Live data from Hacker News

Viewing profile — romka2

romka2

HN member
Joined
Sun, May 07, 2017, 6:12 PM UTC
HN karma
201
Public activity
39 items

About romka2

No profile information was provided.

Recent public activity

  1. story
  2. story
  3. story
  4. story
  5. comment
    Comment #21042252

    You can use powerlevel10k with the default system font. It works fine with urxvt without any twiddling.

  6. comment
    Comment #20092068

    The minor modification would be from `$opts` to `$=opts`. I'd been using ZSH for a year as if it was Bash with better history and much better completions. The difference mentioned …

  7. comment
    Comment #20092035

    Pro tip: s/powerlevel9k/powerlevel10k/. Same config, same theme, but much faster.

  8. comment
    Comment #20059366

    I remember reading this article back when I was still using Bash. I've finally got the history I like only after switching to ZSH. You can get decent history in ZSH just by setting…

  9. comment
    Comment #19628851

    This format looks somewhat underpowered. If one record is corrupted, there is no way to read anything after it. For the same reason there is no lookup/sharding support, such as fin…

  10. comment
    Comment #19628786

    RecordIO supports a form of random access lookup, although it's rarely used. Riegeli supports random access lookups as a first-class operation.

  11. comment
    Comment #19628126

    > That appears to be exactly RecordIO. I suppose you mean "exactly" in a figurative way. Riegeli is definitely inspired by RecordIO and is meant as a successor to it but it's not R…

  12. comment
    Comment #19623710

    There is no listed equivalent of RecordIO. What do people use for high-reliability journals? When I needed something like RecordIO to store market data, I couldn't find anything. S…

  13. comment
    Comment #19612304

    What part of the code do you expect to get faster with `-march=sandybridge -mtune=sandybridge`? > excessive ASM is a mistake these days There is no ASM in the article.

  14. comment
    Comment #19612219

    I would agree with this sentence if "optimizing" were replaced with "making marketing claims". Optimization is the process of finding the fastest implementation. In the case of Lis…

  15. comment
    Comment #19609206

    No, I did the opposite. I made sure the disk caches are warm before each benchmark. Since all versions of ListDir are identical in terms of IO demands, warming up caches is an effe…

  16. comment
    Comment #19605492

    I've posted some details explaining why gitstatusd doesn't use inotify in https://github.com/romkatv/gitstatus/commit/050aaaa04b652e15... . The short version is that the default ma…

  17. comment
    Comment #19604005

    nftw is a wrapper over opendir/readdir, not the other way around.

  18. comment
    Comment #19603983

    Yep, this is mentioned in the doc. > [...] every element in entries has d_type at offset -1. This can be useful to the callers that need to distinguish between regular files and di…

  19. comment
    Comment #19603277

    Yep, every string is potentially an allocation (unless it's short and std::string implements Small String Optimization) plus O(log N) allocations by the vector itself. C++11 didn't…

  20. comment
    Comment #19600081

    gitstatusd calls ListDir in parallel from multiple threads. At least with a fast SSD it's CPU bound. I don't have an HDD to test on.

  21. comment
    Comment #19599921

    That's a cool algorithm. The improvements in the article look pretty small between different versions because CPU time is dominated by system calls that are out of my control. If y…

  22. comment
    Comment #19599835

    > [...] have you tried core.fsmonitor with the watchman hook? I read the docs but it seems like it's not something I can enable on users' machines. Or maybe there is a way to take …

  23. comment
    Comment #19599787

    This directory-listing code only executes on a "cold" run of gitstatusd, which happens when you `cd` into a repo for the first time. Users obviously can tolerate higher prompt late…

  24. comment
    Comment #19599680

    The real implementation of ListDir accepts the descriptor of the directory it needs to list (not parent_fd plus dirname like it's done in the article) and doesn't close it. This is…

  25. comment
    Comment #19598802

    You are quoting a self-deprecating joke. I care deeply about my code being readable, which can be quite challenging when writing high-performance code close to the metal. I'm self-…