[ANN] RebDB v1.1
[1/7] from: atruter::labyrinth::net::au at: 20-Feb-2004 16:20
Latest release includes:
* memo! support
* Client / Server
* SQL dialect
* RebDB*Lite variant (uses save/load with REBOL native values)
* SQL*Plus like command-line (uncomment last line of db.r / db-lite.r)
Release notes and source available at http://www.dobeash.com/RebDB/
Regards,
Ashley
[2/7] from: petr:krenzelok:trz:cz at: 20-Feb-2004 9:10
Ashley Tr=FCter napsal(a):
>Latest release includes:
>* memo! support
<<quoted lines omitted: 3>>
>* SQL*Plus like command-line (uncomment last line of db.r / db-lite.r)
>Release notes and source available at http://www.dobeash.com/RebDB/
my friend looked for small DB, your solutions would be good for him, but
the app is supposed to run on old notebook with only 16MB RAM. Is there
really no way of how to get damned open/skip-or-whatever-would-help to
work? Currently, in rebol, being able to only work with in-memory
storage is starting to be pretty painfull, I meet ppl here or there who
don't want to install full mySQL server and would be fine with some kind
of small rebol db, if that would not be memory based.
-pekr-
[3/7] from: gchiu:compkarori at: 20-Feb-2004 9:36
Petr Krenzelok wrote.. apparently on 20-Feb-2004/9:10:58+1:00
>my friend looked for small DB, your solutions would be good for him, but
>the app is supposed to run on old notebook with only 16MB RAM. Is there
try dbase ...
--
Graham Chiu
http://www.compkarori.com/cerebrus
[4/7] from: petr:krenzelok:trz:cz at: 20-Feb-2004 10:33
Graham Chiu napsal(a):
>Petr Krenzelok wrote.. apparently on 20-Feb-2004/9:10:58+1:00
>
>>my friend looked for small DB, your solutions would be good for him, but
>>the app is supposed to run on old notebook with only 16MB RAM. Is there
>>
>>
>
>try dbase ...
>
do you mean Konstantin's dybase? It requires library component IIRC. But
that is not a much of a problem. I also remember Jeff producing link to
.... was it BerkeleyDB? The problem as I see it is - the set-up and
deployment cost. E.g. mySQL maybe overhead, but it accepts SQL syntax,
so switching to another DB may not influence your scripts so much, even
if there will surely be some work required.
I think that Rebol badly misses its small, native db format, which would
become standard for those, who don't want/can't use in-memory storage
and IOS-like one-record-per file storage is either not an option. I
think RebDB could be the answer, if open/skip would be fixed, but that
is another story. In the time being, I would stick with "standard" -
mySQL and would not support proprietary aproach/syntax, unless thinking
of such solution as being standard and long-term ...
PS: I have nothing against existance of various options, though ...
-pekr-
[5/7] from: gchiu:compkarori at: 20-Feb-2004 10:41
Petr Krenzelok wrote.. apparently on 20-Feb-2004/10:33:08+1:00
>>try dbase ...
>>
>>
>>
>do you mean Konstantin's dybase? It requires library component IIRC. But
no, I meant dbase, or any other odbc available connection
http://www.rebolfrance.net/articles/odbcsockserv/odbcsockserv.html
--
Graham Chiu
http://www.compkarori.com/cerebrus
[6/7] from: atruter:labyrinth:au at: 22-Feb-2004 11:19
Hi Pekr,
> my friend looked for small DB, your solutions would be good for him, but
> the app is supposed to run on old notebook with only 16MB RAM. Is there
> really no way of how to get damned open/skip-or-whatever-would-help to
> work?
We can certainly emulate:
port: open/binary/direct/skip ...
with
port: open/binary/direct ...
copy/part port ...
perhaps replacing large "copy/part" operations with a "copy/part" loop of
some sort, but I don't know how efficient this would be for large files in
a low memory environment. It is certainly something I am looking at
(including small-memory devices) along with the broader issue of, "How can
I reduce REBOL memory usage to the bare minimum?". Using REBOL/Base along
with aggressive 'unset usage is where I am starting from.
> Currently, in rebol, being able to only work with in-memory
> storage is starting to be pretty painfull, I meet ppl here or there who
> don't want to install full mySQL server and would be fine with some kind
> of small rebol db, if that would not be memory based.
My position in a nutshell! The other case where not having direct/skip is
a pain (and copy/part is not an acceptable workaround) is trying to read
JPEG EXIF thumbnails over a dial-up line ... you have to read about 55Kb
of the file just to get the 5Kb thumbnail (160x120) you need, ARGHHH!
Regards,
Ashley
[7/7] from: maarten:vrijheid at: 22-Feb-2004 7:45
We can!
>We can certainly emulate:
>
> port: open/binary/direct/skip ...
>
>with
>
> port: open/binary/direct ...
> copy/part port ...
>
skip-bytes: func
[
{Skips a number of bytes on an open port, to work around the
open/binary/direct skip bug. THIS SHOULDN'T BE NEEDED!!!}
p [port!] {The port to skip on}
i [integer!] {The number of bytes to skip}
/local ofz rem-off
]
[
rem-off: i
until
[
ofz: either rem-off > 100000 [ 100000 ][rem-off]
copy/part p ofz
rem-off: rem-off - ofz
0 >= ofz
]
return p
]
is what I am using in Rebletta, the webserver I'm developing. Here, mem
usage is still low and it skips 100Mb in 0.7 secs (2.2 Ghz PC with WinXP)
Which is slower that a skip would be, but workable for now.
--Maarten
Notes
- Quoted lines have been omitted from some messages.
View the message alone to see the lines that have been omitted