On 6/5/26 8:05 AM, Paul Richard Thomas wrote:
Hello All,
This is the penultimate submodule fix. It is actually straightforward.
However, the version produced by Claude produced error cascades and
modified the existing testcase to check for the error. The latter is
unacceptable because it hides the result of the original fix in all
cases. Instead, I have corrected the error and added the module
prefix, where appropriate.
The attached says the rest. It passes regression testing on FC44/x86_64.
OK for mainline and eventual backporting to gcc-16?
Cheers
Paul
Yes one must be careful of side effects and cascading error. Using the LLM to
assist can be very helpful, but one must monitor. As I use this, I learn to say
things like "remember not to ..." . "Remember" is a key word as it will do so.
(as always, be careful what you say, ;) )
You patch here looks good to go.
Cheers,
Jerry
