pwaller
Repos
223
Followers
210
Following
48

The Go programming language

106161
14944

Fix golang compiled binaries on x86_64 so that they can be packed with UPX

324
19

Show where time is wasted during the context upload of `docker build`

307
15

Generate .goimportsignore and make goimports fast

65
6

An implementation of figlet written in Python

1069
100

Live coding a basic Go compiler with LLVM in 20 minutes

265
10

Events

Unexpectedly high disk usage in .config/Code/User/workspaceStorage

Environment

  • OS and Version: NixOS 22.11
  • VS Code Version: 1.73.1
  • C/C++ Extension Version: v1.11.0
  • Other extensions you installed (and if the issue persists after disabling them): N/A
  • If using SSH remote, specify OS of remote machine: N/A
  • Numerous workspaces

Bug Summary and Steps to Reproduce

I hack on a number of large C/C++ projects including the linux kernel and llvm-project.

What I see is that my $HOME/.config directory contains many large binaries (tens of gigabytes, individual files >2.5GiB each), where I would rather it did not. These are the .config/Code/User/workspaceStorage/<randomhex>/ms-vscode.cpptools/.browse.VC.db files.

I consider XDG_CONFIG_HOME is meant to be the place I can store configuration I intend version control and to backup, but I can't easily do this if it contains a lot of large binaries.

I expected these large binaries to live in XDG_CACHE_HOME, which is where I go to delete things if I'm running out of disk space, the place where things should live which can be recomputed.

As an additional note, I see multiple <randomhex> directories where the workspace.json indicates that these several caches of the large caches belong to the same workspace-on-disk, so there is some unexpected duplication going on as well. The workspace.json is identical. I presume these are outdated and unused, and I would naturally clear them out by blatting my $XDG_CACHE_HOME directory, not something I intend to do with my $XDG_CONFIG_HOME.

Expected behavior

I think that large recomputable binary databases should be stored in $XDG_CACHE_DIR/vscode-cpptools alongside the existing ipch cache.

Code sample and Logs

N/A, but these repos are examples of large ones:

https://github.com/llvm/llvm-project
https://github.com/torvalds/linux

Screenshots

No response

Additional context

No response

Created at 2 days ago
issue comment
steamrt copy of libcurl-gnutls.so.4 segfaults when used in steam fhs-env

Heads up, I possibly observed this issue today for another game on nixpkgs-unstable at this commit: https://github.com/NixOS/nixpkgs/commit/5f588eb4a958f1a526ed8da02d6ea1bea0047b9f (10th Nov).

Your workaround of including curlWithGnuTls.out in the extraPkgs appears to work there for me.

Created at 2 weeks ago
started
Created at 3 weeks ago
issue comment
zig c++ fails to compile for x86_64-linux-gnu: "Undefined symbol: __cxa_thread_atexit_impl"

I've noticed that the build of libcxxabi on my system always reports LIBCXXABI_HAS_CXA_THREAD_ATEXIT_IMPL:INTERNAL=1 even though this isn't present. The cmake--debug-trycompile option shows it produces an archive file which nm reports an U (undefined) reference to __cxa_thread_atexit_impl, but it succeeds in creating this file, and check_library_exists then thinks the symbol is available and subsequently sets HAVE___CXA_THREAD_ATEXIT_IMPL. https://github.com/llvm/llvm-project/blob/d3f5f330671e718a0e28598c412d09e9a3b54273/libcxxabi/cmake/config-ix.cmake#L102 I conclude the detection logic is broken, it looks like check_library_exists may be the wrong cmake function for this.

Created at 1 month ago
delete branch
pwaller delete branch zstd-propagatedbuildinputs
Created at 1 month ago
create branch
pwaller create branch zstd-propagatedbuildinputs
Created at 1 month ago
Created at 1 month ago
opened issue
Journalctl spam with 1/sec 'cupsd[1394]: Expiring subscriptions'

Describe the bug

A clear and concise description of what the bug is.

Steps To Reproduce

Steps to reproduce the behavior:

  1. journalctl
  2. Observe lines of the form 'cupsd[1394]: Expiring subscriptions' continuously once per second since boot.
  3. ...

Expected behavior

No log spam.

Additional context

Here is my NixOS configuration which I could identify as 'obviously' relating to printing. So far as I know, printing and scanning work fine.

{
  services.printing.enable = true;
  services.printing.drivers = [ pkgs.hplip pkgs.hplipWithPlugin ];
  services.avahi.enable = true;
  services.avahi.nssmdns = true;

  networking.firewall.allowedUDPPorts = [
    427
    5353
  ];

  nixpkgs.config.allowUnfreePredicate = pkg: builtins.elem (lib.getName pkg) [
    "hplip"
  ];
}

Notify maintainers

cc @matthewbauer

Metadata

Please run nix-shell -p nix-info --run "nix-info -m" and paste the result.

[root@hostname]# nix run nixpkgs#nix-info -- -m
warning: Nix search path entry '/nix/var/nix/profiles/per-user/root/channels/nixos' does not exist, ignoring
warning: Nix search path entry '/nix/var/nix/profiles/per-user/root/channels' does not exist, ignoring
error: file 'nixpkgs' was not found in the Nix search path (add it using $NIX_PATH or -I)

       at «string»:1:9:

            1| (import <nixpkgs> {}).stdenv.hostPlatform.system
             |         ^
(use '--show-trace' to show detailed location information)
 - system: `0`
 - host os: `Linux 5.19.9, NixOS, 22.05 (Quokka), 22.05.20220925.72783a2`
 - multi-user?: `no`
 - sandbox: `yes`
 - version: `nix-env (Nix) 2.8.1`

Note, I'm using flakes and don't have any channels configured.

Created at 1 month ago
issue comment
systemd confinement does not work well with UMask=0077 + non-root user

I hit this error without using specifying a user nor a umask. I've got:

    confinement = {
      enable = true;
      packages = [ pkgs.git ];
      mode = "chroot-only"; # (no effect).
    };

and a script field only.

I see in the journalctl:

systemd[1]: Started <my service>
systemd[60366]: Failed to create directory at /var/empty/usr: Operation not permitted

nixos-version: 22.05.20220819.00e376e (Quokka).

/var/empty appears to have no write permission bit. Which makes some sense if it is to remain empty...

Created at 1 month ago
Created at 1 month ago
pwaller delete branch add-clangs
Created at 2 months ago
Add armv{7,8} clang-{12,13,14,15}

Most recent force push updates the --gcc-toolchain so they match the versions used by the x86 toolchains as per @partouf's comments on discord.

Created at 2 months ago

Add armv{7,8} clang-{12,13,14,15}

And .1 point releases for 12, since they're available.

Signed-off-by: Peter Waller p@pwaller.net

Created at 2 months ago
Add armv{7,8} clang-{12,13,14,15}

Added.

Created at 2 months ago

Add armv{7,8} clang-{12,13,14,15}

And .1 point releases for 12, since they're available.

Signed-off-by: Peter Waller p@pwaller.net

Created at 2 months ago

Add armv{7,8} clang-{12,13,14,15}

And .1 point releases for 12, since they're available.

Signed-off-by: Peter Waller p@pwaller.net

Created at 2 months ago
Add armv{7,8} clang-{12,13,14,15}

Force update changed --gcc-toolchain for the newly added compilers to use GCC 12.2.0 instead.

Created at 2 months ago

Add armv{7,8} clang-{12,13,14,15}

And .1 point releases for 12, since they're available.

Signed-off-by: Peter Waller p@pwaller.net

Created at 2 months ago
Add armv{7,8} clang-{12,13,14,15}

And .1 point releases for 12, since they're available.

I have not tested this yet. I've taken the set of clang compilers I see available from here:

https://github.com/compiler-explorer/infra/blob/af3d0daba0ab5639ffa5061eed094cc861776fa0/bin/yaml/cpp.yaml#L576-L583

Created at 2 months ago
pwaller create branch add-clangs
Created at 2 months ago
issue comment
Convert README to markdown

Thanks for your contribution. I'm unclear on what improvement this brings. For me the new version slightly less legible: the questions and answers are harder to distinguish. In the original they are grouped together so it's easy to skim just the questions. Some of the file has been made into Markdown and some of it has been put into code blocks. Finally, pyfiglet is all about printing to a fixed-width terminal, so I'm fine with the README appearing on github in a fixed-with form which looks like a terminal as it is.

My decision is not final, but please supply a motivation if you want to make a change like this. I'm going to close this PR for now.

Created at 2 months ago
pull request closed
Convert README to markdown
Created at 2 months ago

Update init.py (#88)

Fix missing possible end spacer on caca2tlf converted fonts. (final '@ ' on every char lines)

Created at 2 months ago
pull request closed
Support caca2tlf converted fonts better

Fix missing possible end spacer on caca2tlf converted fonts. (final '@ ' on every char lines)

Created at 2 months ago

Add more fonts to fonts-contrib (#92)

Added fonts from a node js package specifically https://github.com/patorjk/figlet.js.

The headers in these have been cross referenced with the ones currently available with pyfiglet.version '0.8.post1'.

The names of the fonts have been changed to all lower case and spaces have been replaced by _ to conform to the rest of the fonts in the repo.

figlet,js is currently under the MIT license https://github.com/patorjk/figlet.js/blob/master/LICENSE.txt.

Created at 2 months ago
pull request closed
Added fonts

Added fonts from a node js package — specifically figlet.js — currently under the MIT license.

To eliminate duplicate font files the headers have been cross referenced with the ones currently available in:

pyfiglet.__version__ '0.8.post1'.

The following changes to the font names have been made to conform to the rest of the fonts in the repo:

  • all letters have been altered to lower case
  • spaces have been replaced by _

This additionally includes ansi_shadow.flf much like PR85.

Created at 2 months ago