Is there a way to initialize variables with erl? - erlang

I'd like to do things like L=lists, X=very_long_module_name. There's some modules I use a lot and I'd rather type X:function() than very_long_module_name:function(). The .erlang file seems to be able to run commands but it doesn't keep any variable assignments. For all know you can only use .erlang to add more things to path.

maybe you could use macros? i.e. -DL=lists and then ?L:function() ?

Related

How can a Bazel `repository_rule` adjust a `label_flag` (or a `config_setting` more generally)?

I can create a label_flag in Bazel to allow command line flags to in turn be matched with a config_setting in a Bazel BUILD file.
However, I'd like to not hard-code the default value of the label_flag, and instead compute a good default based on the system when evaluating a repository_rule (or some other part of the WORKSPACE file).
The best (but awful) way I've come up with to do this is to have the default value loaded from a .bzl file that is generated using the template function on the repository_ctx.
I feel like generating a new file by doing textual substitutions probably isn't the right way to do this, but I can't find anything else. Ideas? help?
Generating a bzl file using the repository rule that inspects the host system is the only way to achieve what you need right now. So you're holding it "right" :)

How to deobfuscate this?

I obfuscated this script using some site
But i'm wondering how to deobfuscate it? can someone help me?
i tried using most decompilers and a lot of ways but none has worked
local howtoDEOBFUSCATEthis_Illll='2d2d341be85c64062f4287f90df25edffd08ec003b5d9491e1db542a356f64a488a1015c2a6b6d4596f2fa74fd602d30b0ecb05f4d768cd9b54d8463b39729eb1fe84630c0f8983f1a0087681fe4f2b322450ce07b
something like that for an example.
the whole script: https://pastebin.com/raw/fDGKYrH7
First reformat into a sane layout. a newline before every local and end will do a lot. Then indenting the functions that become visible is pretty easy.
After that use search replace to inline constants. For example: local howtoDEOBFUSCATEthis_IlIlIIlIlIlI=8480; means you can replace every howtoDEOBFUSCATEthis_IlIlIIlIlIlI with 8480. Though be careful about assignments to it. If there are any then it's better to rename the variable something sensible.
If an identifier has no match you can delete the statement.
Eventually you get to functions that are actually used.
Looking at the code it seems to be an interpreter implementation. I believe it's a lua interpreter
Which means that you'll need to verify that and decompile what the interpreter executes.

User name in .bazelrc

I would like to add this to my .bazelrc, but the $(whoami) doesn't expand like if it was in a shell.
startup --output_user_root=/tmp/bazel/out/$(whoami)
It produces the literal result:
/tmp/bazel/out/$(whoami)/faedb999bdce730c9c495251de1ca1a4/execroot/__main__/bazel-out/
Is there any way to do what I want: adding a name/hash to the option in the .bashrc file?
Edit: what I really want is to set the outputRoot to /tmp/bazel/out without using an environment variable and to let bazel create it's user and workspace hash directories there.
You can run Bazel from a wrapper script. In fact, that's exactly what the bazel binary is (at least on Linux): it's a wrapper script that calls bazel-real. You can edit this wrapper script if you like, or rename it to bazel.sh and write your own wrapper.
/usr/bin/bazel is a script which looks for //tools/bazel, and if it exists, calls it. Otherwise, it calls bazel-real. This lets you check Bazel into your repo, or otherwise modify how it gets called. We use that to download a specific version of bazel, extract it, and then call it.
I would recommend creating //tools/bazel, and having that do your modification. It can then either call a versioned version of bazel, or call bazel-real. That keeps your modifications local to your repo rather than global.

How do I find a particular sub directory using Ant and then use it to create a symlink?

I need to create a symlink to a sub-directory using Ant. The issue is that I don't know where the target sub-directory is.
To create a symlink with ant I do this:
<symlink link="${parent.dir}/FOO/linkname" resource="${parent.dir}/BAR/target"/>
But I don't know what BAR is called in advance so I need to do a search for "target" under parent.dir and then pass the one result into the resource.
Is this possible using fileset? Or another way?
It might be possible to use a fileset but that might give you several symlinks or none.
A much better approach is to define the path to BAR in a property. If there is a dynamic part in this path, change the code so that Ant evaluates the dynamic part and everyone else uses Ant's value.
The typical example here is that the path contains a version or timestamp. Define those in your build file so you can use them everywhere. If a Java process needs the values, pass them to the process as a system property (-D...).

Erlang: using include from the console?

The include directive is usually used for a .hrl file at the top of an .erl file.
But, I would like to use include from the Erlang console directly.
I am trying to use some functions in a module. I have compiled the erl file from the console. But, the functions I want to use do not work without access to the hrl file.
Any suggestions?
"But, the functions I want to use do not work without access to the hrl file."
This can't be true, but from this I'll take a shot at guessing that you want access to records in the hrl file that you don't (normally) have in the shell.
If you do rr(MODULE) you will load all records defined in MODULE(including those defined in an include file included by MODULE).
Then you can do everything you need to from the shell.
(Another thing you may possibly want for testing is to add the line -compile(export_all) to your erl file. Ugly, but good sometimes for testing.)
Have you tried the compile:file option? You can pass a list of modules to be included thus:
compile:file("myfile.erl", [{i, "/path/1/"}, {i, "/path/2/"}])
It's worth nothing that jsonerl.hrl doesn't contain any functions. It contains macros. As far as I know, macros are a compile-time-only construct in Erlang.
The easiest way to make them available would be to create a .erl file yourself that actually declares functions that are implemented in terms of the macro. Maybe something like this:
-module(jsonerl_helpers).
-include("jsonerl.hrl").
record_to_struct_f(RecordName, Record) ->
?record_to_struct(RecordName, Record).
... which, after you compile, you could call as:
jsonerl_helpers:record_to_struct_f(RecordName, Record)
I don't know why the author chose to implement those as macros; it seems odd, but I'm sure he had his reasons.

Resources