On Sun, 19 Nov 2023, Jeff Law wrote: > > Verified with the `riscv64-linux-gnu' target and the C language > > testsuite. OK to apply? > Not sure why it is the way it is -- I walked back to Zdenek's change which > introduced the scan-assembler-times and nothing about the -inline argument.
I went through our history beforehand too and found nothing interesting either. My only suspicion has been it may have happened as a conseqence of somewhat confusing regexp(n) TCL documentation just saying: "Determines whether the regular expression exp matches part or all of string and returns 1 if it does, 0 if it does not, unless -inline is specified (see below)." and then you need to dive into the description of `-all' to find out it actually returns the number of matches rather than just 1 or 0: "Causes the regular expression to be matched as many times as possible in the string, returning the total number of matches found." I guess maybe Zdenek missed the part after the comma? > OK, but be on the lookout for scan-asm problems on other targets over the next > few days. Good point. I have grepped our testsuite for instances and found only one (as opposed to numerous non-captured subexpressions), specifically gcc/testsuite/gcc.target/arm/pr53447-5.c, well-documented as working around the quirk. I've posted a change to avoid the quirk with this case: <https://gcc.gnu.org/pipermail/gcc-patches/2023-November/637710.html> and I mean to apply it just before this `scan-assembler-times' fix. Maciej
