> I will try to search the different behaviours
> of php functions with and without BOM.
Cool. I think our main problem with this is the PHP parser itself. One thing we _can_ do for that is come up with a good unixy way of detecting that a file has BOM so we can put it in cvswrappers on SF.net.
The idea being that we should be able to block commits from placing BOM-laden files in CVS. If our CVS tree is free from BOM-marked files, we won't have any problem.
Another way of dealing with it is to let BOM-marked files in, and then do some (horrible output buffering) magic when we include() or require() them to avoid the bogus early headers. Ugggh. Yuck.
General developer forum
UTF-8 and Byte Order Mark
This discussion has been locked so you can no longer reply to it.