<awygle>
anyway if it's not useful oh well, i just happened to come by it and thought of glasgow
<whitequark>
I guess it can be useful to debug fx3-side stuff
<awygle>
is there any documentation on writing applets or do i just start from a sample?
<whitequark>
latter
<awygle>
roger
<awygle>
viz. software, should i be using master or is there a stable/released version?
<whitequark>
master
<awygle>
copy
<awygle>
ugh ubuntu 18.04 only has python 3.6.9. guess it's time to roll to 20.04
<awygle>
oh wait, i am an idiot, disregard.
<awygle>
whenever i make an error in this applet i get, in addition to the real error (twice), an exception in sys.excepthook that i don't have the module named "apt_pkg". any idea which python module i'm missing? i have python-apt
<whitequark>
uhh
<whitequark>
i have no idea? first time i see that
<tnt>
whitequark: I meant python3-xxx vs python-xxx
<awygle>
i'm pretty sure something in the Python Guts wants python3.6 still
<awygle>
even though i update-alternatives-d myself into python3.7
<awygle>
when i changed update-alternatives back it didn't run glasgow anymore so i've probably caused myself some Problems. i may try reinstalling glasgow and if that doesn't work i'll just live with the message lol
<whitequark>
goddamn it
<tnt>
I'd guess the requirement for apt is some magic from the distribution trying to guess which package you're missing in case of error.
<whitequark>
can you use `python3.7 -m pip install -e .` in the glasgow dir
<whitequark>
that should work just fine regardless of the selected alternative
<awygle>
that ran just fine, now i get the apt_pkg complaint regardless of how update-alternatives is set
<awygle>
i don't really have time to screw around with this now so i'll just live with the additional error message, it's not that intrusive
<whitequark>
the only thing worse than python packaging is distro python packaging
<whitequark>
it's truly despair-inducing
<awygle>
zetsuboushita
<whitequark>
yes
<whitequark>
exactly
<d1b2>
<YasirKhilji64> @whitequark what is the purpose of -e . in pip install???
<whitequark>
"editable"
<whitequark>
well, . is installing the current package (the glasgow/software one)
<whitequark>
and -e makes it so that it uses a link to the sources rather than copying them into site-packages
<whitequark>
so when you edit the checked out source code, the changes appear immediately
<whitequark>
it's similar to `python setup.py develop` which you might be familiar with, except they deprecated that one
<d1b2>
<YasirKhilji64> Ah ok
<whitequark>
(and for the onlookers, it's not just deprecated, but also broken in some subtle ways. i recommend always using pip -e . instead, having had the misfortune to discover in which exact subtle ways the brokenness lies)
FFY00 has quit [Ping timeout: 260 seconds]
FFY00 has joined #glasgow
<Stormwind_mobile>
Also, if you didn't know: the package naming convention for python packages seems to be nonuniform, which is capitalized on by hackers who distributed a copy of a package with a very similar name, but with some extra malware.
<Stormwind_mobile>
I learned that there's a name for that: "typo squatting"
key2_ has joined #glasgow
lukego_ has joined #glasgow
awygle_ has joined #glasgow
kmehall_ has joined #glasgow
<d1b2>
<edbordin> Ah yes, my favourite package nunpy
john_k[m] has quit [*.net *.split]
kerel has quit [*.net *.split]
lukego has quit [*.net *.split]
key2 has quit [*.net *.split]
kmehall has quit [*.net *.split]
awygle has quit [*.net *.split]
key2_ is now known as key2
lukego_ is now known as lukego
<apo>
mumpie
kerel has joined #glasgow
john_k[m] has joined #glasgow
Getorix has quit [Read error: Connection reset by peer]
Getorix_ has joined #glasgow
ExeciN has joined #glasgow
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
givewaytomydeskt has quit [Quit: so long king Bowser]
bvernoux has joined #glasgow
awygle_ is now known as awygle
<awygle>
can i run one applet on port A and another on port B?
<awygle>
it kind of seems like that should work based on how the claim stuff works but also seems like no because the bitstream is rebuilt when you run "run"
<d1b2>
<tnt> Nope. Although it's a stated design goal and some stuff is in place (as you noted). Some stuff is still missing.