I've discussed it with Martin, the disagreement was two-fold:
- he felt the bug was too vague,
- I felt setting the importance to low wasn't the right way to
convey that the priority is low (which I agree with, I don't
intend to fix it until the migration to the new design is
completed, including addressing the performance issue)
I answered his concern about vagueness by explaining that I think
there are ~10 at worst 20 such constants and that I don't know
them all.
Here are the ones I can find after a quick grepping:
I've discussed it with Martin, the disagreement was two-fold:
- he felt the bug was too vague,
- I felt setting the importance to low wasn't the right way to
convey that the priority is low (which I agree with, I don't
intend to fix it until the migration to the new design is
completed, including addressing the performance issue)
I answered his concern about vagueness by explaining that I think
there are ~10 at worst 20 such constants and that I don't know
them all.
Here are the ones I can find after a quick grepping:
bzrlib. workingtree. worth_saving_ limit groupcompress. BATCH_SIZE groupcompress. GroupCompressVe rsionedFiles. _DEFAULT_ MAX_BYTES_ TO_INDEX lockdir. _DEFAULT_ TIMEOUT_ SECONDS lockdir. __DEFAULT_ POLL_SECONDS add._DEFAULT_ MAX_FILE_ SIZE
bzrlib.
bzrlib.
bzrlib.
bzrlib.
bzrlib.
I'm sure you know more, feel free to add them (we have time to
list them before this bug is fixed anyway).