<jmairboeck>
Anarchos: in a package? the vim data files are in /system/data/vim/vim91, I don't know if it looks in non-packaged also
diver has quit [Quit: Leaving.]
<Anarchos>
jmairboeck READ-only file system...
<jmairboeck>
that's why I was asking whether you were building a package
<jmairboeck>
the equivalent would be /system/non-packaged/data/vim/vim91, but I don't know if vim uses it then
<Anarchos>
jmairboeck no i just want to install vim syntax files (for metamath)
<Anarchos>
jmairboeck for ocaml there are some lines "set rtp +=..." to link them from wherever they are.
<jmairboeck>
maybe you can also put them into the home directory somewhere, I don't know
janking has joined #haiku
<Anarchos>
jmairboeck i could try but i don't want to clobber ~ with dot files ...
<Anarchos>
i find it inelegant on linux all those .XXX directory which are not yours
<Anarchos>
in your home :)
<jmairboeck>
if it is done correctly, it would be somewhere in ~/config/settings/vim or something like that
<Anarchos>
there is one.
<Anarchos>
i will try
<Anarchos>
jmairboeck it worked. I thought it would have been more difficult :)
AlienSoldier has joined #haiku
diver has joined #haiku
janking has quit [Quit: Vision[]: i've been blurred!]
diver has quit [Remote host closed the connection]
<zdykstra>
Ooh repeatable KDL after unmounting a bfs image file
<Anarchos>
zdykstra did you pass checckfs on it ?
<dovsienko>
beware that may mask the bug
<dovsienko>
consider making a backup before checking the filesystem
<zdykstra>
the image file is ironically how I do my backups
x10z has quit [Quit: My Unrecognized Mac has gone to sleep. ZZZzzz…]
<zdykstra>
I rsync /home to a mounted image, then rsync that image file to a remote host where it gets into ZFS snapshots
<dovsienko>
(in theory, malformed/damaged filesystems are not supposed to crash operating systems, but in practice they do, and such a crash is more a problem of the OS rather than the FS)
<Anarchos>
dovsienko yes i typed too fast, but realized you are correct
dby has quit [Quit: Vision[]: i've been blurred!]
Anarchos has quit [Quit: Vision[]: i've been blurred!]
<dovsienko>
about 20 years ago, when USB ports and memory sticks started to become commonplace, it also became commonplace to fix various OSes so they don't crash or execute external code in kernel context when someone walks up to a computer and plugs a carefully crafted USB storage into it
<gordonjcp>
dovsienko: oh man the fun we used to have with autorun on Windows systems
<dovsienko>
even much more so with various hotplug automount props being in fashion at the time
x10z has joined #haiku
<gordonjcp>
there used to be this great kernel module that would "do stuff" that would only allow a user to become root if a particular USB device was plugged in
<gordonjcp>
didn't matter what the device was
<gordonjcp>
you just told it "if you don't see anything off this list of VIDs, PIDs, and DSNs, then no-one gets root"
<dovsienko>
gordonjcp: I still remember SONY implementing a "security feature" in one of its CD-ROM products that used Windows autorun and could be neutralised as such (hold Shift when inserting the CD)
<gordonjcp>
yeah
<gordonjcp>
it actually overwrote Windows's CDROM drivers
<gordonjcp>
they were very lucky they were able to lawyer their way out of being prevented from trading in the UK
<dovsienko>
and then SONY tried to sue various people on the Internet for saying that the system was not fit for purpose
<gordonjcp>
given that their "copy protection scheme" very comfortably met the requirements for breakin the Computer Misuse Act
<gordonjcp>
they were really seriously lucky
dodo75 has quit [Quit: Vision[]: i've been blurred!]
<dovsienko>
so anyway, if Haiku has a bug along the above lines, it would be best to establish the root cause and to stop having it
<gordonjcp>
at this stage that sounds like "just remove support for CDROMs"
<gordonjcp>
I only own one device with optical media and that's because I still have to cope with DVDs of evidence footage
<gordonjcp>
and everyone's a bit "don't ask don't tell" about why I take my studio PC into the workshop occasionally
<gordonjcp>
"A Wizard Did It" and don't ask questions
xet7 has quit [Quit: Leaving]
<dovsienko>
I own more than one DVD drive, which is convenient for playing real music CDs and real film DVDs (some charity shops have piles of good old movies, some of which I have not ever seen)
<gordonjcp>
I think I have more DVCAM and XDCAM decks than CD or DVD decks
<gordonjcp>
oh, my laptop has a DVD drive
<gordonjcp>
I have more VHS decks than DVD drives
<dovsienko>
gordonjcp: if you have really old computer/radio hardware that you cannot force yourself to bin, there's the Museum of Communication in Burntisland, they consider taking new exhibits on a case-by-case basis
<gordonjcp>
dovsienko: I've already donated stuff to them
<gordonjcp>
also had the Rockwell AIM-65 that someone donated to them running
<dovsienko>
I've never dealt with any of those
<gordonjcp>
basically a dev board for all the Rockwell 65xx family chips
<gordonjcp>
not just the 6502 but the various IO support chips
<gordonjcp>
my dad had one
<gordonjcp>
I should try to find my Acorn Atom
x10z has quit [Quit: My Unrecognized Mac has gone to sleep. ZZZzzz…]
hightower2 has joined #haiku
vezhlys has quit [Remote host closed the connection]
x10z has joined #haiku
<dovsienko>
where I grew up computers were rare, with the only exception of ZX Spectrum clones perhaps. even to buy a computer-related book you'd have to travel a couple hours to the book shop in the nearest big city
<dovsienko>
so I tried to swallow any book that passed my way (that was years before first contact with the Internet), one of them fortunately being "the complete idiot's guide to Unix" followed by the white book by Evi Nemett a few years later
<dovsienko>
by the time I had my own PC and someone talked me into trying Linux on it, I already had a good enough idea of what to expect from the OS
<gordonjcp>
dovsienko: just a few miles down the road from me and a few miles up the road from you, ZX Spectrums were ubiquitous when I was at school
<dovsienko>
long story short, my expertise lays more in the software domain rather than hardware
<gordonjcp>
literally everyone had one
<gordonjcp>
they were made in the Timex factory in Dundee, a big chunk of which survives as some sort of business unit space
<gordonjcp>
so everyone's dad "knew someone"
<dovsienko>
you told this story a couple weeks before, I think
<gordonjcp>
you \*could\* buy them in the shops but you could also buy them in most pubs, notably the Planet Bar in Lochee
<gordonjcp>
possibly
<gordonjcp>
anyway I had a few Spectrums
<gordonjcp>
that's how I got into this mess
<gordonjcp>
I read about the various Eastern Bloc ZX Spectrum clones in the 80s, they were an obvious one to smuggle
<gordonjcp>
like you weren't walking through customs with a BBC Micro under your jacket
<gordonjcp>
I often wondered how clever the folk were that reverse-engineered the ULA, probably cleverer than the folk that designed it
<dovsienko>
the what?
<gordonjcp>
there was a big custom chip inside
<gordonjcp>
the Ferranti "ULA", Uncommitted Logic Array
<gordonjcp>
they made a big slab of gates but left unconnected, which you'd then put a couple of metal layers on to wire up
<gordonjcp>
CPLDs and FPGAs weren't really a thing then
<dovsienko>
oh. I don't think there was one in my Spectrum clone
<gordonjcp>
all the address decoding and IO was done in the ULA including the screen
<gordonjcp>
your clone probably did it in about a dozen 74LS chips
<dovsienko>
it had a genuine Z80, an UV-erasable 16KB ROM and a load of Soviet 14=pin or 16-pin ICs, and that's it
<gordonjcp>
yup
janking has joined #haiku
janking has quit [Quit: Vision[]: i've been blurred!]
HaikuUser has joined #haiku
<scanty>
boo z80. yay 6502
jmairboeck has quit [Quit: Konversation terminated!]