Clearly, we need to publicize the gdb/python integration a bit more. It is easy to get, and reasonably functional.
I’ve been thinking a bit about how to get the word out better. What is most effective? Here are some ideas:
- A series of “how-to” blog entries (what do you want to do?)
- A “how-to” article in a magazine (what do you read?)
- A talk at a conference (which ones?)
I’ll probably do some or all of these — but what would be most useful to you?
I’m hoping we can ship a Python-enabled gdb in F11. Hopefully that will boost adoption. I’m also planning to ship a suite of libstdc++ pretty-printers in F11, so even if you don’t write any Python yourself, you can still benefit. (For those not following the progress, we have a feature that lets you write custom visualizers based on type; this makes printing a std::vector, or whatever, much simpler.)
And, by the way — Dodji is indeed awesome. Go give him an ohloh kudo.
8 Comments
Please track this as a feature for this for Fedora 11!
Talk to John Poelstra, and see:
https://fedoraproject.org/wiki/Features/Dashboard
I think a series of blog entries with simple examples will be most effective. Organize them in such a way that they can easily be looked up and/or turned into part of the manual. Start with the “OK, this is how you get started, fetch, build, run hello.py”, then each day introduce a fun command, convenience function, or explain what the python representation of core gdb data structures are and how you would use/manipulate them. Keep the posts short, but each one a practical complete example and you will most likely get people coming back for more.
I have to admit that I don’t actually read any paper magazins anymore. But if you manage to get something published in something like LWN that would certainly bring you lots of deep technical readers.
Fosdem is obviously my favourite, but that is on the wrong side of the ocean for you 🙂 The GCC summit seems appropriate for this kind of thing if you want something specialized. But you can do a lightning talk anywhere. Andrew Haley once did a “look me being a wizard with gdb” then minutes hands on and it was great to see someone actually drill down and come up with useful info on a program.
I’d like an article series on LWN, a bit like Ulrich Drepper did with his memory articles (although maybe a bit shorter).
I’ve not yet gotten used to screencasts, but perhaps that’s something to consider as well to demonstrate neat GDB features.
// Simon
[…] All hail Archer and Tom Tromeys python gdb integration: […]
I think both a lwn.net article and tracking this as a Fedora 11 feature are excellent suggestions.
I’d vote for an LWN article too, even if it in turn pointed to the series of blog entries. I very seldom bother to watch videos or screencasts — stick to text which is easier to produce and consume, and works much better with search engines too.
BTW … very keen to see the end result. Thanks!
[…] you are working on something new and exciting and want to make sure it brings attention to Fedora and the people working on it, please consider […]
Your expertise is desperately needed here:
http://stackoverflow.com/questions/311840/tool-to-trace-local-function-calls-in-linux
and/or here:
http://stackoverflow.com/questions/311948
It is amazing this is so hard with Linux. (or is it?)