view tests/test-obsolete-bounds-checking.t @ 50400:95acba2c29f6

encoding: avoid quadratic time complexity when json-encoding non-UTF8 strings

Apparently the code uses "+=" with a bytes object, which is linear-time, so the
whole encoding is quadratic-time. This patch makes us use a bytearray object,
instead, which has a(n amortized-)constant-time append operation.

The encoding is still not particularly fast, but at least a 10MB file
takes tens of seconds, not many hours to encode.

Create a repo, set the username to something more than 255 bytes, then run hg amend on it.

  $ unset HGUSER
  $ cat >> $HGRCPATH << EOF
  > [ui]
  > username = aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa <[e-mail]>
  > [extensions]
  > amend =
  > [experimental]
  > evolution.createmarkers=True
  > evolution.exchange=True
  > EOF
  $ hg init tmpa
  $ cd tmpa
  $ echo a > a
  $ hg add
  adding a
  $ hg commit -m "Initial commit"
  $ echo a >> a
  $ hg amend 2>&1 | egrep -v '^(\*\*|  )'
  transaction abort!
  rollback completed
  Traceback (most recent call last):
  *ProgrammingError: obsstore metadata value cannot be longer than 255 bytes (value "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa <[e-mail]>" for key "user" is 285 bytes) (glob)