On 3/7/22 15:23, Mark Wielaard wrote: > Hi, > > On Mon, 2022-03-07 at 10:57 +0100, Arthur Cohen wrote: >> This is due to the lexer's fix not being in yet. I think it should >> be >> resolved later as soon as bors merges the fix, but I'll check just in >> case. > > According to: > https://builder.wildebeest.org/buildbot/#/builders?tags=gccrust > > It was fixed by: > > commit 7a3c935c0f220835c001307944587a84c5af0192 > Author: Philip Herron > Date: Fri Mar 4 13:45:34 2022 +0000 > > Check if this constant item might already be compiled > > Which is slightly odd, because that isn't a lexer fix. > So I suspect the bug is still there, but is hidden again. I've checked the master branch, and the lexer is fixed there. I've checked out the commit you mentionned, and running git blame on it shows that the fix was introduced in 45eac5686867: commit 45eac568686745af73848f6c238fefcd87e315de (HEAD, origin/lexer-string-lifetime, lexer-string-lifetime) Author: Arthur Cohen Date: Wed Mar 2 15:18:40 2022 +0100 lexer: Add reference and warning documentation Fixes the -fself-test invalid memory accesses and adds documentation regarding a possible future fix. Co-authored-by: tschwinge Co-authored-by: philberty Which is the proper commit. I suspect this might be due to bors, our merging bot on github, and me asking it to merge multiple pull requests this morning (around 4 or 5). Not really sure. But the bug is fixed and properly placed in the history :) All the best, Arthur