Federico G. Schwindt
Federico G. Schwindt
I haven't looked at this in detail yet but it occurs to me we could move it to util/regexp-assemble/regexp-920120.data to help readability, at least for the non-assembled version (assuming the...
Having had another look, I realise the pattern presented is generated by a script. My bad, sorry. Perhaps we should include your script in a similar way we include `regexp-assemble.pl`...
Thanks for updating the PR. Have you seen this part of my comment? > As for the pattern, I believe you could extract the common suffix as well (^|). >...
> Regarding extracting ^ Right. I understand the nuances of ^. I was a bit hasty here and didn't analyse the pattern in detail before I suggested the move though....
For a simpler test limited to pcre, you can run `pcretest -TM ` using the different patterns and inputs.
It's too early to tell either way (whether this will be merged) but the more information we have, the more informed the decision will be. > Perf discussion continued This...
I stand corrected. As you said, there will be no difference in logging with this change.
I'd like to chime in. There is a lot to unpack on this single issue. In the future it might be easier to have multiple issues so they are easier...
I believe this fell through the cracks. Might be a good time to revisit.
Today's meeting outcome: Update commit message to explain why we reverting these changes (including why we double decode) before merging.