Next Taifun

With Higos there's the next Taifun coming soon. Now that we know that hoarding will be a problem at our local store we should buy some stuff soon in case the canteens shut down.


Today brought a big disappointment. When going to Xiawucha (read: Afternoon Tea, 下午茶) at 5 today French Toasts (read: sāidōsih, 西多士) where out. What a pity (倒霉). They still had noodles with sausages (but no ham, 火腿) and a chicken wing though.

Cantonese finals wanted (follow-up to 'Infrequent syllables of Cantonese')

Recently I was looking into Infrequent syllables of Cantonese that I found in the table provided by the Research Centre for Humanities Computing of the Chinese University of Hong Kong (CUHK) and the Unihan table provided by Unicode (see the combined table in Jyutping syllable table).

For some of this syllables I now found some more sources, that give information about syllable initials and finals. So instead of going through syllables I'll provide the information for syllable finals. The source used is:

Robert S. Bauer, Paul K. Benedikt: Modern Cantonese Phonology (摩登廣州話語音學). Walter de Gruyter, 1997, ISBN 3-11-014893-5. Google has a scanned version (preview).

  • eu according to Bauer & Benedikt is pronounced /ɛw/,
  • em is pronounced /ɛːm/,
  • en is pronounced /ɛːn/, this syllable is not found in Unihan, the CUHK table includes it but gives no example,
  • ep is pronounced /ɛːp/,
  • et is pronounced /ɛːt/,

For all those finals no mapping is given to Cantonese Yale by this book. Furthermore the book 'Speak Cantonese' by Huang (Parker Po-fei Huang, Gerard P. Kok: Speak Cantonese (Book I). Revised Edition, Yale University, 1999, ISBN 0-88710-094-5) doesn't give any finals that could be mapped to those Jyutping finals.

There are still finals found partially in the CUHK and Unihan tables:

  • um, found in the CUHK table without example, not in Unihan,
  • up, found in the CUHK table without example, not in Unihan, too,
  • oet, found in the CUHK table without example, though Unihan has character 㖀 for loet,
  • oei, only found in Unihan with characters 唳, 捩 for loei,
  • om, only found in Unihan with character 媕 for om,

These finals are not listed in Bauer & Benedikt so no IPA can be given.

Update: Cantonese: A Comprehensive Grammar (Stephen Matthews, Virginia Yip, Routledge, 1994, ISBN 0-415-08945-X) gives finals -em, -up, -et, -en, -um for Cantonese Yale (p. 20, chapter 1.3.1) and states that they are used in loan-words or as onomatopaeic sounds and gives examples lem, bup, wet, fen and bum.

Python doctest and Unicode

Today I found out about Python's doctest module. "The doctest module searches for pieces of text that look like interactive Python sessions, and then executes those sessions to verify that they work exactly as shown."

In the code I currently write I have a lot of docstrings I would like to constantly check, so that would help me a lot. Running it on my code though results in an UnicodeEncodeError:

  File "/usr/lib/python2.5/", line 1815, in testmod
  File "/usr/lib/python2.5/", line 1361, in run
    return self.__run(test, compileflags, out)
  File "/usr/lib/python2.5/", line 1215, in __run
    self.report_start(out, test, example)
  File "/usr/lib/python2.5/", line 1123, in report_start
    'Expecting:\n' + _indent(example.want))
UnicodeEncodeError: 'ascii' codec can't encode character u'\xe1' in position 64: ordinal not in range(128)

It seems other people have the same problem: [1], [2].

After Python and Exceptions containing Unicode messages that is already the second Unicode problem in Python's core that I come across. Personal feeling: I do think that nowadays it isn't appropriate anymore for a serious programming language to have those problems.

Upgrading Drupal

I decided to upgrade Drupal today mostly because of security fixes. First to 5.10, then to 6 as I was already at it. It turns out that Drupal 6 needs at least MySQL 4.1 but our RZ still runs 4.0 which is bugging me for more than two years. So I needed to revert back to 5.10. Suddenly my content was gone. Instead I would see "n/a" ("n/v" in the translated version). Google didn't give results that delt with my problem. After working on this for more than 2 hours I solved it: "Problem Exists Between Chair And Keyboard."

Doing a backup I do a full SQL dump. When restoring I can't use locking as our RZ wouldn't give us locking rights (which makes me patch Drupal for more than two years). I do a "grep -v 'LOCK TABLES' though I should've done a "egrep -v '^LOCK TABLES" as I deleted the hugh table of content: table node_revisions was completely missing.

Syndicate content