On Fri, Jun 13, 2003 at 11:50:55PM -0500, Manoj Srivastava wrote: > Well, there were reasons, should you care. I held off the > transition for a long time. Unfortunately, the C++ scanners generated > by the old flex would not work with the gcc version in > unstable. I held off on the transition as long as I could, trying to > patch better support in old flex, but my scaffolding of fixes was > groaning under its own weight, and there were bug reports pointing > out the shiny new compiler support in upstream flex. > > The new Flex supported the compiler we were shipping; it had a > nice new test suite, and I incorporated passing the test suite as a > part of the debian build process; I looked at the tests, and they > seemd far more extensive than anything we ever had before (all we had > before was flex to be able to compile itself). The temptation of that > test suite was the last straw. > > I must confess to being startled by the magnitude of changes > without a bump in the version number. > > In any case, I see this division of flex-old and flex as the > best solution to the issue at hand. The new flex has exciting new > features, but breaks with POSIX, and is not backwards compatible. > It's scanners ought to work on most modern OS's. > > Think of this as part of the g++ transition. I for one wish you had said all of the above back in April instead of "The behaviour of flex has changed in the latest release, yes." and adopting the fierce "NOT ONE STEP BACKWARD" stance you have taken prior to this message. And I hope we all learn from this experience that one's upstream version numbers sometimes need to be taken with a truck or two of salt. -- G. Branden Robinson | It was a typical net.exercise -- a Debian GNU/Linux | screaming mob pounding on a greasy branden@debian.org | spot on the pavement, where used to http://people.debian.org/~branden/ | lie the carcass of a dead horse.
Attachment:
pgpserUcUoAXL.pgp
Description: PGP signature