#2003: piraha_everywhere Zelmani stress test ---------------------+------------------------------------------------------ Reporter: rhaas | Owner: sbrandt Type: defect | Status: assigned Priority: major | Milestone: Component: Cactus | Version: development version Resolution: | Keywords: piraha ---------------------+------------------------------------------------------
Comment (by sbrandt):
What would it take to change this behavior to the one you request? Piraha is not currently a two pass parser. Every token has its own pattern, e.g. name=[a-zA-Z0-9_]+ or something equivalent. Since it would be impractical to change every token to allow \\n\s* in the middle, the only way to support this feature would be to implement a pass prior to the Piraha parsing which would do nothing except remove the \\n\s* patterns--except it wouldn't be quite that simple because simply substituting would throw the line numbers off. Perhaps it could insert a number of blank lines after the next \n to compensate. Maybe it wouldn't be that hard to do what I just described or that costly to the CPU, but... well, when I see that token broken in the middle with a \ for no reason it makes me cringe. I feel like I'm Detective Monk and you asked me to eat a sandwich that just fell on the floor. :)
Again, I wrote the text and didn't imagine it would be interpreted that way. I don't see why the fix can't be made to the documentation rather than the code.