On Mon, Feb 3, 2025, at 10:24 PM, Zeffie via Bug reports for the GNU Bourne Again SHell wrote: > Focus on $BASH_VERSION: > While you argue about testing for $BASH_VERSION, this point isn’t > central to the bug report. The issue is about terminal reinitialization > in sh‑mode (when Bash is invoked via its symlink as "sh") and preserving > the echoctl flag—not about whether the shell reports its version. Your > discussion on this aspect, therefore, doesn’t address the technical > problem at hand.
You're the one who brought up BASH_VERSION first by making the incorrect assertion that bash does not set it when run as "sh". No one actually claimed that it has anything to do with the core issue, yet you remain laser-focused on it. I expect Chet only printed it to show you what version of bash he was testing. > Common System Configuration: > Additionally, on many systems, /bin/sh is actually a symlink to dash, > not Bash. For example, on some distributions you might see: > > -rwxr-xr-x 1 root root 1235608 Apr 18 2022 /bin/bash > lrwxrwxrwx 1 root root 4 Nov 24 2019 /bin/sh -> dash > > This configuration underscores that tests or assertions based solely > on Bash-specific variables or behaviors are not generally applicable to > the broader ecosystem where /bin/sh may not even be Bash. You're absolutely right: sometimes /bin/sh is not bash, but another shell. You're so close to getting it. > Relevance to the Bug Report Scenario: > The bug report describes a scenario in which the terminal's echoctl > flag is overridden in sh‑mode, leading to missing visual feedback (e.g., > "^C" not appearing when CTRL‑C is pressed). Focusing on tests that rely > solely on $BASH_VERSION misses the point. The critical matter is how > terminal attributes are reset when Bash is run in sh‑mode. > > [...] > > In summary, while your testing and analogies may be well intended, they > do not directly address the core issue documented in the bug report. The > focus should remain on the behavior of Bash in sh‑mode—specifically, how > terminal settings like echoctl are handled—rather than on aspects that > are either tangential (such as version reporting). No, you are actually the one who is missing the point, and it's caused most if not all of the miscommunication in this thread. I thought I went a little overboard about this in my other message, but apparently not overboard enough. I'll try again. [Jack Torrance mode] The bug report is about FreeBSD ash, not bash. The bug report is about FreeBSD ash, not bash. The bug report is about FreeBSD ash, not bash. The bug report is about FreeBSD ash, not bash. The bug report is about FreeBSD ash, not bash. The bug report is about FreeBSD ash, not bash. The bug report is about FreeBSD ash, not bash. The bug report is about FreeBSD ash, not bash. The bug report is about FreeBSD ash, not bash. The bug report is about FreeBSD ash, not bash. HEEEEEERE'S JOHNNY -- vq
