and friends.
>>
>>
>> This is the API I prefer: units. IMHO, it is more important to keep
>> consistency with Elixir libraries.
>>
>> -bt
>>
>> On Thu, Mar 7, 2024 at 11:02 PM José Valim wrote:
>>
>>> It is worth noting that D
I’ll open one now. I’m also happy to investigate adding this,
> but will need a nudge as to where to start looking. Not high priority, but
> I think will be a great quality of life improvement for users.
>
> On Apr 2, 2024, at 5:40 PM, José Valim wrote:
>
> We could try that. Please open up
ant remote
> calls at that point?
>
> On Apr 2, 2024, at 5:33 PM, José Valim wrote:
>
> Unfortunately this is hard. The simplest way to implement this would be by
> checking on every remote call if the module is available and if the
> function is macro, which would considerably slow down
Unfortunately this is hard. The simplest way to implement this would be by
checking on every remote call if the module is available and if the
function is macro, which would considerably slow down the compiler. The
reason why we have "require" is exactly so we don't need to pay this price.
We
I don't see a strong need to make this part of ExUnit, especially because:
1. We should avoid mutating the global state in tests
2. We should avoid reading system environment variables in code (use
config/runtime.exs instead)
Those can add to their own suites if necessary. :)
On Wed, Mar 27,
I would love to see a PR that implements translation for gen_statem, so we
can evaluate if it makes sense to merge it into Elixir.
On Wed, Mar 27, 2024 at 4:10 PM Cameron Duley
wrote:
> Several popular Elixir libraries use gen_statem internally, given it's
> very useful for managing data over
best data structure for the actual use
> case)
>
> That's all I found.
>
> Best regards,
> Oliver
>
> On Wednesday, March 20, 2024 at 11:40:15 AM UTC+1 José Valim wrote:
>
>> Can you please provide a link to the previous discussions? I recall
>> deal
Can you please provide a link to the previous discussions? I recall dealing
with some complexities around finding and not finding elements as well.
Thanks!
On Wed, Mar 20, 2024 at 11:37 AM 'oliver@googlemail.com' via
elixir-lang-core wrote:
> Hello, okay I checked.
>
> Well, there was a
improved the performance overall.
>
> I think I didn't then go, and profile specifically if removing the version
> check alone will improve the performance by itself. So all I have to back
> up that the version check is the root cause, is fprof.
>
> On Friday, March 15, 2024 at 8:2
ns
>> Regex.run/21.98 M 504.74 ns ±5851.21% 416 ns
>> 542 ns
>>
>> Comparison:
>> :re.run/3 2.88 M
>> Regex.run/21.98 M - 1.46x slower +157.84 ns
>> ```
>> On Friday 15 March 2024 at 07:20:11 UTC+1 jan.k...@gma
>> I read the commit, and I don't it fixes what our actual problem was. See
>>> my comment above. The problem is the actual call to :re.version, not the
>>> recompilation of the regex
>>>
>>> On Thursday, March 14, 2024 at 4:37:43 PM UTC+1 José Valim wrote:
l call to :re.version, not the
> recompilation of the regex
>
> On Thursday, March 14, 2024 at 4:37:43 PM UTC+1 José Valim wrote:
>
>> I have pushed a fix to main. But also note we provide precompiled Elixir
>> versions per OTP version. Using a matching version will alway
I have pushed a fix to main. But also note we provide precompiled Elixir
versions per OTP version. Using a matching version will always give you the
best results and that's not only about regexes. :)
On Thu, Mar 14, 2024 at 2:20 PM Jan Krüger wrote:
> I've recently had to work on a code base
Elixir (also relevant for
>> extending the use of Duration later on)
>> - Reduced cognitive load as the time units are always spelled the same
>> regardless of the context
>>
>> The reasons for singular do outweigh the reasons for plural, so unless
>> we'r
keys exclusive to durations. Another reason
> for going with plurals is that it _should_ make migrating from some
> libraries to the standard library relatively straight forward (with the
> exception of microseconds).
>
> On Wednesday, March 6, 2024 at 11:07:52 PM UTC+1 José Valim
After a quick glance on other programming languages, it seems Python, Java,
Rust, and C# all have plural names. Erlang also uses plural in its helper
functions in the timer module. So we might want to follow suit.
On Wed, Mar 6, 2024 at 23:03 José Valim wrote:
> We discussed plural vs singu
ficial for everybody.
>
> PS. I would expect plural in the duration fields.
>
> On Wed, Mar 6, 2024 at 8:23 PM José Valim wrote:
>
>> The main argument for having it in core is:
>>
>> * It integrates directly with the Calendar behaviour
>> * We could pr
The main argument for having it in core is:
* It integrates directly with the Calendar behaviour
* We could provide built-in sigils in the future to create readable
durations, such as ~P[3 hours and 10 minutes]
* Postgrex, Explorer, CLDR, etc all implement their own version of
durations
The low-level integration bits are part of Erlang. If this functionality
exists in Erlang today, we could surface it up, but I am afraid it doesn't
right now. I believe there is a plan to provide more file system bindings
there in the future, but there is no precise schedule.
On Sat, Mar 2, 2024
the
> size—so I'm proposing that addition. Just the single guard!
> On Thursday, December 24, 2020 at 1:37:40 AM UTC-6 José Valim wrote:
>
>> Because if I have a MapSet, I would prefer to call MapSet API to make the
>> intent clearer.
>>
>> On Thu, Dec 24, 202
as maps are missing one very important
>>> function needed for implementing Lua and that is to be able to efficiently
>>> step through a tree. I need a 'first' function which returns the first
>>> key-value pair and then a 'next' which returns the next pair after a given
>
her
>> in the docs.
>>
>> Thank you for your ongoing development of the language!
>>
>> On Tuesday, February 27, 2024 at 12:34:19 PM UTC+1 José Valim wrote:
>>
>>> Doing this now would certainly be a breaking change as it would break
>>> all usages o
Doing this now would certainly be a breaking change as it would break all
usages of mix release today.
The easiest way to go about this is to add a "path: " key to your release
configuration in your mix.exs, in which you can employ any triple target
that you want.
AJ
>>
>> Le mercredi 21 février 2024 à 20:21:56 UTC+9, José Valim a écrit :
>>
>>> +1 for the proposal, but it has to be implemented more efficiently than
>>> what was described here and we also need to add drop!. PRs welcome.
>>>
>>> On Wed
+1 for the proposal, but it has to be implemented more efficiently than
what was described here and we also need to add drop!. PRs welcome.
On Wed, Feb 21, 2024 at 10:59 AM Artur Plysiuk wrote:
> I just came here to add the same proposal.
>
> четвер, 28 грудня 2023 р. о 11:40:07 UTC+2 DaTrader
Correct, Elixir rarely adds aliases, because it adds to the learning curve
(and Enum is already a large module) and it also eventually becomes a
struggle for consistency (this code base must “only use select” or “only
use filter”).
Plus I personally dislike select because it partially conflicts
My general gut feeling is: compared to the amount of code in Erlang/OTP,
Elixir, dependencies, and the code that Ash generates, the amount of code
in Ash used for code generation is going to be minimal. And all of that to
save memory on a box that most likely still needs to load hundreds of rows
ped lambda calculus. Avoiding assignments feels like a
highly arbitrary restriction to impose on your code. :)
On Fri, Jan 26, 2024 at 23:28 José Valim wrote:
> One of the goals behind Elixir is to provide a certain amount of syntax
> that translates to a clear abstract syntax tree. T
One of the goals behind Elixir is to provide a certain amount of syntax
that translates to a clear abstract syntax tree. The syntax available to
create the language should also be available for others to extend language.
Operators are generally tricky because they are hardcoded in the language.
This is complicated because "truncate" is about removing the precision, and
we cannot have a minute precision. The way to support it properly would be
to allow seconds to be nil and then we would print it as "2022-12-31
13:45", without the seconds component, but I think this would be a large
> 1. Do you think anyone would purposefully name a variable a type? (false
positives.)
Yes, we do it in Elixir all the time. Because if I am implementing
String.downcase/1, the name `string` is a very good name for the argument:
def downcase(string)
And since we use names in the docs, calling
> you think would be the level of effort to have some rough take on it? If
> it's a rather small change I would consider building my own fork just to
> use personally.
>
> my three ideas.
> 1) per warning type suppression
> 2) warning suppression entirely
> 3) separate war
Please see: https://github.com/elixir-nx/nx
The forum may be a better place to ask questions, as this mailing list is
for language implementation discussion and features.
On Tuesday, January 9, 2024 at 3:40:03 PM UTC+1 jonkl...@gmail.com wrote:
> I was wondering if there was anyway to allow
You can do a pass after the fact replacing <<0>> by the replacement
character.
On Sun, Dec 17, 2023 at 10:43 AM Daniel Kukula
wrote:
> Any chance to have an option to also escape null byte in
> String.replace_invalid ??
>
> In utf-8 <<0>> is a valid character but there are problems with it:
> -
In addition to what Wojtek said and this being a breaking change, you can
pattern match on the empty list vs [head | tail] directly, without the
overhead of calling a function to just pattern match on ok/error.
--
You received this message because you are subscribed to the Google Groups
Correct. When it comes to strings, Elixir generally assumes that the string
has been validated before entering the system. If all functions were to
validate they are indeed strings, it will become quite expensive.
That said, I think your solution of changing the spec is the correct one,
This feature is now in main and it happens automatically.
On Tue, Nov 7, 2023 at 12:36 Dario Heinisch
wrote:
> Hi,
>
> Running mix format on larger projects can be slow since it checks every
> file. I don't think there is currently a way to run it on changed files
> only. Users could run `mix
We will inspect the function too. :) And we will see how it evolves.
On Thu, Nov 2, 2023 at 1:01 PM Panagiotis Nezis wrote:
> What if an alias is a function?
>
> On Thursday, November 2, 2023 at 1:49:43 PM UTC+2 José Valim wrote:
>
>> Before going down this road, I would firs
Before going down this road, I would first like to add "mix help --aliases"
and we could show the alias definition (i.e. what it invokes). I think it
would already be a massive improvement to the UX. Would you like to give
this a try?
On Thu, Nov 2, 2023 at 12:45 PM Panagiotis Nezis wrote:
>
src/bytes/bytes_test.go;l=1157
>
> On Tue, Oct 31, 2023 at 2:09 PM José Valim wrote:
>
>> Does the specification provide tests for us to include? Otherwise we can
>> include enough tests for full line coverage and a “brute force”/property
>> test commented out.
ly uncertainty - How much is prudent and in what manner?
> On Tuesday, October 31, 2023 at 12:35:48 PM UTC-4 José Valim wrote:
>
>> Folks, I am following up on this, where did we land?
>>
>> The new implementation is roughly ~70LOC for UTF-8, so at first I don't
>> see an iss
module makes good sense too as José says but
> that's a longer "sales" and implementation cycle.
>
> On Saturday, October 7, 2023 at 7:40:50 PM UTC+11 José Valim wrote:
>
>> Hi Cameron,
>>
>> If the goal is to include this handling for UTF-16 and UTF-32, I
Unfortunately we should not rely on the Docs chunk, because it is not
expected to be available. We could add infrastructure to this but I would
say only: :macros give a good enough approximation (and they'd cover
patterns and guards).
On Sunday, October 22, 2023 at 7:31:27 AM UTC+2
It delegates to Phoenix.Template.
On Mon, Oct 23, 2023 at 07:48 'Alexander Steppke' via elixir-lang-core <
elixir-lang-core@googlegroups.com> wrote:
> Thanks for the feedback, I've completely forgotten about
> `__mix_recompile__?`! @Jose: I couldn't find anything implementing it in
> LiveView or
Timex.shift is probably the best reference:
https://hexdocs.pm/timex/Timex.html#shift/2
Keep in mind it has many pitfalls, especially related to timezones and
multiple calendars, so it is probably necessary to draft a proposal before.
There is also a debate of which module it should be defined
There is a callback that you implement for your modules called
__mix_recompile__? Which you can implement with your own logic to detect
how files and directories change. You can See how embed_templates does it
for examples. :)
On Thu, Oct 19, 2023 at 11:01 'Alexander Steppke' via elixir-lang-core
Our goal is to revisit this relatively soon. If the type system work
succeeds, we will introduce a new syntax for defining structs and we will
make documentation a core part of field names.
On Thu, Oct 19, 2023 at 07:59 Wojtek Mach wrote:
> > PS. Having already typed this, it could be possible
Enum always returns lists by design, because it was designed to be a single
abstraction that can:
1. enumerate elements while skipping
2. support both eager and lazy with the same protocol
3. support halting (such as take)
4. support zipping
5. does not cause dangling resources (so you can work
Hi Cameron,
If the goal is to include this handling for UTF-16 and UTF-32, I suggest
proposing this to Erlang/OTP as new functions in the "unicode" module.
Otherwise, Elixir only has facilities to deal with UTF-8. You could propose
such a feature in their issues tracker.
Also note that "rolling
This is a small function that can be part of your application or a package,
we do not plan to add it to the standard library at the moment. Thank you.
:)
On Wed, Sep 27, 2023 at 10:40 PM Weslei Juan Novaes Pereira <
weslei...@gmail.com> wrote:
> The current Map.from_struct/1 works only in the
For what is worth, the error message on main says:
---
iex(1)> [1, 2, 3][0]
** (ArgumentError) the Access module does not support accessing lists by
index, got: 0
Accessing a list by index is typically discouraged in Elixir, instead we
prefer to use the Enum module to manipulate lists as a
Index-based access on a list is very frequently an anti-pattern. Allowing
list[i] means it will be easier to write non-efficient versions of
algorithms without a second thought for those coming from imperative
languages, instead of using the functions in Enum. It means someone can
write this:
for
Sounds good to me. PR welcome!
On Mon, Sep 18, 2023 at 6:41 PM Michal Śledź
wrote:
> System.put_env/1 erases given key when a value is nil.
>
> The proposal is to make System.put_env/2 symmetric with the
> System.put_env/1 and also erase a key on the nil value.
>
> The spec would change from:
>
This is a bug. We should emit the trace before they are inlined. Can you
please do a bug report?
On Mon, Sep 11, 2023 at 1:55 AM Mitchell Hanberg <
the.mitch.hanb...@gmail.com> wrote:
> Currently, by the time *_function and *_macro traces are emitted, I
> believe that inlined functions have
The Elixir Team has no plans to participate but other projects in the
community may be interested. I recommend starting a discussion on Elixir
Forum. The Erlang Ecosystem Foundation was also used to participate in the
past.
On Tue, Sep 12, 2023 at 5:27 AM Apoorv Gupta
wrote:
> Will elixir be
If we want to add more parallelism later, then it is worth discussing
future developments of the API now, so we don't put ourselves into a
corner. If the answer is no, it is clear we are only parallelizing tests or
parallelizing modules, but never both. And if we want to do both in the
future, we
> 3. return both modules from ExUnit.Server.take_async_modules/1
> 3. running
>1. spawn_tests/3 before spawn_modules/2
> 2. spawn_tests/3 inside spawn_modules/2 if it's a per_test async
> module
>
> I'd prefer 1.2 + 2.2 + 3.1:
>
>- save asy
Yes, it is desirable and it has come up in the past:
https://github.com/elixir-lang/elixir/pull/11949#issuecomment-1177262901
Although I think async: :per_module is what most people want, since the
tests in the same module tend to access the same resource, opting-in for it
to be per test will be
A PR is welcome!
On Mon, Aug 28, 2023 at 10:37 PM Anderson Cook
wrote:
> Under the hood `Access.at` calls `Enum.at` which takes an optional
> default. I propose that it accept the same optional default, much like
> `Access.key`. This should be a minimally-invasive addition.
>
> --
> You
mix deps.tree could help you here. My suspicion is either a path or a git
dependency, as their deps are not in the lock file. A branch called
"allow-poison-5" is highly suspicious.
We could also introduce a mix deps.why or a mix deps.tree --filter option
to help narrow down those cases. Proposals
If someone submits a PR for Erlang/OTP, they will automatically work for
Elixir, so please go down that road. :)
On Mon, Jul 17, 2023 at 03:58 Fernando Canizo
wrote:
> Request (TL;DR):
>
> Please add support for Home and End keys in `iex` shell.
>
> I know one can use CTRL-A and CTRL-E
Phoenix standardized on setup, but it does not include the compile command.
Not sure how useful build with compile is, if you run mix test instead of
mix compile, build won’t help it.
On Fri, Jul 14, 2023 at 09:11 Andrea Leopardi wrote:
> I personally like the idea of a mix build alias defined
: is not an operator at the user level for JS but it behaves like one
syntactically. You can add or remove spaces on either side and it works.
That’s not true for Ruby or Elixir as moving the spaces around is either
invalid or has a different meaning.
--
You received this message because you are
n that some such
>>> syntax should only work for structs: with Elixir LS today, starting to type
>>> a `key:` in a struct/map literal does indeed suggest from the list of known
>>> struct keys. I don't see this being impossible in LS tooling today, but
>>> als
ts: with Elixir LS today, starting to type
> a `key:` in a struct/map literal does indeed suggest from the list of known
> struct keys. I don't see this being impossible in LS tooling today, but
> also don't know much about what is possible with the language server
> protocol today. :)
&g
> waiting for Elixir 2.0 and/or atom garbage collection was the right play
> here.
>
> On Thursday, June 29, 2023 at 2:33:22 AM UTC-5 José Valim wrote:
>
>> > I would argue that if we want to support only atoms, but make it clear
>> that the syntax only appli
s, and offers a syntax construct that can be used in more places
> (as a specific example, error = "rate limit exceeded"; $:error # return
> error tuple. Apologies if it feels like I am trying to torpedo other
> solutions, that is not my intent at all.
> On Thursday, June
Hi Chris Keele, thank you for the excellent proposal. I just want to add
that I agree with Paul that we don't need to support both strings and
atoms, but it must be clear that it applies to either strings or atoms (if
it supports only one of them) and the reason for that is because otherwise
it
There is a very extensive proposal and discussion here:
https://elixirforum.com/t/proposal-add-field-puns-map-shorthand-to-elixir/15452
There is also very clear lines draw by myself on what this feature should
look like:
https://groups.google.com/g/elixir-lang-core/c/XxnrGgZsyVc/m/KgJwV8e-CgAJ
re sensitive to such things.
>
> And possibly a relevant question in both cases: How well does this work
> with inspect limits that end up eliding parts of the data?
>
> On Sunday, May 28, 2023 at 12:35:40 PM UTC-4 José Valim wrote:
>
>> If we are going down this path, then my sugge
If we are going down this path, then my suggestion would be to add this to
all of Elixir itself, starting with function clause errors. :) If someone
wants to explore this path, please go ahead!
On Sun, May 28, 2023 at 5:13 PM Ben Wilson wrote:
> Agreed. Can the formatter get invoked on output
I am not sure we want to encourage those functions. The mentioned languages
have a static type system, which means you know the shape upfront.
In Elixir, we don’t, so we want to prefer pattern matching as much as
possible, as it also helps assert on tuple size and help catch mistakes.
On Mon,
I agree this is a good proposal but Wojtek unfortunately touched on a point
that I don't know how to address.
There are currently 4 *_lazy functions and, if we add this functionality to
one of them, it would soon be expected that all of them support it. The
problem is that the functions that
Uppercase sigils do not supported interpolation, so that's a no-go.
Note you can also get the compile time behaviour today by doing @uri
URI.new!("...").
On Thu, May 18, 2023 at 4:25 PM Ben Wilson wrote:
> Question: Would this sigil support interpolation? Is that too clever?
>
>
We discussed it internally and the only concern at the moment is if it
won't look inconsistent if we have ~UTC but still keep ~D / ~T / ~N.
On Tue, May 16, 2023 at 9:59 AM Wojtek Mach wrote:
> Hi, I'd like to propose deprecating ~U in favour of ~UTC.
>
> I don't think the change would
ems etc. etc.
>>
>> But because of a lot of questions and doubts it's clear that's the
>> requester responsibility to propose detailed description of the API, take
>> into account all pros and cons, describe how they will affect the whole
>> ecosystem and whether the requeste
One addition: “features” makes sense for Rust because the contents of its
“module body” cannot be dynamic as in Elixir. So if they want to provide
this feature in the first place, it must be done as part of the compiler.
Elixir can execute any Elixir code when defining modules, which is why it
is
> I get that. I am coming from a different perspective than yours that
would help manage such dependencies based on how people deal with the
dependencies today, such as using Code.ensure_loaded; and **try** to tackle
the problem with a more "robust" idea.
Here is the issue: the proposal has not
ending on B and C, whereas B also
>>>>> optionally depends on C. Then A might be included as `{:a, "~> …", config:
>>>>> [b: true, sigils: false]}` and then A would know at the compilation stage
>>>>> it should “flag C as used, and flag B as used
mat` friendly because then its much
> much easier to remove it
>
> On Tuesday, May 9, 2023 at 9:51:54 AM UTC-6 José Valim wrote:
>
>> There is a "structs: false" option but I am not sure if that is supported
>> when diffing.
>>
>> But also keep in min
There is a "structs: false" option but I am not sure if that is supported
when diffing.
But also keep in mind there are structures that represent memory data (such
as PIDs and REFs) and those can never really be copy and pasted. And
sometimes, like above, it shows in a special style because it is
+1 for more tools for working with CLIs but that's definitely outside of
the scope of the language itself. For more information, we have a document
on the development guidelines and team:
https://elixir-lang.org/development.html - Discussions about additional
tooling to the ecosystem are probably
lists is realized.
>
> On Friday, April 28, 2023 at 9:19:35 AM UTC-5 José Valim wrote:
>
>> I am wondering if we should allow keyword lists in those cases instead
>> but it can be confusing if folks then use the direct Erlang APIs.
>>
>> On Fri, Apr 28, 2023 at 3
l#module-boot-configuration
>
> Say I want to change the file path for different envs.
>
>
> José Valim schrieb am Donnerstag, 30. März 2023 um 12:34:54 UTC+2:
>
>> Can you please provide examples of where this may happen?
>>
>> On Thu, Mar 30, 2023 at 12:08
Oh, nice!
The builds are automated, so a PR that adds signatures would be welcome.
Although, afaik, you should be safe with a SHA512 checksum.
On Tue, Apr 25, 2023 at 08:19 Benedikt Reinartz wrote:
> Hi,
>
> I have been working on making Elixir available to rebar3 users without the
> need to
I would love to see a PR that tries approach 1. And we can expand from
that. :) Thank you.
On Wed, Apr 19, 2023 at 2:40 PM Bruce Tate wrote:
> I have been going over my notes and noting the places my students have
> trouble with standard Elixir documentation. One of them is this example
> code
Agreed. I just wouldn't call it to_server but "state". Easier to bikeshed
on code, so a PR is welcome. :)
On Mon, Apr 10, 2023 at 4:14 PM Bruce Tate wrote:
> I would like to propose that we change the base example for GenServer. For
> those not familiar with it, here it is:
>
> defmodule
Hi Yordis, thank you for the proposal.
I would like to see it expanded to say how such module you would look like
and how it would interact with all the other Calendar modules.
Kip also had a library where the dates and datetimes themselves could be
duration and I would want to see them explored
Hi Zach, thanks for the proposal! A PR for "fetch_alias_as" is welcome, we
even use private APIs in Kino.VegaLite and friends which we could stop
using if this function existed. :)
On Sat, Apr 1, 2023 at 5:48 PM Zach Allaun wrote:
> Upon reflection, the change that I'm proposing to
Can you please provide examples of where this may happen?
On Thu, Mar 30, 2023 at 12:08 PM lostko...@gmail.com
wrote:
> We'll be seeing deeper integration with :logger in the upcoming elixir
> version, but :logger is using maps for configuration, which don't merge in
> config, but require
The point of returning gap or ambiguous is exactly that they may or may not
be errors and it is up to you to handle them. For example, some application
may choose to surface those to the user and ask more information to resolve
the ambiguity. In a statically typed language would have to handle
Is this feature supported in Erlang? I am afraid it is not, which means we
can’t support it.
On Thu, Mar 30, 2023 at 06:03 Rafael Willians Moraes
wrote:
> Right now it's possible to declare multiple clauses for `@spec` as we can
> see in Typespecs documentation
>
"mix deps.unlock --unused" is generally the way to go about it. You can
also add "mix deps.unlock --check-unused" to CI.
On Sat, Mar 18, 2023 at 2:13 PM Jonathan Arnett
wrote:
> I fully agree that this needs to be more discoverable. Every year or so,
> I remove a dependency and then have to go
Hi Austin and Zach,
For the Bamboo Phoenix case, they should declare phoenix_view as an
optional dependency. Then either conditionally compile code or use
"@compile {:no_warn_undefined, Phoenix.View}" in the relevant module. Note
Elixir v1.14 introduce a --no-optional-deps flag to help library
Yes, even another library. You can run it in test while you are in test.
And the error message says it must be a relative path pointing inside the
current directory, IIRC.
On Fri, Mar 17, 2023 at 20:26 Brandon Gillespie wrote:
> On 3/16/23 5:00 PM, José Valim wrote:
>
> > You sh
You shouldn't need to bundle test code into your production code.
1. Deps are compiled in prod but in_umbrella deps are compiled in the same
environment as the current project. You can achieve this by specifying the
env option to the dependency, such as "env: Mix.env()", and it works for
any
Can you please provide a small app that reproduces the issue?
Even if the module comes from a transitive dependency, Elixir should be
able to see it and avoid the warning.
On Fri, Mar 10, 2023 at 5:00 PM Brandon Gillespie wrote:
> I'm in the process of breaking our larger/monolithic app into
re checks are only inclusive, it
> should be fine. But you can also think if a library named A assumes that
> dependency C is compiled without some flag, and library B assumes C is
> compiled with said flag, you will end up with conflicting behaviour.
>
> I don't have answers to those questions but
> needed to get the desired feature. What I would like to have is to focus on
> the feature itself, leaving deps and their versions to library maintainers.
>
>
>
>
>
> wtorek, 7 marca 2023 o 14:45:03 UTC+1 José Valim napisał(a):
>
>> Hi Michał,
>>
>>
Hi Michał,
Thanks for the proposal. Your initial description makes me think there may
exist bugs which we would need to investigate first.
2. Users of a library with optional dependencies have to include all
optional dependencies in their mix.exs
This should not be required. You only need to
1 - 100 of 1118 matches
Mail list logo