Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Apr 20 17:38
    cgay commented #15
  • Apr 20 17:37
    cgay commented #17
  • Apr 20 17:36
    cgay edited #19
  • Apr 20 14:36
    cgay opened #19
  • Apr 19 02:55
    housel synchronize #1388
  • Apr 19 02:51
    housel synchronize #1388
  • Apr 18 02:21

    cgay on master

    cl-strings: delete everything e… Merge pull request #1387 from c… (compare)

  • Apr 18 02:21
    cgay closed #1387
  • Apr 17 22:45
    housel commented #1387
  • Apr 17 22:23
    housel opened #1388
  • Apr 17 17:49
    cgay synchronize #1387
  • Apr 17 17:17
    cgay opened #1387
  • Apr 17 16:45
    cgay commented #1269
  • Apr 11 20:02

    cgay on master

    Add an assertion for empty stri… (compare)

  • Apr 11 19:46
    cgay commented #1386
  • Apr 11 19:43
    cgay labeled #1386
  • Apr 11 19:42
    cgay opened #1386
  • Apr 11 16:36

    cgay on master

    Fix broken link to DRM errata Merge pull request #123 from fr… (compare)

  • Apr 11 16:36
    cgay closed #123
  • Apr 11 12:17
    fraya opened #123
Peter S. Housel
@housel
apply-mep? Ok, I'll take a look at that
@waywardmonkeys That's from a syntax error within the define interface form in melange?
Bruce Mitchener
@waywardmonkeys
@housel Yeah ... from a trailing comma in a list of things.
I think that the ud2 usually comes from falling off the edge of a return path.
Peter S. Housel
@housel
Yes, that's right
Peter S. Housel
@housel
      // Unhandled case
      ins--block(be, default-bb);
      ins--call-intrinsic(be, "llvm.trap", vector());
      ins--unreachable(be);
Carl Gay
@cgay
Wish I had looked this up years ago... for branch in $(git b --merged master); do git b -d $branch; done
Also wish I hadn't just deleted the master branch. An apparent flaw in my strategy.
Peter Hull
@pedro-w
Reminds me of that picture of a man sawing through the branch he's sitting on
Carl Gay
@cgay
lol
Luckily it's not hard to glue the master branch back on.
Carl Gay
@cgay
WARNING: unused .dylan file: /home/cgay/dylan/pkg/xml-parser/head/src/collect.dylan
Added that warning a few weeks ago for reasons I don't even remember, and it's turning up other unused stuff.
Carl Gay
@cgay
Current status: trying to figure out why Sphinx wants to read sources of submodule documentation for no apparent reason.
Peter S. Housel
@housel
Submodule documentation?
Carl Gay
@cgay
RST files in libraries that are submoduled in.
I'm guessing that there's some config option that says "go find all .rst files and assume they should be included" or something like that, because I can't find any references that would pull them in on purpose.
Peter S. Housel
@housel
I’d noticed it aggressively scanning for .rst files I didn’t want to include, but I hadn’t noticed it being that aggressive
Carl Gay
@cgay
I'm trying to setup a system to document all libraries together, whether they're in OD or not, and for example it seems to be finding the concurrency docs twice because it's a submodule of http.
Potentially less painful path: get packages working fully so we can get rid of submodules, then work on a doc system.
Carl Gay
@cgay
Even the test files: /home/cgay/dylan/workspaces/doc/http/submodules/sphinx-extensions/feed/tests/root/index.rst:13: WARNING: Unknown directive type "latest"
Carl Gay
@cgay
Just out of curiosity, is anyone here able to read this config file comfortably? For me the contrast is beyond hopeless. https://www.sphinx-doc.org/en/master/usage/configuration.html?highlight=conf.py#example-of-configuration-file
Peter S. Housel
@housel
It’s fine if you care nothing about the comments and only want to read the code
Carl Gay
@cgay
haha, yes
Carl Gay
@cgay
Fernando Raya
@fraya
Help, please. Is == the default function to compare a key in <table>? Does make(<object-table>, key-test: \=) change the default comparison function?
Peter S. Housel
@housel
No, you need to subclass <table> and override table-protocol, since you need to supply both an equality test function and a hash function
Carl Gay
@cgay
Carl Gay
@cgay
Basically, I think if I clone everything without submodules and then just generate the top-level index appropriately it should work.
Peter S. Housel
@housel
The “Indices and Tables” in the middle looks a little weird
Carl Gay
@cgay
I assume that's because of the ``:ref:genindex```` in the http and testworks docs and I'll have to figure out how to deal with it. I imagine we'll want to be able to publish docs as part of the shared documentation site and also on their own web pages, so just removing that probably isn't a complete solution.
Carl Gay
@cgay
hmm, how did I not write better tests for subcommand parsing?
guess i got tied up in fixing up existing tests and forgot to write new ones for new functionality.
Fernando Raya
@fraya
@cgay @housel Thank you. Somehow my thanks didn't show in gitter.
Fernando Raya
@fraya

I'm trying to make a table for <character> following the <string-table> example:

define sealed class <character-table> (<table>)
end;

define sealed domain make (singleton(<character-table>));
define sealed domain initialize(<character-table>);

define sealed method table-protocol
    (table :: <character-table>)
 => (test :: <function>, hash :: <function>)
  values(method (x :: <character>, y :: <character>) x = y end, object-hash)
end method table-protocol;

If I import table-extensions the compiler show this warning:

Serious warning - Unexpected token "::".
                 --
          (table :: <character-table>)
                 --

If I remove table-extensions and put the line like in

define sealed method table-protocol
  (table :: <character-table>); // note the ; like <string-table> example

show this warning:

      --
       => (test :: <function>, hash :: <function>)
       --

I'm a little confused. Removing the ; it compiles without warnings.

Peter S. Housel
@housel
You have a return value spec (table :: <character-table>) => (test :: <function>, hash :: <function>) you can put ; at the end of that but not between the parameters and the return values
Carl Gay
@cgay
@fraya the first issue, Unexpected token "::", is probably because you're using an older version of table-extensions, in which table is a macro. Pretty sure I fixed that after our most recent release, so you'd have to build from source to avoid it. But you can just name the parameter tbl instead.
(I feel like this is really a bug in our parsing technology though, because you can't have a macro call in a parameter list anyway.)
(Er, in that position in a parameter list.)
Fernando Raya
@fraya
@housel Sorry, you are right. I saw the ; after the parameter but in the example It was after the return. Thank you.
@cgay Ok. I'll build from source. Thank you.
One more question: It is necessary in values the anonymous method or I can use \=?
Peter S. Housel
@housel
You can use \=
Of course equivalent <character> values are always == so you don't need to subclass for that, the default <object-table> will work fine in that case
Fernando Raya
@fraya
@housel It make sense. Thanks.
yc
@cloutiy
Hi is there an "awesome dylan" curated list of libraries, projects etc...?
Peter S. Housel
@housel
Not yet, other than the blanket assumption that everything under https://github.com/dylan-lang or https://github.com/dylan-foundry is awesome :grin:
yc
@cloutiy
Lol obviously.