hashing tests fail on 32-bit linux #354
Labels
bug
Indicates an unexpected problem or unintended behavior
Comments
|
This is a regression. |
JeffBezanson
added a commit
that referenced
this issue
Jan 28, 2012
==(Rational{T<:Integer},Number) at rational.j:105
==(T<:Number,T<:Number) at promotion.j:75
==(Rational{T<:Integer},Integer) at rational.j:98
which in turn caused issue #354 together with the recent changes to
unsigned integer arithmetic.
This is tricky because while (Rational,Integer)<:(Rational,Number),
(Rational,Number) has overlap with (T,T) but (Rational,Integer) doesn't.
The new sort order is better, and exposed 2 new method ambiguities in
rational which are also fixed here.
StefanKarpinski
pushed a commit
that referenced
this issue
Feb 8, 2018
* Support comparison between all Period subtypes Note that defining `==` methods with `Base.:(==)` does not work on Julia 0.4 * Add entry to README
cmcaine
pushed a commit
to cmcaine/julia
that referenced
this issue
Nov 11, 2022
* Add authors and contributors to Practice Exercises For each Practice Exercise, we've looked at the commit history of its files to see which commits touched that Practice Exercise. Each commit that we could associate with a Practice Exercise was used to determine authorship/contributorship. ### Authors 1. Find the Git commit author of the oldest commit linked to that Practice Exercise 2. Find the GitHub username for the Git commit author of that commit 3. Add the GitHub username of the Git commit author to the `authors` key in the `.meta/config.json` file ### Contributors 1. Find the Git commit authors and any co-authors of all but the oldest commit linked to that Practice Exercise. If there is only one commit, there won't be any contributors. 1. Exclude the Git commit author and any co-authors of the oldest commit from the list of Git commit authors (an author is _not_ also a contributor) 2. Find the GitHub usernames for the Git commit authors and any co-authors of those commits 3. Add the GitHub usernames of the Git commit authors and any co-authors to the `contributor` key in the `.meta/config.json` file We used the GitHub GraphQL API to find the username of a commit author or any co-authors. In some cases though, a username cannot be found for a commit (e.g. due to the user account no longer existing), in which case the commit was skipped. ## Renames There are a small number of Practice Exercises that might have been renamed at some point. You can ask Git to "follow" a file over its renames using `git log --follow <file>`, which will also return commits made before renames. Unfortunately, Git does not store renames, it just stores the contents of the renamed files and tries to guess if a file was renamed by looking at its contents. This _can_ (and will) lead to false positives, where Git will think a file has been renamed whereas it hasn't. As we don't want to have incorrect authors/contributors for exercises, we're ignoring renames. The only exception to this are known exercise renames: - `bracket-push` was renamed to `matching-brackets` - `retree` was renamed to `satellite` - `resistor-colors` was renamed to `resistor-color-duo` - `kindergarden-garden` was renamed to `kindergarten-garden` ## Exclusions There are some commits that we skipped over, and which thus didn't influence the authors/contributors list: - Commits authored by `dependabot[bot]`, `dependabot-preview[bot]` or `github-actions[bot]` - Bulk update PRs made by `ErikSchierboom` or `kytrinx` to update the track * Restore formatting * Manually fix authors A few noteworthy changes: * JuliaLang#162 has manually been excluded as it was a mass-PR done by changing a global template * PRs to several exercises that changed the exercise itself were all included * All reviews were included as contribution; reviewers where added where missing; however due to GitHub's UI hiding reviewers in the history, I may have missed reviews done on PRs that weren't on the initial PR of an exercise * 6c8a1db caused weirdness due to having two parents from merging the default branch into the PR branch. This has manually been resolved as well If you feel like you've been removed or added unjustly, let me know and I'll re-add you in a follow-up PR. Sorry if that happened! Co-authored-by: Sascha Mann <git@mail.saschamann.eu>
Keno
pushed a commit
that referenced
this issue
Oct 9, 2023
This supports SSAValue args for cglobal while also fixing a bug in which a GotoIfNot ended up going to the wrong statement. The latter was a consequence of incorrect :call unnesting. Fixes #455 Fixes #454 Fixes #415 Fixes JuliaDebug/Debugger.jl#275 Improves #354
Keno
pushed a commit
that referenced
this issue
Oct 9, 2023
This has an llvmcall doing an atomic operation...just avoid the hassle and call it directly. Fixes #354
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
$ make test
JULIA test/default
* core
* numbers
* strings
* corelib
* hashing
assertion failed: ||(!(isequal(a,b)),hash(a)==hash(b))
in error, /home/viral/julia/j/error.j:5
in runtests, lib.j:3
in load, src/boot.j:193
in process_options, /home/viral/julia/j/client.j:88
in _start, /home/viral/julia/j/client.j:122
at hashing.j:18
at default.j:6
make[1]: *** [default] Error 1
make: *** [test] Error 2
The text was updated successfully, but these errors were encountered: