____ _ _ | _ \ _ _ _ __ | |_(_)_ __ ___ ___ | |_) | | | | '_ \| __| | '_ ` _ \ / _ \ | _ <| |_| | | | | |_| | | | | | | __/ |_| \_\\__,_|_| |_|\__|_|_| |_| |_|\___| ____ _ | _ \ ___| |__ _ _ __ _ __ _ ___ _ __ | | | |/ _ \ '_ \| | | |/ _` |/ _` |/ _ \ '__| | |_| | __/ |_) | |_| | (_| | (_| | __/ | |____/ \___|_.__/ \__,_|\__, |\__, |\___|_| |___/ |___/
Runtime::Debugger - Easy to use REPL with existing lexical support and DWIM tab completion.
(emphasis on "existing" since I have not yet found this support in other modules).
In a script:
use Runtime::Debugger; repl;
On the commandline:
perl -MRuntime::Debugger -E 'repl'
Same, but with some variables to play with:
perl -MRuntime::Debugger -E 'my $str1 = "Func"; our $str2 = "Func2"; my @arr1 = "arr-1"; our @arr2 = "arr-2"; my %hash1 = qw(hash 1); our %hash2 = qw(hash 2); my $coderef = sub { "code-ref: @_" }; {package My; sub Func{"My-Func"} sub Func2{"My-Func2"}} my $obj = bless {}, "My"; repl; say $@'
Test command:
RUNTIME_DEBUGGER_DEBUG=2 perl -Ilib/ -MRuntime::Debugger -E 'my @a = 1..2; my %h = qw( a 11 b 22 ); my $v = 222; my $o = bless {a => 11}, "A"; my $ar = \@a; my $hr = \%h; use warnings FATAL => "all"; eval{ say qr<$hr-\>{b}> }; say "222"'
"What? Another debugger? What about ... ?"
The standard perl debugger (perl5db.pl) is a powerful tool.
perl5db.pl
Using per5db.pl, one would normally be able to do this:
per5db.pl
# Insert a breakpoint in your code: $DB::single = 1; # Then run the perl debugger to navigate there quickly: PERLDBOPT='Nonstop' perl -d my_script
If that works for you, then dont' bother with this module! (joke. still try it.)
This is a great and extendable module!
Unfortunately, I did not find a way to get the lexical variables in a scope. (maybe I missed a plugin?!)
Example:
perl -MDevel::REPL -E ' my $my_var = 111; # I want to access this our $our_var = 222; # and this. my $repl = Devel::REPL->new; $repl->load_plugin($_) for qw( History LexEnv DDS Colors Completion CompletionDriver::INC CompletionDriver::LexEnv CompletionDriver::Keywords CompletionDriver::Methods ); $repl->run; '
Sample Output:
$ print $my_var Compile error: Global symbol "$my_var" requires explicit package name ... $ print $our_var Compile error: Global symbol "$our_var" requires explicit package name ...
This module also looked nice, but same issue.
perl -MReply -E ' my $var=111; Reply->new->run; '
> print $var 1 > my $var2 = 222 222 > print $var2 1
While debugging some long-running, perl, Selenium test files, I basically got bored during the long waits, and created a simple Read Evaluate Print Loop (REPL) to avoid the annoyong waits between test tries.
Originally I would have a hot key command to drop in a snippet of code like this into my test code to essentially insert a breakpoint/pause.
One can then examine what's going on in that area of code.
Originally the repl code snippet was something as simple as this:
while(1){ my $in = <STDIN>; chomp $in; last if $in eq 'q'; eval $in; }
With that small snippet I could pause in a long running test (which I didn't write) and try out commands to help me to understand what needs to be updated in the test (like a ->click into a field before text could be entered).
And I was quite satisfied.
From there, this module increased in features such as using Term::ReadLine for readline support,tab completion, and history (up arrow).
Term::ReadLine
This module has changed in its approach quite a few times since it turns out to be quite tricky to perform eval_in_scope.
eval_in_scope
To make usage of this module as simple as possible, I tried my hand at source filters.
My idea was that by simply adding this line of code:
use Runtime::Debugger;
That would use a source filter to add in the REPL code.
This solution was great, but source filters can only be applied at COMPILE TIME (That was new to me as well).
Unfortunately, the tests I am dealing with are read as a string then evaled.
So, source filters, despite how clean they would make my solution, would not work for my use cases.
Next idea.
Then I decided to go back to using a command like:
use Runtime::Debugger; eval run;
Where run would basically generates the REPL code and eval would use the current scope to apply the code.
Side note: other Debuggers I had tried before this one, do not update lexical variables in the current scope. So this, I think, is unique in this debugger.
I learned later that eval run would under certain circumstances not work:
eval run
First call would print 111, while the exact same eval line would print undef afterwards.
sub { my $v = 111; eval q( # my $v = $v; # Kind of a fix. eval 'say $v'; # 111 eval 'say $v'; # undef ); }->();
Using eval run is still possible (for now).
Just be aware that it does not evaluate correctly under certain circumstances.
Simply add these lines:
This will basically insert a read, evaluate, print loop (REPL).
This should work for more cases (just try not to use nasty perl magic).
To reach the current solution, it was essential to go back to the main goal.
And the goal/idea is simple, be able to evaluate an expression in a specific scope/context.
Basically looking for something like:
peek_my(SCOPE)
But instead for eval:
eval_in_scope(SCOPE)
Given eval_in_scope(1), that would evaluate an expression, but in a scope/context one level higher.
eval_in_scope(1)
In order to eval a string of perl code correctly, we need to figure out at which level the variable is located.
Thats not hard to do: just look through increasing caller() levels until finding the first whose package name is not thia module's.
caller()
Given the scope level, peek_my/our is utilized to grab all the variables in that scope.
Having these variables:
my $var = 111; our $var = 222;
There can only be a single variable (glob) of a name. When multiple, the lexical one would be used.
Then we need to preprocess the piece of perl code that would be evaled.
At this stage variables would be replaced which their equivalent representation at found in peek_my/our.
This code:
say $var
Might be replaced with something like this:
say ${$PEEKS{'$var'}}
This transformation would normally be down seamlessly and hidden from the user.
Finally, eval the string.
And we pretend to have done eval_in_scope.
One idea would be to create an XS function which can perform an eval in a specific scope, but without the translation magic that is currently being done.
This might appear like peek_my, but for eval. So something like this:
eval_in_scope("STRING_TO_EVAL", SCOPE_LEVEL);
DEPRECATED! (Use repl instead)
repl
Runs the REPL.
Sets $@ to the exit reason like 'INT' (Control-C) or 'q' (Normal exit/quit).
$@
Note: This method is more stable than repl(), but at the same time has limits. See also
Works like eval, but without the lossy bug
Transform variables in a code string into references to the same variable as found with peek_my/our.
Try to insert the peek_my/our references (peeks) only when needed (should appear natural to the user).
Ok to transform:
say "@a"
NOT ok to transform:
say "%h"
This module has rich, DWIM tab completion support:
Press TAB when: - No input - view commands and variables. - After arrow ("->") - to auto append either a "{" or "[" or "(". (Depends on variable type) - After a hash) - show keys. - Otherwise - show variables.
Wrapper to simplify completion function.
Input:
words => ARRAYREF, # What to look for. partial => STRING, # Default: "" - What you typed so far. prepend => "STRING", # Default: "" - prepend to each possiblity. nospace => 0, # Default: "0" - will not append a space after a completion.
Returns the possible matches:
Show help section.
All commands run in the debugger are saved locally and loaded next time the module is loaded.
Can use hist to show a history of commands.
By default will show 20 commands:
hist
Same thing:
hist 20
Can show more:
hist 50
Data::Dumper::Dump anything.
You can use "d" as a print command which can show a simple or complex data structure.
d 123 d [1, 2, 3]
Data::Printer::p
You can use "p" as a print command which can show a simple or complex data structure with colors.
Some example uses:
p 123 p [1, 2, 3] p $scalar p \@array p \%hash p $object
Returns a unique list of elements.
List::Util in lower than v5.26 does not provide a unique function.
Internal use.
Install required library:
sudo apt install libreadline-dev
Enable this environmental variable to show debugging information:
RUNTIME_DEBUGGER_DEBUG=1
Why not perl debugger?
Why not Devel::REPL?
Why not Reply?
Tim Potapov, <tim.potapov[AT]gmail.com> 🐪🥷
<tim.potapov[AT]gmail.com>
Doing a Control-C may occassionally break the output in your terminal (exit with 'q' when possible).
Simply run any one of these:
reset tset stty echo
Currently it is not possible to create new lexicals (my) variables.
You can create new global variables by:
- Default $var=123 - Using our $our $var=123 - Given the full path $My::var = 123
inside a long running (and perhaps complicated) script, a variable may become undef.
This piece of code demonstrates the problem with using c<eval run>.
sub Func { my ($code) = @_; $code->(); } Func( sub{ my $v2 = 222; # This causes issues. use Runtime::Debugger; eval run; # Whereas, this one works. use Runtime::Debugger; repl; });
This issue is described here https://www.perlmonks.org/?node_id=11158351
Please report any (other) bugs or feature requests to https://github.com/poti1/runtime-debugger/issues.
You can find documentation for this module with the perldoc command.
perldoc Runtime::Debugger
You can also look for information at:
https://metacpan.org/pod/Runtime::Debugger
https://github.com/poti1/runtime-debugger
This software is Copyright (c) 2022 by Tim Potapov.
This is free software, licensed under:
The Artistic License 2.0 (GPL Compatible)
To install Runtime::Debugger, copy and paste the appropriate command in to your terminal.
cpanm
cpanm Runtime::Debugger
CPAN shell
perl -MCPAN -e shell install Runtime::Debugger
For more information on module installation, please visit the detailed CPAN module installation guide.