<OscarL>
>>>> 02:28 <win8linux[m]> Why does the VMware Add-ons package have a VirtualBox driver?
<OscarL>
win8linux[m]: VBox can make use of the vmware graphics driver (assuming the VBox is setup to use the VMSVGA graphics adapter.
<win8linux[m]>
Actually, that step of using the VMSVGA graphics adapter doesn't seem necessary anymore.
<win8linux[m]>
Well, at least when it comes to automatically resizing the resolution.
<OscarL>
but truth to be told, I've only use the "VBox guess additions" package partially (after unpacking it, and only use what I KNOW works)
<OscarL>
at a point I even attempted to build the VBbox recipe locally, with the intention to see if I could either fix, or at least disable the broken bits.... but that proven to be too much for both my hardware (PC), and my brain.
<OscarL>
using (under VBox on Win10)... only the VMWare drivers, and the VBoxTray app (needs vboxdev driver installed)... seems to be the best I can do.
<OscarL>
ocassionally, the clipboard integration (part of that vboxdev/VBoxTray combo)... either crashes, or freezes apps on the host OS.
<OscarL>
I still use it most of the time while on Win10, as VBox generates less issues on my old hardware than with VMware 16 (which seems to eat far more RAM).
<OscarL>
I will be happy to help, wherever I can, with any efforts on making better VBox guest additions for Haiku.
<OscarL>
FWIW... the "step" of using vmware driver under VBox, in *my particular* case, comes down to using the same installs for both bare-metal, and VMs. Using the VMware driver allows me to use a common resolution, that the VBox driver/"hardware" refuses to accept".
<OscarL>
changing the subject a bit... seems I can fully utilize usb_rndis now with my old Android 4.0 phone... as long as I follow some strict steps :-D
<OscarL>
was even able to update to latest nigthly! Marvelous news for me... and a huge "THANK YOU FOR YOUR WORK" to PulkoMandy... regarding the usb_rndis driver.
<OscarL>
Don't even look at it wrongly, tough... because it likes to KDL on this customized driver! :-D
erysdren has quit [Quit: Konversation terminated!]
<OscarL>
I wish I was smart enough to help make it better.
<OscarL>
enough verbiage from me for one day. Be well you all!
* OscarL
HATES "*** failed to commit transaction: Failed to change the package activation in packagefs: Name in use"
* QuARC[m]
has reinstalled haiku on its laptop in BIOS mode
<OscarL>
That eerror stops further systems updates.
<QuARC[m]>
why do haiku icons look so good
<OscarL>
couldn't pkgman/package_fs just "overwrite" the .hpkg from the proper server?
tuaris has quit [Read error: Connection reset by peer]
<OscarL>
or something like that.... not shure how to properly handle that edge-case.
<OscarL>
(in this case... trying to install/update python3.10.13-2... on a system that has it already installed... because I was the bloddly one that did the PR for it!.
<OscarL>
s/bloddly/bloody/ (I guess... English is hard)
<augiedoggie>
are you doing afull sync
<augiedoggie>
?
* OscarL
humbly tries "pkgman update" instead, following augiedoggie's lead...
<OscarL>
Wonderful!. That worked! Thanks a bunch augiedoggie!
<augiedoggie>
nice
<OscarL>
but in the end..... WHY?
<OscarL>
:-D
<augiedoggie>
full sync sees that the packages are from different repos
<augiedoggie>
it tries to sync with the registered repos
<OscarL>
I see. But it sounds like one of those: "it makes technical sense, but final users should not care, or even know about that 'diference'", no?
<OscarL>
(I've hit this way too many times... you could say :-P)
<augiedoggie>
that's why full-sync should be avoided
<augiedoggie>
or treated more as a rescue option instead of the regular way to update
<OscarL>
btw... always glad to be schooled by you, augiedoggie! (I mean it... you're awesome!)
x10z has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<augiedoggie>
it's just my opinion, don't take it as facts :P
<OscarL>
I'd be willing to bet that your experience beats my hunches at least 8 out of 10 times :-P
<OscarL>
(those odds... still beign heavily biased to *my side* :-D)
OscarL has quit [Quit: "what a quitter!"]
Begasus has joined #haiku
<Begasus>
g'morning peeps
v_harkonnen has joined #haiku
KitsunePrefecture has joined #haiku
<KitsunePrefecture>
Ahoy folks !
<KitsunePrefecture>
I have an issue during installation
<KitsunePrefecture>
I wanted to create an exact copy of my R1B4 install on a Kingston thumbdrive
<KitsunePrefecture>
to the same Kingston product thumbdrive
<KitsunePrefecture>
the goal to change this secondary one to a nightly - changing the Haiku repo in it after the installation finished
<KitsunePrefecture>
unfortunately I got an I/O error during install, so install interrupted
<KitsunePrefecture>
as I had USB disk errors previously but mainly with USB 3.§ thumbdrive
<KitsunePrefecture>
I constantly monitored syslog during install.
<KitsunePrefecture>
The following lines appeared :
<KitsunePrefecture>
KERN: usb error ehci 2: qtd (0xd85380) error: 0x9c002d40
<KitsunePrefecture>
KERN: usb_disk: operation 0x28 failed at the SCSI level
<KitsunePrefecture>
Until now I had not experienced such errors on USB 2.0 thumbdrives using Haiku
<KitsunePrefecture>
but often on a Kingston 3.0 Datatravaler - I gave up on that using as an install target - it would have been my main target for Haiku 64 install
<KitsunePrefecture>
but all kind of attempt to install Haiku on it failed during copy from install media : nor USB drive neither DVD source had not helped
<KitsunePrefecture>
now I would like to create a bleding edge testin environment
<KitsunePrefecture>
on a same USB key like my daily driver
<Begasus>
KitsunePrefecture, please use some past site instead of spamming the channel :)
<KitsunePrefecture>
as I could not install Haiku 32 bit on it as known error - DriveSetup had not set bootable during install - then copy finished successfully to this drive
<KitsunePrefecture>
there were no read or write error at all
<KitsunePrefecture>
ok, I finished
<Begasus>
KitsunePrefecture, is it 32bit compatible?
<Begasus>
I can't install 32bit here either
<KitsunePrefecture>
sorry, Where to ?
<KitsunePrefecture>
what is 32bit compatible ?
<Begasus>
hardware
<Begasus>
on this laptop*
<KitsunePrefecture>
64 and 32 bit compatible
<KitsunePrefecture>
I was on 32bit Haiku earlier with this Dell WST grade laptop
<KitsunePrefecture>
Dell Precision M6700
<KitsunePrefecture>
my issues related only to usb thumbdrives generally
<KitsunePrefecture>
install both failed when I tried to install lately
<KitsunePrefecture>
I have no issue when I used dd to copy the install image to thumbdrive
<KitsunePrefecture>
but there is no 14GB size image
<KitsunePrefecture>
so once, only once I could install Haiku 64 bit onto this usb 2.0 thumbdrive, wherenow the install size is 14 GB
<KitsunePrefecture>
from this installer purpose drive I could not install Haiku to a 3.0 drive
<KitsunePrefecture>
now, as it was succeed earlier to install Haiku 64 bit to a similar drive - I attempted to install Haiku 64 bit here
<KitsunePrefecture>
also as I could not use as installer
<KitsunePrefecture>
it became my daily driver
<KitsunePrefecture>
it is frustrating to me
<KitsunePrefecture>
the installation IS NOT repreducible lately for me
<KitsunePrefecture>
fails at copying
<KitsunePrefecture>
for me .. it is not rocket science - but blocking me extremely
<KitsunePrefecture>
finally I try to create a 14 GB image to use dd to copy it
<KitsunePrefecture>
I tried dd-ing partitions but that failed - I think I had not used a well parametered command
<KitsunePrefecture>
or how I attempted is not possible
<KitsunePrefecture>
I still had not educated well related to dd, it seems :))
vdamewood has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<KitsunePrefecture>
anyway it's sunday, I may give up on it
tuaris has quit [Read error: Connection reset by peer]
FreeFull has joined #haiku
ClaudioM has joined #haiku
HaikuUser has joined #haiku
HaikuUser has quit []
<nephele>
Hello
BrunoSpr has joined #haiku
<BrunoSpr>
hi all
<BrunoSpr>
What is the "Windows"-Key good for, same for AltGr?
itaipu has quit [Ping timeout: 480 seconds]
v_harkonnen has quit [Ping timeout: 480 seconds]
<nephele>
BrunoSpr, You can hold it to use tab stacking
<BrunoSpr>
AltGr is used in Artpaint for deselecting, just found out
mmu_man has quit [Remote host closed the connection]
<BrunoSpr>
nephele hello, ah yes I forgot this one
<BrunoSpr>
I removed the left Windows-key for gaming!
<BrunoSpr>
Looks like i thought: PowerStatus makes new and alot problems, and keeps @Waddlesplash busy, with this enhancement!
<phschafft>
erysdren: wrote a new shell ;)
<erysdren>
oh hell yeah
<phschafft>
it now supports backspace. ;)
<phschafft>
that was like the number 1 request by all testers ;)
<erysdren>
lol
<nephele>
what about delete
<phschafft>
if (c == BS || c == DEL) {
<phschafft>
erysdren: it also supports now symbolic names, and most basic io stream redirection.
BrunoSpr has quit [Quit: Vision[]: Ich wurde gewaschen!]
<phschafft>
nephele: does that answer your question?
<nephele>
phschafft, if it's both in the same if statement... maybe, sounds like you implemented delete to behave as backspace, and not as delete ;)
<phschafft>
nephele: yes, but there are two reasons for that.
<phschafft>
the first being me not supporting line edit beside backspace. so there never is a character under the cursor.
<phschafft>
the second is that some terminals have BS and DEL swapped.
<phschafft>
so this way they both work.
<nephele>
not sure what you are targeting wih a new shell :)
<nephele>
For me, if i were to write a new shell i would kick the vt100 layer to the curb
<phschafft>
nephele: ah, I guess you don't have the context. This is NOT for a desktop or something. this is specifically for a micro controller based system.
<nephele>
ah. okay. Even in that context, especially when you control the shell completely. I would not use a terminal emulator. But it depends on how much time you have too :)
<phschafft>
how exactly do you mean 'would not use a terminal emulator'?
<nephele>
posix based shells are expected to run in a terminal emulator that is somehow compatible to the vt100 and derivitives (and later specs)
<phschafft>
you mean a PTY.
<phschafft>
and how do you want to skip that?
<nephele>
not only, also the actuall "emulator", that is the user visible part which will for example change how data is rendered based on control sequences
<nephele>
what do you mean with skip it?
<phschafft>
I wonder how else a shell would interact with the user.
<phschafft>
what your idea is.
<phschafft>
you said what not to do, now I wonder what your suggestion is.
<nephele>
I'll switch trains in some minutes, i'll elaborate then if that's fine :)
<phschafft>
sure. IRC is async.
<phschafft>
also if anyone wants to have a demo, let me know. :)
<nephele>
mostly expected to be "realtime" ;)
<nephele>
(and I'll loose the connection once the laptop suspends, so see you in a bit)
nephele has quit [Quit: This computer has gone to sleep]
<waddlesplash>
AkshayWarrier[m]: that's probably not the real error, the real problem is likely further up
nosycat has joined #haiku
<waddlesplash>
PulkoMandy: win8linux[m] I think this is a global rate limit on the server that kallisti5 added to deal with bots
<AkshayWarrier[m]>
waddlesplash: Oh, I'll try a rebuild and send more of the output
<PulkoMandy>
Itys always the users not in the us or eu hitting that limit, so I doubt it's just a global limit...
mmu_man has joined #haiku
<waddlesplash>
PulkoMandy: odd. well kallisti5 should know and be able to investigate
mmu_man has quit [Remote host closed the connection]
mmu_man has joined #haiku
nephele has joined #haiku
<nephele>
phschafft, here i am again.
mmu_man has quit [Remote host closed the connection]
<nephele>
there was a discussion on the forum related to this too
<nephele>
but basically it boils down to removing the posix compatible commands and stdout/stderr concepts, instead using structured message passing between components of a "commandline", making the Window that displays the output capable of showing/interpreting that data and such
<nephele>
pulkomandy in that thread suggested commands returning a BView for example
<nephele>
for me one big motivation for structured data is for example that say a compiler output saying which files are generated can give just the proper paths of the files, and all the internal state of in which directory the compiler was can be ignored
<nephele>
(so no more "generated ../../../../Some/directory" (where .Some/ was already where the compiler was for instance too)
<phschafft>
yes.
<phschafft>
nephele: and this is exactly what this shell does.
<nephele>
But you still want to run in a vt100 emulator?
<phschafft>
it adapts the physical TTY (a real TTY!) to a message passing.
<nephele>
Ah, I suppose SerialConnect is your client then if you have a real tty?
<phschafft>
I mean all this does is to parse a command line and call a function.
<phschafft>
and at some point you need to convert keyboard input into that structure to be proccessed by whatever part of your system.
v_harkonnen has joined #haiku
HaikuUser has joined #haiku
<nephele>
Well, If you learn anything interesting about the design i would apreciate if i could read some of it to take it into consideration when I design a shell for Haiku :)
HaikuUser has quit [Remote host closed the connection]
HaikuUser has joined #haiku
<phschafft>
nephele: feel invites to our presentations at work where I at random intervals share the results from our research. :)
HaikuUser has quit []
HaikuUser has joined #haiku
HaikuUser has quit []
<nephele>
Just carry a clipboard and walk inside? ;)
<phschafft>
basically. ;)
<phschafft>
na, we have limited seats, but I have annouced things here in the channel in the past. (just don't want to go too offtopic and ad-spammy)
<nephele>
I don't actually know what/who you work for
<phschafft>
nephele: Löwenfelsen, a little german software devlopment company
<phschafft>
we work closly with the Xiph.Org Foundation.
<phschafft>
**development
<nephele>
Hmm, brandenburg
<phschafft>
(all our presentations are virtual, as our company is basically home-office only)
<nephele>
ah okay. Well, depending on when that is would definetely be interesting
<phschafft>
generally fridays somewhere in the UTC-afternoon.
<phschafft>
but it depends on who has intrest, so sometimes we shuffle things a bit around to work better.
<nosycat>
Xiph.org is a good cause.
<scanty>
wow, playing flacs on haiku uses almost no cpu
<nosycat>
Haiku is famously efficient at playing media files.
<nosycat>
An artifact of its architecture, I think/
<nephele>
... of just using ffmpeg behind the scenes?
<erysdren>
:P
<scanty>
well, whatever it is, it's quite nice.
<scanty>
cpu usage is basically immesurable
<scanty>
(sp?)
<nephele>
phschafft, now that comment about the .ogg file makes more sense :)
<scanty>
immeasurable.
<phschafft>
nephele: ;)
<nephele>
Today at fosdem was quite nice, spoke to a lot of people about various things, also to some about Haiku :D
<Begasus>
closing down in a bit, just need to show scanty something first :P
<scanty>
nice Begasus..... now, where is my Kontact? ^_^
<Begasus>
not working (no idea how to setup this thing), but looks nice in a dark color :D https://0x0.st/Hk2J.png
mmu_man has joined #haiku
<scanty>
looks, good... doesn't work?
<Begasus>
got that thing nailed nephele :)
<Begasus>
doesn't retreive mails, folders aren't set (and it doesn't want me changing paths) ...
<scanty>
i see.
<scanty>
anyone working on it?
<Begasus>
nice for a show-case, but further then that ... :)
<Begasus>
not that I know of
<scanty>
mmm.
<Begasus>
maybe 3dEyes is already way ahead of me here :)
<scanty>
so, he's working on it?
dpirate_ has joined #haiku
<phschafft>
nephele: one of my problems with what is in that forum post is that it seems to focus only on user interaction, completly ignoring e.g. scripting.
<Begasus>
not that I know of too :) I know he replied (somewhere) that he still haden't figured out qt6webengine
<scanty>
i see.
<scanty>
guess i just have to keep waiting.
<Begasus>
k, time to grab a bite here
<Begasus>
cu peeps!
Begasus has quit [Quit: Vision[]: i've been blurred!]
<nephele>
phschafft, Which post? for scripting itself I don't think there is that much difference, just that it is structured data (and for example stuff like quoting beeing hard can be fixed in a script)
HaikuUser has joined #haiku
HaikuUser has quit []
<phschafft>
nephele: if you return data in a form that is for user view how can I use that again as input?
dpirate has quit [Ping timeout: 480 seconds]
<nephele>
You don't. The data is returned in a form that is structured, the application should just format it so that it makes sense for users
<nephele>
or maybe I am misunderstanding your question?
<phschafft>
here having a structure data as a result of find would be very nice.
<phschafft>
so we could have an actual list, not a stream of bytes.
<nephele>
yes, I agree on that point
<phschafft>
my point is however that whatever you return it must be in a way that xargs can read it (with xargs like being replaced by some shell construct here anyway)
<phschafft>
so whatever find returns must be both machine readable and user renderable.
<nephele>
I can think of two aproaches here: one would be to have negotiation for the format like we have for copy/paste, but this might become complex the bigger your "chain" becomes
<nephele>
and the second would be having only one structured format that is defined which find will return that is machine readable
<phschafft>
yes.
<nephele>
and if this is send to the equivalent to stdout it would be the responsibility of the user visible part to actually render this
* phschafft
nods.
<phschafft>
we do this using roles.
mmu_man has quit [Remote host closed the connection]
<phschafft>
so basically on any output port (in our system you can have any number) you return a value. that value is then checked against it's roles and rendered accordingly.
<phschafft>
for example if it is a file of some kind you might want a preview.
<phschafft>
if it's a contact information you might want to have a business card style rendering.
<nephele>
indeed. I think for this case the replicant concept could come in Handy, since we do already have applications in Haiku capable of displaying such types of data for instance
* phschafft
nods.
<nephele>
although, it does not have to be a file that is returned, it could be a contact data type directly without ever touching the disk
<phschafft>
yes.
<phschafft>
same here.
B2IA has quit [Quit: Vision[]: i've been blurred!]
<nephele>
would your roles be equivalent to say mimetypes?
<nephele>
it sounds to me like that would be an analogous concept that could work for Haiku
<phschafft>
media-subtypes (what people commonly call mime types) are roles.
<phschafft>
over here the call to ask if something has a role is DOES() (see perldoc UNIVERSAL ;)
mmu_man has joined #haiku
B2IA has joined #haiku
<phschafft>
so basically I would ask $obj->DOES('image/png')
<nephele>
i suppose another term makes sense if you don't use email, but then it's called this in haiku everywhere :g
<phschafft>
but I could also e.g. check for $obj->DOES('proto-file') (anything that seems to be a file)
<nephele>
i've not heard media-subtype before, thoguh
<phschafft>
or $obj->DOES(any => ['person', 'account']) (anything that is a person or an user account)
<nephele>
nice
<phschafft>
any object can have any number of roles, and roles can imply each other.
<phschafft>
for example a inode is also a proto-file
<phschafft>
+n
<scanty>
wow, my emulator compiles in 38seconds, including lto
<phschafft>
:))
<nephele>
so, similar to drag/drop, copy/paste negotiation? :)
<nephele>
I'm really liking how much of the concept haiku uses for gui applications come in handy in such a system too
<phschafft>
nephele: the point here is that it is more push based, so no negotiation. but there are concepts for that. like all the proto-* types and proxy objects and stuff like that.
<nephele>
So, i would push all presentations that make conceptually and the client application has to deal with that?
<nephele>
sense conceptually*
<nephele>
say for the find application: the list of files, and a second part that is a set of errors perhaps (if any entry was forbidden to view), and a third part that sais how many were returned etc.
itaipu has joined #haiku
<phschafft>
find's primary output is surely the list. I think we agree on that.
mmu_man has quit [Remote host closed the connection]
<nephele>
yes
<phschafft>
as find lists actual files, it would be a list of the actual inodes.
<phschafft>
(not filenames or such)
<phschafft>
it may have a second list with errors, or maybe another output that only contains the UNIX like status code.
<phschafft>
so, within one system there should only be one super type of errors and one super type of status codes.
<phschafft>
and a renderer should implement things on that level.
<phschafft>
at least.
<nephele>
I think the concept of the unix error codes are a bit imprecise, it is not really easy to handle. It might make sense to have a type of error-list that includes which command encountered the error that is just passed along, with appending more errors to the chain so it could be viewed at the end
<phschafft>
nephele: I don't argue for or against status codes. just that if you use any existing software as part of it that software will return such a status code.
<nephele>
at least for a user figuring out what a status code means, and from which app it comes, is very hard in the current posix like shell enviroment
<phschafft>
and the OS will then translate that 'legacy' code into an object passed via the API.
<nephele>
That would be one way to do it, I would have done this the other way around, that no "legacy" apps are in this system, and explicit support has to be added. So the posix system can live side-by-side with this one if needed
<phschafft>
that is a design question for the OS.
<nephele>
Yes
<phschafft>
I just mean that you might have something at least like a status code.
<phschafft>
and as this has mash structure you can have things like 'this status code is a type of a successful status code'
<nephele>
one motivation for me for this is that the posix commands are not really readable, for most you just have to know what they mean, and their names cannot be localized. So i would replace the set of functions they provide with locsalizeable alternative
<phschafft>
or 'this is a type of error'
<phschafft>
then even if you don't know what that status code means you can do the basic things:
<phschafft>
you can report it as error or success. you can also read it's name and description from the system database.
<nephele>
in posix api there is only one sucessfull status code and the rest is an error :g
<phschafft>
provide tooltips to documentation, display linked icons along side with it...
<nephele>
We could probably use the haiku errors if we want to?
<phschafft>
all that without support for that specific code and/or it's actual type.
<phschafft>
because all those are so meta level that you can do that on any type of object.
<phschafft>
nephele: sure.
<phschafft>
just subclass something like 'status code' ;)
<nephele>
Hmm, wrapping "legacy" commands with a wrapper would definetely be doable
<nephele>
but that might be an area that would be better left to specific wrappers, as some commands might have to be wrapped differently
<nephele>
I suppose this also depends on wether the system only sets up these connections or if it is actually the middleman that handles them
<phschafft>
for example on this system I'm working on you can just pass stdin/out/err via a port.
<phschafft>
so a legacy subsystem would just export that this program needs those.
nosycat has quit [Quit: Leaving]
<phschafft>
and if a the program has some 'basic' support it might declare more infos on the use of that or if it might maybe not need them.
<phschafft>
and a fully integrated program might skip those concepts.
mmu_man has joined #haiku
<phschafft>
for example 'date' could declare that it needs stdout and stderr, but not stdin.
<nephele>
You might need a translator/wrapper then, I would not expect applications in the new system to neccesarily support the stdout/stderr equicalent roles
<phschafft>
and 'vim' could declare that it requires a controling terminal or a X11 connection or ...
<phschafft>
nephele: still no problem. erysdren's VM already works in such a way that it is run on behalf of something else (similar to shebangs).
<nephele>
yes, not a big problem. But still a bit extra work if you want to take legacy apps with you ;)
<phschafft>
I'm with you, just that I think it's hard to NOT support them. ;)
<nephele>
Haha, yeah. It would be hard to design it in a way that supporting them becomes impossible
<nephele>
but I don't want to design it on the other hand that supporting those is a first priority
<phschafft>
I mean more like without them you need to reach a critical mass first. and that is really hard.
<phschafft>
specifically with users that want /their/ software.
<nephele>
That is true
<phschafft>
'But I need word 95!!!!'
<nephele>
I think it would make sense to have an api that existing clients can implement to work in this system on their own
<nephele>
So if you need the posix-esque stuff you start it "in bash". and if you do not you can use the new system
<phschafft>
also keep in mind that a program might have any number of APIs implemented in it's binary.
<nephele>
the posix equivalent commands atleast should have to be implemented completely in the new system
<phschafft>
there is no problem with that.
<nephele>
yes
<phschafft>
generally I also feel like all the POSIX stuff should go into a library. and I was pleased to see that Haiku is ahead in that regard.
<phschafft>
(but I think that it should also be a seperate library on *any* POSIX compatible system)
<nephele>
For the (far) future, in my mind, it should also be possible to just deinstall the posix layer for haiku if you don't need it ;)
<phschafft>
generally happy with that. but not sure if that is something that will work in reality.
<nephele>
with seperate library you meen libroot?
<phschafft>
plus there might be linterlinked stuff like the things both C and POSIX define.
<phschafft>
like stdio,
<phschafft>
so if you want to be C conform you need to have some of that.
mmu_man has quit [Remote host closed the connection]
<nephele>
yes, indeed
<phschafft>
and it would be hard to really define if it should be in libC or libPOSIX ;)
<phschafft>
which means that those might depend on each other.
<nephele>
(or just have it in both with the same implementation :P)
<phschafft>
BUT that doesn't mean that a program that doesn't need that could just skip linking libc.
<erysdren>
anyone here tried that QBE compiler backend for anything?
<erysdren>
i just ported it to Haiku, but it's been around for a while
HaikuUser has joined #haiku
HaikuUser has quit []
<nephele>
I have not heard of it before, interesting project
<erysdren>
yeah i've been toying with it, it's neat. i've had mental designs for a programming language for a while, and i think QBE could allow me to prototype a compiler
<phschafft>
my demo system is now free again if anyone wants to try it out. :)
<phschafft>
(prerequirement: socat installed)
mmu_man has joined #haiku
<nephele>
I still need to design one too... for my game :D
B2IA has quit [Quit: Vision[]: i've been blurred!]
B2IA has joined #haiku
mmu_man has quit [Remote host closed the connection]