5 ms·

Why would you be surprised? Out of nearly a decade of using Bash on Linux I don't think I have ever run into a bash bug or issue. There's been bugs with utiliti

Why would you be surprised? Out of nearly a decade of using Bash on Linux I don't think I have ever run into a bash bug or issue. There's been bugs with utilities like sudo of course, but those aren't Bash.

chasil 5y ago
From the anguish of those who try to implement the POSIX shell, I understand that edge cases are far from difficult to find. The Almquist shell (Debian Dash) is cited for not using lex/yacc to implement the parser, and described as untrustworthy. https://archive.fosdem.org/2018/schedule/event/code_parsing_posix_s_hell/ https://archive.fosdem.org/2018/schedule/event/code_parsing_... A major goal of this later ADA implementation is readability, in addition to (Windows) performance and ADA's safety. https://archive.fosdem.org/2019/schedule/event/ada_shell/ https://archive.fosdem.org/2019/schedule/event/ada_shell/
SavantIdiot 5y ago
"I don't think I have ever run into a bash bug or issue." That is not a statement about a formally verifiable program that I would base a billion-dollar program on.
- varjag 5y ago
  Ariane 5 bug was formally verified.
  - sigzero 5y ago
    No it really wasn't.
- count 5y ago
  Formally verified programs often still have bugs. AWS S2N, for example, currently has hundreds of open issues and identified bugs, despite being the subject of a continuous formal verification program. Bash powers probably hundreds of billions of dollars worth of critical infrastructure...
- pabs3 5y ago
  There are attempts at formal analysis of shell scripts, for eg: https://colis.irif.fr/ https://colis.irif.fr/ https://github.com/colis-anr/morbig https://github.com/colis-anr/morbig
- nexuist 5y ago
  My point being that whatever other software runs on this thing is far more likely to contain hidden bugs than bash has given its age and widespread distribution. It seems weird to single out bash as opposed to the hundreds of other random packages on any given distro.
bmoyles 5y ago
It happens... I can think of a couple that bit me in the last 5 years or so. Some of them felt like bugs but were flagged as new features in later releases like this one from 4.4... spent way too long trying to figure out why this worked under 4.4 but not 4.3 and didn't even consider a bug, let alone a feature at the time. I was convinced it was something I did wrong :P "a. Using ${a[@]} or ${a[*]} with an array without any assigned elements when the nounset option is enabled no longer throws an unbound variable error."
pabs3 5y ago
There are design flaws and edge cases in bash though, lots of them come up when trying to automatically show the exit code of the last command and when trying to automatically show whether or not the output of the last command ended in a line ending character.
qiqitori 5y ago
Not sure if this counts as "running into an issue", but this bash issue is pretty famous: https://en.wikipedia.org/wiki/Shellshock_(software_bug) https://en.wikipedia.org/wiki/Shellshock_(software_bug)

chasil 5y ago

From the anguish of those who try to implement the POSIX shell, I understand that edge cases are far from difficult to find. The Almquist shell (Debian Dash) is cited for not using lex/yacc to implement the parser, and described as untrustworthy. https://archive.fosdem.org/2018/schedule/event/code_parsing_posix_s_hell/ https://archive.fosdem.org/2018/schedule/event/code_parsing_... A major goal of this later ADA implementation is readability, in addition to (Windows) performance and ADA's safety. https://archive.fosdem.org/2019/schedule/event/ada_shell/ https://archive.fosdem.org/2019/schedule/event/ada_shell/

SavantIdiot 5y ago

"I don't think I have ever run into a bash bug or issue." That is not a statement about a formally verifiable program that I would base a billion-dollar program on.

varjag 5y ago
Ariane 5 bug was formally verified.
- sigzero 5y ago
  No it really wasn't.
count 5y ago
Formally verified programs often still have bugs. AWS S2N, for example, currently has hundreds of open issues and identified bugs, despite being the subject of a continuous formal verification program. Bash powers probably hundreds of billions of dollars worth of critical infrastructure...
pabs3 5y ago
There are attempts at formal analysis of shell scripts, for eg: https://colis.irif.fr/ https://colis.irif.fr/ https://github.com/colis-anr/morbig https://github.com/colis-anr/morbig
nexuist 5y ago
My point being that whatever other software runs on this thing is far more likely to contain hidden bugs than bash has given its age and widespread distribution. It seems weird to single out bash as opposed to the hundreds of other random packages on any given distro.

bmoyles 5y ago

It happens... I can think of a couple that bit me in the last 5 years or so. Some of them felt like bugs but were flagged as new features in later releases like this one from 4.4... spent way too long trying to figure out why this worked under 4.4 but not 4.3 and didn't even consider a bug, let alone a feature at the time. I was convinced it was something I did wrong :P "a. Using ${a[@]} or ${a[*]} with an array without any assigned elements when the nounset option is enabled no longer throws an unbound variable error."

pabs3 5y ago

There are design flaws and edge cases in bash though, lots of them come up when trying to automatically show the exit code of the last command and when trying to automatically show whether or not the output of the last command ended in a line ending character.

qiqitori 5y ago

Not sure if this counts as "running into an issue", but this bash issue is pretty famous: https://en.wikipedia.org/wiki/Shellshock_(software_bug) https://en.wikipedia.org/wiki/Shellshock_(software_bug)