<azonenberg>
the polyimide is too thin for a nice wide low-loss trace lol
<azonenberg>
You'll get less shielding from the DUT that way though
<azonenberg>
But it will be more flexible
<azonenberg>
And i guess in real world use, you'd want to have this probe be curved up away from the DUT anyway, like how lecroy illustrates in their wavelink examples
<azonenberg>
Having a little rubber sleeve around the probe body would also help keep it away from interference sources
<azonenberg>
actually it looks like it will be hard to get below ~69 ohms that way. So i'm in an awkward middle ground
<azonenberg>
Guess I'm back to GCPW, just using a thinner center conductor than i wanted
<Bird|otherbox>
so, that was actually remarkably useless -- cppcheck just spat out a message about a #error coming out of freetype's ftconfig.h:
<azonenberg>
Lol. time to write your own test logic or something?
<Bird|otherbox>
/usr/include/freetype2/freetype/config/ftconfig.h:99:0: error: #error "Unsupported size of `int' type!" [preprocessorErrorDirective]
<Bird|otherbox>
there
<Bird|otherbox>
that's the problem
<Bird|otherbox>
and it happens for *every single file in the project* AFAICT
<azonenberg>
interesting, there's probably a way to make cppcheck ignore some #define's
<azonenberg>
or force certain #define values
<azonenberg>
also i'm thinking the analysis should be a separate target that is NOT built by default
<azonenberg>
judging by how long cppcheck took to run on the whole project i don't want a trivial "make" to build it
<azonenberg>
it should be something we do during CI builds etc
<Bird|otherbox>
OK, trying a flag for that now before I add stuff permanently into the CMakeLists for this sort of thing
<Bird|otherbox>
should it be at the toplevel of the project only or some sort of thing that's set in the cache then acted upon by subdirectories?
<azonenberg>
i'm thinking "make analysis" should run cppcheck and possibly other tools on the entire project
<azonenberg>
from the top level
<Bird|otherbox>
right
<Bird|otherbox>
hm....the built-in CMake support would look more like a cmake -DANALYZE ..; make
<Bird|otherbox>
because I'd have to set a variable inside the CMakeLists
<azonenberg>
Bird|otherbox: that doesnt seem to make sense because cppcheck doesnt support incremental analysis
<azonenberg>
and forcing re-analysis of the entire project every compile is just stupid
<azonenberg>
even for dev builds
<azonenberg>
the ideal, i guess, would be if you could use some temp files or something
<azonenberg>
to identify which source files have changed since the last analysis run
<azonenberg>
and rerun analysis only on those
<azonenberg>
but that would be a lot of work and at this stage of development i dont think ,akes sense
<Bird|otherbox>
I think that's what the CMake support does already, because it just runs the analysis for what needs to be recompiled
<azonenberg>
oh well in that case
<azonenberg>
then i guess just figure out what cppcheck options you need to fix that error
<Bird|otherbox>
yeah, I think I got it licked
Degi_ has joined #scopehal
Degi has quit [Ping timeout: 272 seconds]
Degi_ is now known as Degi
electronic_eel has quit [Ping timeout: 260 seconds]
electronic_eel has joined #scopehal
<azonenberg>
Think I'm gonna write a SWD protocol decode soonish
<azonenberg>
i'm about to start spinning up the MEAD boards for testing tomorrow, and i'll be flashing them all over SWD
<azonenberg>
seems like a nice excuse
<monochroma>
oo
<azonenberg>
i have a basic jtag decode but i want to improve on it
<azonenberg>
What might be nice to do also is implement at least some upper level ADIv5 decodes too
<azonenberg>
(So you can figure out what registers proprietary debug tools are poking etc)
<azonenberg>
I also want to improve the JTAG protocol decode to use the full protocol analyzer framework
<azonenberg>
and then start poking at the vivado ILA more
<azonenberg>
Also on my todo for the weekend is putting a bit more polish on the tek5/6 driver and adding more trigger sources for both tek and lecroy
bvernoux has joined #scopehal
bvernoux1 has quit [Ping timeout: 240 seconds]
Nero_ has joined #scopehal
_whitelogger has joined #scopehal
sorear has quit [Ping timeout: 272 seconds]
sorear has joined #scopehal
bvernoux has quit [Quit: Leaving]
bvernoux has joined #scopehal
_whitelogger has joined #scopehal
agg has quit [Quit: WeeChat 1.8]
adamgreig has joined #scopehal
juli965 has joined #scopehal
deltab has quit [Ping timeout: 265 seconds]
deltab has joined #scopehal
adamgreig is now known as agg
<azonenberg>
Assembling MEAD prototype with production board and new enclosure
<azonenberg>
The screw holes arent threaded/drilled deep enough
<azonenberg>
SO i have to use washers to shim up the screws
<azonenberg>
i think i can make it work with shorter screws and no changes to the mechanical design
<azonenberg>
the new connector position is MUCH better
<azonenberg>
After fixing the two missing components (10 ohm LCD backlight series resistor and ferrite filtering the main power input) the pod comes up OK and displays correct text on the LCD
<azonenberg>
before i go any further i'm going to take a detour to write a SWD decoder while i have the stlink on it
<bvernoux>
ha yes nice
<azonenberg>
Going to try and design the lexan overlay for the front panel today and send out one prototype order
<azonenberg>
then make some mechanical tweaks to the enclosure
<azonenberg>
also see if i can find some slightly shorter screws that won't need washers