We think this may be an issue with 'hop apply all'
also. It would explain some of the oddities we've
seen crop up around here.
hop ketchup writes all the changes it's going to apply
to the database, then applies them. If for any reason
the process is aborted in the middle (ctrl-c), hop
lists all the changes as being applied, but in actual
fact they haven't been.
Logged In: YES
user_id=161477
The entire hop system works this way. It's bad. But it
means a major rewrite to fix it.
Logged In: YES
user_id=161477
I've changed "ketchup" in the version that AccuMap uses to
do a "smarter" ketchup. It also does away with
the "passtheketchup". You just run a "hop pass all" in the
relevent build tree, and it will pass "all APPLIED changes
in that tree, which HAVE NOT YET been passed". To ensure
that any particular change does not get passed, UNAPPLY it
before running the "pass all"
I'll implement the same "smarter" ketchup for
Solaris/Cygwin now
Logged In: YES
user_id=161477
Ok. The "smarter" ketchup has been implemented, tested
under cygwin, and checked in.
Logged In: YES
user_id=209227
In general this applies to all change apply unapply
activity.