Thread: --enable-locale and SET command ...
Morning ... Have a client that is asking us to enable localization support so that he can deal with turkish characters ... docs seem a bit sparse on this (or I'm not looking in the right place), but I'm guessing I want to use --enable-locale vs --enable-multibyte? *raised eyebrow* From the docs for v7.0, found at: http://www.postgresql.org/users-lounge/7.0/docs/postgres/x16981.htm It appears that I can't set this on a per-database? Its the whole server? I've looked at the SET command, but I think that might be where I'm missing something ... Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy Systems Administrator @ hub.org primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
Hmm, I guess they're trying to be misleading when they say "Up to 4GB of memory (PC100 ----> SDRAM <-----)"? Hmm, there isn't any mention of the nead for RDRAM... Shame on these companies for trying to screw people. - Mitch ----- Original Message ----- From: The Hermit Hacker <scrappy@hub.org> To: <pgsql-hackers@postgresql.org> Sent: Thursday, May 04, 2000 10:29 AM Subject: [HACKERS] --enable-locale and SET command ... > > Morning ... > > Have a client that is asking us to enable localization support so > that he can deal with turkish characters ... docs seem a bit sparse on > this (or I'm not looking in the right place), but I'm guessing I want to > use --enable-locale vs --enable-multibyte? *raised eyebrow* > > From the docs for v7.0, found at: > > http://www.postgresql.org/users-lounge/7.0/docs/postgres/x16981.htm > > It appears that I can't set this on a per-database? Its the whole > server? > > I've looked at the SET command, but I think that might be where > I'm missing something ... > > Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy > Systems Administrator @ hub.org > primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org >
At 10:47 AM 5/4/00 -0400, Mitch Vincent wrote: >Hmm, I guess they're trying to be misleading when they say "Up to 4GB of >memory (PC100 ----> SDRAM <-----)"? > >Hmm, there isn't any mention of the nead for RDRAM... Shame on these >companies for trying to screw people. No, they're not being misleading...most of the boards either come without the memory hub, in which case you need RDRAM, or with the memory hub (which is a separate chip that translates wide SDRAM data into narrow, serialized RDRAM data, which is why it's slower). The latter boards only take SDRAM. Some folks offer hybrid solutions. ASUS has one involving a board you plug into an RDRAM socket. The board has the memory hub chip on it, and you plug your SDRAM into that. Some offer i820 boards with two RDRAM and two SDRAM slots, and you choose which you want to use. That kind of thing. As far as 4GB SDRAM, the i840 offers two memory channels and two memory hubs, which means in principle twice as much SDRAM as with a traditional BX or i820 board, i.e. 4GB - if they provide enough slots. So it sounds like you're looking at a board with the i840 chipset, memory hub, and SDRAM DIMM slots. The memory hub wasn't part of Intel's original plan, it was added last year before rollout because RDRAM availability and price sucks. Mobo makers have jumped on the memory hub solution because they know buyers don't want to pay $800 for a 128MB RDRAM DIMM. Current RDRAM offers no real advantage to PC100/BX for most applications, but it has lots of upside potential. But RDRAM on an i840 board DOES offer an advantage because of the 10%-ish penalty associated with the memory hub/SDRAM solution. Ugh. - Don Baccus, Portland OR <dhogaza@pacifier.com> Nature photos, on-line guides, Pacific Northwest Rare Bird Alert Serviceand other goodies at http://donb.photo.net.
> Have a client that is asking us to enable localization support so > that he can deal with turkish characters ... What is the encoding for turkish? If it's sort of ISO 8859, you could INSERT/SELECT turkish characters with standard PostgreSQL. >docs seem a bit sparse on > this (or I'm not looking in the right place), but I'm guessing I want to > use --enable-locale vs --enable-multibyte? *raised eyebrow* If you want to sort the result correctly, go with --enable-locale (of course if the turkish LOCALE database in your system is correct). On the other hand, if you want to use Unicode, --enable-locale will not help. If the encoding for Unicode is UTF-8, probably you could live with --enable-multibyte. This one will give correct results for LIKE and regex search, but for the sort order, I'm not sure. Anyway, the decision really depends on the encoding. Can you tell me more details? > It appears that I can't set this on a per-database? Its the whole > server? They are for the compile time options. So they are for the whole server, of courese. No? -- Tatsuo Ishii
Tatsuo Ishii wrote: > > > Have a client that is asking us to enable localization support so > > that he can deal with turkish characters ... Turkish is a roman alphabet with some extra "European characters" like a C-cedilla. Probably is done in 7 or 8 bits, but I don't know particulars... - Thomas -- Thomas Lockhart lockhart@alumni.caltech.edu South Pasadena, California
> Turkish is a roman alphabet with some extra "European characters" like > a C-cedilla. Probably is done in 7 or 8 bits, but I don't know > particulars... After reading a book regarding charsets, it seems that ISO 8859-3 or 8859-9 is for Turkish (two standards being coexisting is probably coming from some historical reasons). Marc, can you confirm which one is used for Turkish in your case? -- Tatsuo Ishii