Bradford Larsen

@bradlarsen

software toolsmith — static analysis, compilers, automated bug finding

Massachusetts, USA
Vrijeme pridruživanja: siječanj 2009.

Tweetovi

Blokirali ste korisnika/cu @bradlarsen

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @bradlarsen

  1. Prikvačeni tweet
    28. kol 2018.

    Got a C++ project? If you can build with Clang or GCC 4.9+, building your test suite with Address Sanitizer is perhaps the easiest thing you can do to extract additional bug-detecting power from your tests.

    Prikaži ovu nit
    Poništi
  2. 2. sij

    Periodic reminder that ripgrep from is terrific software:

    Poništi
  3. 5. stu 2019.

    i make it a game now to unfollow one account on twitter each time I open it

    Poništi
  4. proslijedio/la je Tweet
    4. ruj 2019.

    My tribute to my friend Dan Garcia, who was confirmed as one of those who passed away when the Conception caught fire off Santa Cruz Island. Farewell, friend.

    Prikaži ovu nit
    Poništi
  5. 21. kol 2019.

    Periodic reminder that you can get quite far identifying software system performance problems without needing fancy tools (though they can help)

    Poništi
  6. 19. srp 2019.

    Did you ever stop to think how the two construction-related idioms — “nailed it” and “screwed it” — have nearly opposite meanings?

    Poništi
  7. 8. srp 2019.

    I’m happy to say I’m joining !

    Poništi
  8. 18. lip 2019.

    I once showed up for an interview in the center of the city at 9:30, and the office was empty except for the receptionist. “Wait, who are you?” Interview coordinator had confirmed schedule with me, but then they went on vacation without telling anyone else.

    Poništi
  9. 17. lip 2019.

    Conjecture: a web site that has logic to give a response saying “cross-site scripting attempt blocked” is almost certainly actually vulnerable to cross-site scripting

    Poništi
  10. proslijedio/la je Tweet
    31. svi 2019.

    Let’s do this: One Like, One Undefined Behavior. I plan on running this through the weekend at least.

    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet
    24. svi 2019.

    Do you believe our software tooling is complete? Is there something totally missing from it? Not "it could do X better", more "nothing does X at all."

    Poništi
  12. 9. svi 2019.

    A heuristic: when you find yourself saying or writing “methodology”, if “method” work just as well, use it.

    Poništi
  13. 25. tra 2019.

    Do you think to include yourself in your threat modeling exercises?

    Poništi
  14. 19. tra 2019.

    Is there any research on making understandable static analysis reports? Or on how to triage reports effectively?

    Poništi
  15. 18. tra 2019.

    Hey software folks! When you download binaries artifacts, source tarballs, etc, do you verify the pgp signatures / hash digests of what you download?

    Poništi
  16. 17. tra 2019.

    how are you even supposed to use gpg anyway

    Poništi
  17. 12. tra 2019.

    This doesn’t seem to be documented anywhere, and it isn't obvious from option names that this gives you machine-readable report output.

    Prikaži ovu nit
    Poništi
  18. 12. tra 2019.

    Oh wow! If you run clang-tidy with `-export-fixes=FILE.yml`, the reports are additionally emitted in YAML format. It seems like even reports without suggested fixes are included in the YAML output. So it seems it IS possible to get machine-readable output from clang-tidy.

    Prikaži ovu nit
    Poništi
  19. 11. tra 2019.

    It sounds like Amazon Batch might be the simplest way to get started, without having to code up custom work queuing etc. I will report back at some point after experimenting.

    Prikaži ovu nit
    Poništi
  20. 10. tra 2019.

    Each running Docker container uses less than 4GB peak memory (probably less than 1GB, but I’d have to check). The containers make no network requests, and rely on nothing that’s not already built into the image. I’m dealing with an embarrassingly parallel batch processing task.

    Prikaži ovu nit
    Poništi
  21. 10. tra 2019.

    I have a Docker image that takes an input string, runs sequentially for 15m, then dumps 100MB of output to a well-known location. I have 75k input strings. On my piddly hardware, it’d take 100 days of real time. What’s the easiest way to run this batch workload in the cloud?

    Prikaži ovu nit
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·