Sometimes things are missing from my irb history. This has been happening for years and I haven't really noticed a pattern. I'll use irb for a while, control-d out of it, come back in a few minutes later, and the commands I was just using aren't there, but commands from earlier in the day or week are there. This is mostly/always in a rails console.
Here's my .irbrc
I could be wrong, but reading over http://trac.greenstone.org/browser/extensions/gsdl-video/trunk/installed/cmdline/lib/ruby/1.8/irb/ext/save-history.rb?rev=18425
makes me think that the last irb process will overwrite the other processes history with its own.
So if you run:
irb
irb #in another term window
"a string" #then Ctrl-D in the first
#Ctrl-D in the second
then your .irb-history should be unchanged, as the initial, backgrounded irb will have loaded the old history, then overwritten the new history as it exits.
ie: Somewhat foolishly: open('w') rather than open('a') is the culprit.
You might consider writing a patch and commiting it for the greater good if this strongly irritates you.
Related
If I'm at a Bash prompt and I type bash I can then detect that I'm at the 2nd level by typing echo $SHLVL.
Suppose I'm in a pry session and I type pry again. How can I detect that I'm at the 2nd level down? How can I detect the Pry level? Nothing listed in (Pry.methods - Object.methods).sort seems to have anything useful.
This is for testing some code for a pry-related project, where I need detect the level.
If you call caller within a pry session within a pry session, then you will see a list of commands. Among them, you should be able to find the part that corresponds to a nested pry call. Find the crucial line that is related to each pry session call, and you will know your level. As far as I checked, you should find two occurrences of a line like:
"/usr/local/lib/ruby/gems/2.2.0/gems/pry-0.10.1/lib/pry/repl.rb:67:in `repl'"
Count such lines.
I have a ruby script, apparently correct, that sometimes stops working (probably on some calls to Postgresql through the pg gem). The problem is that it freezes but doesn't produce any error, so I can't see the line number and I always have to isolate the line by using puts "ok1", puts "ok2", etc. and see where the script stops.
Is there any better way to see the current line being executed (without changing the script)? And maybe the current stack?
You could investigate ruby-debug a project that has been rewritten several times for several different versions of ruby, should allow you to step through your code line by line. I personally prefer printf debugging in a lot of cases though. Also, if I had to take an absolutely random guess at your problem, I might investigate whether or not you're running into a race condition and/or deadlock in your DB.
I notice that the commands I have in my shell scripts never get added to the history list. Understandably, most people wouldn't want it to be, but for someone who does, is there a way to do it?
Thanks.
Edit:
Sorry about the extremely late reply.
I have a script that conveniently combines some statements that ultimately result in lynx opening a document. That document is in a dir several directories below the current one.
Now, I usually end up closing lynx to open another document in the current dir and need to keep switching back and forth between the two. I could do this by having another window open, but since I'm mostly on telnet, and the switches aren't too frequent, I don't want to do that.
So, in order to get back to lynx from the other document, I end up having to re-type the lynx command, with the (long) path/filename. In this case, of course, lynx isn't stored in the command history.
This is what I want added to the history, so that I can get back to it easily.
Call it laziness, but hey, if it teaches me a new command....
Cheers.
As #tripleee pointed out, which problem are you actually trying to solve? It's perfectly possible to include any shell code in the history, but above some level of complexity it's much better to keep them in separate shell scripts.
If you want to keep multi-line commands in the history as they are, you might want to try out shopt -s lithist, but that means searching through history will only return one line at a time.
I'm developing a ruby app with Vim in one screen, and in another I'm using irb to test the code I'm writing in Vim.
For the sake of this example, we'll say there are only 2 irb commands to test the code that I've written, a load and a method call.
Generally, it's more than 2 commands - if its just 2 I would tap the up arrow twice and hit enter, but generally its around 5 or 6 and that's up to 36 key presses just to run the last 6 commands. There must be an easier way.
Is there an easy way to execute the last X commands from the irb readline history buffer in the order they were executed, where X is the number of commands you want to run? Something like;
run_last_x_commands(i)
I'm pretty sure it would be a custom written piece of code, I'm just wondering if anyone already solved this problem or if this is something I'll need to write myself.
You may be interested in running Vim in IRB.
Gem.
Source.
I would really store your test script in a file. It is less convenient than just typing it out, but at the same time, it's easier to edit and work with. Plus, if you're trying to define classes, you often get issues with IRB not able to redefine a class without doing things like Object.remove_const('MyClass'). I've had the same problem you're having, and I decided against using IRB for this task.
There may be some readline magic to do what you want though. Readline has all kinds of tricks and shortcuts I don't know about.
I would like to talk to a interactive bash process.
Here is an example, so you know what I want to archieve:
Program starts a new bash process.
User types "ls" into my program.
Program sends this command to the bash process.
Program reads all available output of the bash (including the prompt) and displays it back to the user.
GOTO 1
As you can guess, there is much room for nifty manipulations here and there... ;-)
It would be wonderful if this also worked for subprocesses (started by the bash process) and curses-based programs.
I would like to implement this functionality in Ruby, and already have experimented with IO.popen, but strange things happen. You are also welcome to do this in other languages.
Ok, I've found a solution. This work pretty nicely, you can even start vim inside it :-)
require "pty"
system("stty raw -echo")
PTY.spawn("bash -i") do |pin, pout|
Thread.new do
loop do
pout.print STDIN.getc.chr
end
end
loop do
print pin.sysread(512)
STDOUT.flush
end
end
This does the following:
enable character-wise input (limited to UNIXoids, I'm afraid)
create a pseudo-TTY, start a interactive bash session inside
forward each character from STDIN to the bash
print every output back to the user
Have you tried using the Session gem?
http://rubygems.org/gems/session
https://github.com/ahoward/session (Homepage with introduction.)
I don't have any experience with it, but the README seems to describe what you want. It's description says, "session kicks the ass", so it should be fun/productive to play with it in any case.