- cross-posted to:
- programmerhumor@lemmy.ml
- cross-posted to:
- programmerhumor@lemmy.ml
You must log in or register to comment.
Maybe applies more to regex, the write only language.
The copy paste language. AI writes better regex than I do
and you won’t get better if you use ai for it
Meh I rarely use it. Even if I don’t use AI I wouldn’t get better at it, since I will forget everything the next time I will use it.
Back when I did a lot of Perl, those were okay-ish to parse. Nowadays, not so much. I guess it’s like Bash. If you write a lot of it (maybe some people do), it’s probably simple. If it’s only once every six months or less, eeehhh…
It all boils down to familiarity, which comes from repetitiveness.
So true. Every time I have to look up how to write a bash for loop. Where does the semicolon go? Where is the newline? Is it terminated with
done? Or withend? The worst part with bash is that when you do it wrong, most of the time there is no error but something completely wrong happens.I can only remember this because I initially didn’t learn about
xargs— so any time I need to loop over something I tend to usefor var in $(cmd)instead ofcmd | xargs. It’s more verbose but somewhat more flexible IMHO.So I run loops a lot on the command line, not just in shell scripts.
It all makes sense when you think about the way it will be parsed. I prefer to use newlines instead of semicolons to show the blocks more clearly.
for file in *.txt do cat "$file" doneThe
doanddoneserve as the loop block delimiters. Such as{and}in many other languages. The shell parser couldn’t know where stuff starts/ends.Edit: I agree that the
then/fi,do/donecase/esacare very inconsistent.Also to fail early and raise errors on uninitialized variables, I recommend to add this to the beginning of your bash scripts:
set -euo pipefailOr only this for regular sh scripts:
set -eu-e: Exit on error-u: Error on access to undefined variable-o pipefail: Abort pipeline early if any part of it fails.There is also
-xthat can be very useful for debugging as it shows a trace of every command and result as it is executed.set -euo pipefailFun fact, if you’re forced to write against POSIX shell, you aren’t allowed to use these options, since they’re not a thing, which is (part of) the reason why for example Google doesn’t allow any shell language but bash, lol.
Btw, all three set options given above are included in POSIX since 2024: https://pubs.opengroup.org/onlinepubs/9799919799/
Ooh, you’re totally right!! I forgot about that since it’s not in the older versions.
Me with powershell. I’ll write a pretty complex script, not write powershell for 3 months, come back and have to completely relearn it.
And I thought I was the only one… for smaller bash scripts chatGPT/Deepseek does a good enough job at it. Though I still haven’t tried VScode’s copilot on bash scripts. I have only tried it wirh C code and it kiiiinda did an ass job at helping…
AI does decently enough on scripting languages if you spell it out enough for it lol, but IMO it tends to not do so well when it comes to compiled languages
I’ve tried Python with VScode Copilot (Claude) and it did pretty good
That’s because scripted languages are more forgiving in general.
I was chalking it up to some scripting languages just tending to be more popular (like python) and thus having more training data for them to draw from
But that’s a good point too lol
Both can be true, Python does have a lot of examples floating online.
every control structure should end in the backwards spelling of how they started
Once you get used to it it is kind of fun.
Shame about
dothough.it could have been
notsince there’s notry.
PSA: Run ShellCheck on your shell scripts. It turns up a shocking number of programming errors. https://www.shellcheck.net/
Thank you for this. About a year ago I came across ShellCheck thanks to a comment just like this on Reddit. I also happened to be getting towards the end of a project which included hundreds of lines of shell scripts across dozens of files.
It turns out that despite my workplace having done quite a bit of shell scripting for previous projects, no one had heard about Shell Check. We had been using similar analysis tools for other languages but nothing for shell scripts. As you say, it turned up a huge number of errors, including some pretty spicy ones when we first started using it. It was genuinely surprising to see how many unique and terrible ways the scripts could have failed.
I wish it had a more comprehensive auto correct feature. I maintain a huge bash repository and have tried to use it, and it common makes mistakes. None of us maintainers have time to rewrite the scripts to match standards.
I honestly think autocorrecting your scripts would do more harm than good. ShellCheck tells you about potential issues, but It’s up to you to determine the correct behavior.
For example, how could it know whether
cat $fooshould becat "$foo", or whether the script actually relies on word splitting? It’s possible thatintentionally contains multiple paths.Maybe there are autofixable errors I’m not thinking of.
FYI, it’s possible to gradually adopt ShellCheck by setting
--severity=errorand working your way down to warnings and so on. Alternatively, you can add one-offcomments before offending lines to silence warnings.For example, how could it know whether
cat $fooshould becat "$foo", or whether the script actually relies on word splitting? It’s possible thatintentionally contains multiple paths.Last time I used ShellCheck (yesterday funnily enough) I had written
ports+=($(get_elixir_ports))to split the input sinceget_elixir_portsreturns a string of space separated ports. It worked exactly as intended, but ShellCheck still recommended to make the splitting explicit rather than implicit.The ShellCheck docs recommended
IFS=" " read -r -a elixir_ports <<< "(get_elixir_ports)" ports+=("${elixir_ports[@]}")
Then you’ll have to find the time later when this leads to bugs. If you write against bash while declaring it POSIX shell, but then a random system’s
shdoesn’t implement a certain thing, you’ll be SOL. Or what exactly do you mean by “match standards”?
Wait im not the only one? I think i relearned bash more times than i can remember.
That’s why I use nushell. Very convinient for writing scripts that you can understand. Obviously, it cannot beat Python in terms of prototyping, but at least I don’t have to relearn it everytime.
We have someone at work who uses it and he’s constantly having tooling issues due to compatibility problems, so… yeah.
I’m sure it’s fine for sticking in the shebang and writing your own one-off personal scripts, but I would never actually main it. Too much ecosystem relies on bash/posix stuff.
So the alternative is:
- either an obtuse script that works everywhere, or
- a legible script that only works on your machine…
Ruby and calling bash like this
`cat a.txt`a script that only works on your machine
That’s why docker exists :D
I am of the opinion that production software shouldn’t be written in shell languages. If it’s something which needs to be redistributed, I would write it in python or something
For a bit of glue, a shell script is fine. A start script, some small utility gadget…
With python, you’re not even sure that the right version is installed unless you ship it with the script.
I try to write things to be cross-platform; with node builds, I avoid anything using shell scripting so that we can support Windows builds as well. As such, I usually write the deployment scripts in Node itself, but sometimes python if it’s supported by our particular CI/CD pipeline
I keep forgetting windows exists.
Most common development platform in the world
I quit using it in the WfW days and never looked back.
I tend to write anything for distribution in Rust or something that compiles to a standalone binary. Python does not an easily redistributable application make lol
Yeah but then you either need to compile and redistribute binaries for several platforms, or make sure that each target user has rust/cargo installed. Plus some devs don’t trust compiled binaries in something like an npm package
Nu is great. Using it since many years. Clearly superior shell. Only problem is, that it constantly faces breaking changes and you therefore need to frequently update your modules.
They’ve slowed down with those a bit recently, haven’t they?
Not really. They’ve been on the stabilising path for about two years now, removing stuff like dataframes from the default feature set to be able to focus on stabilising the whole core language, but 1.0 isn’t out yet and the minor version just went three digits.
And it’s good that way. The POSIX CLI is a clusterfuck because it got standardised before it got stabilised.
dd’s syntax is just the peak of the iceberg, there, you gotta take out the nail scissors and manicure the whole lawn before promising that things won’t change.Even in its current state it’s probably less work for many scripts, though. That is, updating things, especially if you version-lock (hello, nixos) will be less of a headache than writing
shcould ever be. nushell is a really nice language, occasionally a bit verbose but never in the boilerplate for boilerplate’s sake way, but in the “In two weeks I’ll be glad it’s not perl” way. Things like command line parsing are ludicrously convenient (though please nushell people land collecting repeated arguments into lists).Yesterday, I upgraded from
0.101.0to0.102.0anddate to-tablewas replaced equally (actually better) withinto record, however it was not documented well in the error. Had to research for 5 to 10 minutes, which does not sound much, but if you get this like every second version, the amount of time adds up quickly.Actually had been deprecated beforehand, you should have gotten a warning. The deprecation cycle certainly is quite short, I’m still on 0.100.0, If I were to upgrade now I’d jump the version with the warning.
Not a problem for me in Nix, seems like a skill issue /j
Knowing that there is still a bash script i wrote around 5 years ago still running the entirety of my high scool lab makes me sorry for the poor bastard that will need to fix those hieroglyphs as soon as some package breaks the script. I hate that i used bash, but it was the easiest option at the time on that desolate server.
Bash scripts survive because often times they are the easiest option on an abandoned server
There’s always the old piece of wisdom from the Unix jungle: “If you write a complex shellscript, sooner or later you’ll wish you wrote it in a real programming language.”
I wrote a huge PowerShell script over the past few years. I was like “Ooh, guess this is a resume item if anyone asks me if I know PowerShell.” …around the beginning of the year I rewrote the bloody thing in Python and I have zero regrets. It’s no longer a Big Mush of Stuff That Does a Thing. It’s got object orientation now. Design patterns. Things in independent units. Shit like that.
I consider python a scripting language too.
I use it for scripting too. I don’t need Python as much as before nowaday.
They’re all programming languages, they all have their places.
All scripting languages are programming languages but not all programming languages are scripting languages
I initially read “UNIX jungle” as “UNIX jingle” and thought I had been really missing out!
You have, look up the SuSE songs.
Regex
This is one of the best uses for LLM’s imo. They do all my regex for me.
No. Learn it properly once and you’re good. Also it’s super handy in vim.
interns gonna intern
twitch
You always forget regex syntax?
I’ve always found it simple to understand and remember. Even over many years and decades, I’ve never had issues reading or writing simple regex syntax (excluding the flags and shorthands) even after long regex breaks.
It’s not about the syntax itself, it’s about which syntax to use. There are different ones and remembering which one is for which language is tough.
This is exactly it. Regex is super simple. The difficulty is maintaining a mental mapping between language/util <-> regex engine <-> engine syntax & character class names. It gets worse when utils also conditionally enable extended syntaxes with flags or options.
The hardest part is remembering whether you need to use
\wor[:alnum:].Way too few utils actually mention which syntax they use too. Most just say something accepts a “regular expression”, which is totally ambiguous.
I give you that, true. I wish vim had PCRE
There is the “very magic” mode for vim regexes. It’s not the exact PCRE syntax, but it’s pretty close. You only need to add \v before the expression to use it. There is no permanent mode / option though. (I think you can remap the commands, like / to /\v)
Don’t let the gatekeepers keep you out. This site helps.
Chatgpt helps even more
I know that LLMs are probably very helpful for people who are just getting started, but you will never understand it if you can’t grasp the fundamentals. Don’t let “AI” make you lazy. If you do use LLMs make sure you understand the output it’s giving you enough to replicate it yourself.
This may not be applicable to you specifically, but I think this is nice info to have here for others.
Most of regex is pretty basic and easy to learn, it’s the look ahead and look behind that are the killers imo
(?=)for positive lookahead and(?!)for negative lookahead. Stick a<in the middle for lookbehind.
You get used to it, I don’t even see the code—I just see: group… pattern… read-ahead…
I feel same with regex…
Bash substitution is regex-level wizardry.
Slapping a $ before an environment variable name is “wizardry?”
interns be interning
Not quite that, but more the
${variable##.*}sort of thing.Back then, a pain in the ass. Nowadays, I just let an AI handle that. I used this crap for years and years and still cannot remember, which symbols you need in which order. And why should I remember? I’m not the computer. The computer should know, not me.
Right, so that’s just the string manipulation functions. I already posted a link to the bible for this following a different reply to the same comment to which you replied.
Nope, the whole
${variable/regex/replacement}syntaxAll the string manipulation functions are easy: https://tldp.org/LDP/abs/html/string-manipulation.html
This one is my bookmark
Every time I have to do this I always go here. I can never remember the prefix suffix parts next time I do parameter substitution.
I have a confession to make: Unless shell script is absolutely required, I just use Python for all my automation needs.
My google history hits for powershell for loop is is in the dozens.
