[Swan-dev] debugging our test runner

D. Hugh Redelmeier hugh at mimosa.com
Fri Feb 1 06:33:19 UTC 2019

I decided to take a poke at this since I had a hung run sitting around for 
24 or so hours.

How does one debug a running python program?

Googling tells me that it isn't easy.


Some brave souls use gdb, thus they have to study the entrails of the
interpreter.  I'm not that desperate.

pdb can attach to a running program IF you installed breakpoints
before you ran it.  That doesn't count.

I also found PyDev:
Sounds good.

Then I found that it was part of Eclipse.  I've never used Eclipse
because it seems like a monster, but if the monster can eat bugs, why

My computer and I thrashed about in eclipse for a while.  What a

The term from which I invoked eclipse got a message.  I didn't
preserve it.

Would anyone else have run eclipse from a terminal?  If not, how would
they see this message?

The message said that I needed to run a pkcon command to install
something for building a project.  When I issued the pkcon command
(sudoed, even though the message didn't suggest this), it wanted to
remove some packages, including anaconda.  No thanks!

I followed the blog post for running PyDev.  It took a lot of fumbling
to find where to tell it the location of the python interpreter.  It
was quite insistent that I set this but gave no assistance in locating
the setting.  Being a GUI, I am unlikely to remember how I did it.

I did try "help" but that did nothing.  I'm not complaining that it
didn't address my question, I'm complaining that it was a no-op!

Then the moment of truth.  Attach to the running program and look at a

At which point it hit a bug and asked if I'd like to report it to
three or four different folks.  I said "sure".

Then the bug reporter failed in a couple of ways: It got some dbus
thing to time out and ask me (every few seconds) if I wanted to wait
or proceed.  And the bug reporter failing provoked another bug
reporter to come up.

I gave up.  What a shambles.

More information about the Swan-dev mailing list