Change 1552
Quick and kludgey fix to bug introduced by change 1540.
Fix probably
has to do with combination of using filelog output to fill in the
name of a FileHead, using tags to run filelogs asynchronously, and
preloading a bunch of FileHead pointers at once.
The kludge: store the name which was originally used to create the
FileHead and use that in comparisons.
The better fix (to be done later): call WaitTag() if needed to fill
in the FileHead and get the name of the file. The problem with the
current kludge is that it still won't work if the file's "oname"
isn't in depot syntax, which is what change 1540 was supposed to
address in the first place. Grahr.
Fix probably
has to do with combination of using filelog output to fill in the
name of a FileHead, using tags to run filelogs asynchronously, and
preloading a bunch of FileHead pointers at once.
The kludge: store the name which was originally used to create the
FileHead and use that in comparisons.
The better fix (to be done later): call WaitTag() if needed to fill
in the FileHead and get the name of the file. The problem with the
current kludge is that it still won't work if the file's "oname"
isn't in depot syntax, which is what change 1540 was supposed to
address in the first place. Grahr.