<wiebel[m]>
Have you proposed the idea of such an applet or actual code you might want to link to?
<whitequark[cis]>
it was an idea that's been discussed
<whitequark[cis]>
usbip is notoriously complex and unreliable (as in "can hang or crash your kernel") so I'm not willing to support that either
<whitequark[cis]>
but of course someone could do either of those in a fork
<wiebel[m]>
I used usbip in the past with great success, but I can totally see how this brings also complexity in the completely wrong direction.
<whitequark[cis]>
the usbip 'happy path' usually works fine. but if you ever have an error in whatever's on the other side (or in the network, if you're using it over actual IP) it works less fine
miek has joined #glasgow
andymandias has quit [Remote host closed the connection]
<joshua_>
yeah, usbip is the level of stability where it would not be merged into the kernel if it were proposed today
<joshua_>
I don't recommend it as a good option, only as an option when all of your other options have run out. this is why this proposed applet was to be named 'wafflehouse', to indicate ('obviously') that it was truly a bad idea.
<wiebel[m]>
I'm sorry but I don't get the cultural konotaleo konotationtion of the word wafflehouse
<wiebel[m]>
Nice, someone's got to fix the not very cohesive entries in the urban dictionary. Or is ud not a thing anymore?
<wiebel[m]>
Still looks cozy to me. eg
<wiebel[m]>
s/ /\*/, s//\*/
<joshua_>
I mean the thing is that waffle house is actually good. I have been on a business trip to Atlanta where the absolute best part of my trip was leaving my jackass coworkers and going to get a waffle for dinner at 9pm. but it is *also* where you go to pick a fight and throw a chair at someone because it is 4am and you are heavily intoxicated.
<joshua_>
which is approximately how I feel about the original concept that I had, which was that a userspace applet would serve two MPSSE-implementing USB interfaces (and, presumably, also the other FTDI modes), with the MPSSE engine running on Glasgow
<whitequark[cis]>
i think the wafflehouse concept is really cool and i've worked on things that are worse, personally
<wiebel[m]>
The notion of, "being open no matter what" draws a pretty sharp picture of what's going on.
<whitequark[cis]>
and if not for the usbip part of it, i might even consider having it as a part of the main stack; usb device emulation is not that far out
<whitequark[cis]>
unfortunately, usbip is kind of really bad
<joshua_>
yeah, it would actually be useful in a lot of ways, to be clear
<joshua_>
(also being able to do the fabled quad-MPSSE USB device would be really nice)
<wiebel[m]>
Is the facedancer approach somehow manageable or is it even worse?
galibert[m] has joined #glasgow
<galibert[m]>
Is there any alternative for random device simulation ?
<whitequark[cis]>
what does facedancer even do
<joshua_>
but yes, it is something I wish to do, like, if I took a few month sabbatical and packed up my lab and went to an airbnb in the hills to work on 'lol' projects. it would be a fun weekend hack
<wiebel[m]>
I guess I grossly underestimate the effort for those implementations.
<wiebel[m]>
From a first glance facedancer provides USB device emulation in python.
<wiebel[m]>
Ment for mitm debugging.
<wiebel[m]>
I'm still wondering in which way the renumberation from the fx2 is such a major complication, can't it "simply" be resetted by a vendor command or such and then boots directly via USB. Or was that the way you initially envisioned?
<whitequark[cis]>
I do not wish to be the point of support for this functionality. I do not wish to develop it, to answer questions about it, to keep backwards compat for it, to document it, to think about API design for it, to distribute and deploy it
<whitequark[cis]>
every single step of that is going to be in some way miserable, and combined, they are exponentially worse
<whitequark[cis]>
that's not to say that I can't technically do it. I can, of course. I just don't want to, because it will cause me immense suffering
<wiebel[m]>
This project has such a long history for practically being around for a month. That is kind of confusing and to be frank quite intimidating, beside the sheer complexity of the code, so please excuse me asking questions that have long been answered. Is there some condensed historical reading to be found, besides the closed issues?
<whitequark[cis]>
'practically being around for a month'?
<whitequark[cis]>
I'm confused
<wiebel[m]>
Well the majority of the glasgows has been delivered in the last months haven't they?
<whitequark[cis]>
people have been building their own for close to 5 years now
<whitequark[cis]>
I never planned to sell hardware, on the basis of not having the resources or, frankly, desire to do so
<wiebel[m]>
At leas I have received mine not that long ago. And following the project purely theoretically is, well not exactly exciting.
<wiebel[m]>
s/leas/least/
<whitequark[cis]>
the project has been usable since ~2018
<wiebel[m]>
Don't get me wrong I'm not complaining it's bliss to work with such a polished framework.
<whitequark[cis]>
as for historical reading: not really, it's mostly in my head
<wiebel[m]>
Of course it has been around since 2018. But I don't expect the number of users having built it doesn't even come close to the number of people having received in the last few month.
<wiebel[m]>
s/doesn't/does/
<whitequark[cis]>
I'd expect on the order of ~100 people to have built their own or built a few for others
<whitequark[cis]>
which is significantly lower than the 1b2 production run, but not completely incomparable
Attie[m] has joined #glasgow
<Attie[m]>
don't forget I shipped ~50 some time back (I forgot how many exactly)
<whitequark[cis]>
re: questions being long answered: it's difficult to conclusively answer why something is difficult to someone with no background, especially if it concerns not a one-time development, but a dynamic process of maintenance done by real, specific people
<whitequark[cis]>
this is why I prefer saying "I do not wish to do it" over "it is hard to do so"
<wiebel[m]>
s/doesn't/does/, s/month/months/
<Attie[m]>
I think it's important to remember that this project is both the hardware and software parts... if you want to repurpose the hardware, enumerate differently, etc... then that's a new concept, and it therefore stands separately to "Glasgow"
<wiebel[m]>
Please don't get me wrong. I'm not complaining about the timeline of your production or delivery. I'm happy as can be to have it.
andymandias_ has joined #glasgow
<wiebel[m]>
That is a very good point and to be honest you put out a piece of software that makes it very hard to say "Screw it I write my own" one would be stupid to do so. I really hope to master some of it to maybe even contribute at some point. But it seems a long way away. Granted I probably lack the skill and time.