AltME groups: search

weird, I had no problem when I upgraded-did you do a fresh install 
?

I think that Cyphre or someone else could be contracted, to improve 
feeling to AltME in a big way - e.g. list-box with available worlds 
is simply awful - typical VID bug - you can't escape it by pressing 
escape, you can't escape it by clicking outside, you can't escape 
it in any way ... that just makes so bad impression here ....

I think that small things could make a real difference ... and I 
vote for some button to de-red groups at once :-)

is there a list of changes somewhere?

I kind of miss a list of changes in a window, when asking whether 
you want to upgrade

+1 on everything Pekr just said, actually I'd give it at least a 
+10 ;-)

same here but I'm using a wireless card in my laptop at a truck stop 
and it show 1.2.8 here

playing around with file sharing here... is it wise to just allow 
to run DOS batch files without permission? when I upload a .bat file 
and click on it from another host, it immediately runs.

Not a good idea IMO. I guess its altme-filesharing and they use 'run 
for launch? aybe they should launch through the browser isntead, 
would that protect?

anton:


1. on Machine A, an XP box, I create a bat file and add it to my 
file sharing list on my world

2. on Machine B, also an XP box, I open AltME on that same world 
with the same user account and notice that a new file is ready for 
download.
3. on Machine B, I click it and it downloads.

4. on Machine B, I click it again, and it immediately opens a DOS 
shell window and runs.

A bit of a security hole.

but that 's it like telling, then someone sends me format.exe in 
email, I save it, and once I click it, it formats my HD :-) Or am 
I missing something? How is that supposed to be a security hole?

there isn't even a requester saying "do you really want to run this 
app? could be dangerous. your computer could go boom." like webbrowsers 
do nowadays.

i guess he simply uses an os-call to launch the right app. Where 
the app is choosen by the os, based on filetype. And a *.bat runs..

I would log it as a bug

is it a font issue?

http://www.rebol.net/builds/lib-dep.html

how about the above link?

I saw 'freetype' as a requirement

YEah, this is an ongoing problem, we need a wiki for this.

the problem is related to X-windows DPI settings, which seems to 
change on a whim with varying window managers and distributions.

Anyone know how many world servers AltMe World servers will communicate 
while polling for changes?  I am just running a demo of a packet 
sniffer package  PRTG Packet Grapher and I get about 6-12 IP addresses 
connecting regularly with my computer.  The IP addresses all seem 
to end in either .69 or .223.

This should not be related with Altme at least I have have only one 
regular Altme connection (to www.rebol.net:5400) But I'm not running 
any world. If you have a world, you probably will see people connected 
to your world (altme) server

Yes, I have a world running but it is just for personal chat between 
family and doesn't get used very often.

This is a rather annoying bug with 1.2.8 ... not sure if it were 
present before .. but if you have something in the clipboard, then 
right click no longer copies anything to the clipboard.

Yes, I've noticed something has changed there. (Probably should check 
the AltME changes log for a different behaviour.)

It may be that it was just a different installation that did it ..

I would like to share a slideshow.

is it rebol on os x or altme, pasting code into altme doesn't keep 
newlines and tabs, if it is an altme issue then a fix would be very 
welcome 8)

Louis:  This is not my world, you would have to ask Carl.

Will, pasted from what?  That is the key.  Please look at the contents 
of your paste in a hex editor, and tell me if it has CR, LF, LF + 
CR,. or CR + LF.

I have a Mac, but today I'm away from it.

The source does matter, try writting a couple of lines with TextEdit, 
and also try copying and pasting from a website.

Tried a variety of combination 8)

>> t/text: rejoin ["this is" newline "a" newline "test"]   
== "this is^/a^/test"
>> show t

That works

>> t/text
== "this is^/a^/test"

This is from Notepad in WindowsXP.

Only the last test above was done in Windows XP. The others were 
done in OSX.


When I paste from TextEdit in MacOSX into a VID textarea, ^M is pasted, 
and it doesn't work right. When I paste from Notepad in Windows into 
a VID textarea, ^/ is pasted and it works correctly. So the results 
are different, yes.


When I copy and paste from a VID textarea to TextEdit, it looks alright, 
but this might be due to how TextEdit handles different line ending 
chars the same. I can't see the line endings in TextEdit. When I 
save the TextEdit file to disk in Windows Latin 1 format and read 
it in REBOL, the line endings are correct.


When I copy and paste between two VID text areas in OSX, it looks 
alright.


OSX Rebol/View has also another bug in which, when using cursor up 
or down in a text area, the cursor also moves one char to the left.

when doing a 'read clipboard:// in OSX, I still get the ^M. so it 
might not be VID after all.

I don't think this is an AltME problem, but a rebol/clipboard general 
problem. But I think that Mac Roman text formatting does use a different 
line ending and that comes through with the clipboard into rebol.

another issue is how text encoding is done. I just had a conversation 
in Danish with Geomol. While Danish letters look right on a mac it 
will look different in Windows and vice versa.

¾¿Œ <--- on a mac

I had a little script I was gonna post.  Maybe rebol.org instead.

REBOL /View (or AltMe)  just crashed on me..when I do the following 
on WinXP.

1. My Computer->Properties
2. Advanced tab->Performance

3. Change from "Let Windows choose what's best for my computer" to 
"Adjust for best performance'

4. Apply..... WinXP 'pops' up a dialog saying /View has crashed and 
ask for information to be sent over to MS.

Can someone try the above and see if it duplicates?

NOTE: I have AltMe running when I did that.

It all depends.


 250,000 integers vs 250,000 multi-column rows with large string values
	client device with 16MB RAM vs 'server' with 4GB RAM
	'local' query vs pulling the data over a network


There are plenty of optimization strategies. One technique, if concurrency 
is not an issue for you, is to have your query return 250,000 rowids 
then page through rows based on simple 'rowid in (...)' type queries. 
It's fast and efficient, but not too great if others are modifying 
the same table(s) at the same time.

I have no concurrency issue. Yes, this might be a good solution too.

Could you use a cursor?

(That would work with many other SQLs....Not sure about MySQL's support 
for them)

Yes, but SQLite doesn't has this concept. So you can only move forward 
through a result set. But the set isn't known at this stage.

That's a pity.

What's the best way to get back the columns of a table ONCE? I just 
need the column names in left-to-right order inside the table.

In general, the best strategy can be to rename the table, create 
a new table with the old name and the new schema, and then make an 
import query to transform and transfer the data from the old table 
to the new. Then you can drop the old table.

If you have a complex app, it's impossible to get the tables right 
the first step. What I have done so far is to only put data into 
columns that is required for SELECT statements. All other data is 
stored as a rebol block in one generic _DATA column.

Hmm... sounds good but is more like a brute force approach.

Just to repeat it here: What's the best way to get back the columns 
of a table ONCE? I just need the column names in left-to-right order 
inside the table.

By in left-to-right order, do you mean having the names returned 
in a single row, or will having them in a single column do?

I have a table with columns: [A B C D AA] and I just need to get 
back them in order [A B C D AA] and not something like [A AA B C 
D]

I think I can use as statement "SELECT * from <my-table-name> LIMIT 
1" to avoid doing a query that returns 100.000 records just to get 
the column names.

It returns info about a table, and by default returns info about 
the columns of the table, in the order they are declared in the table, 
including column names. From there, you can extract the names quite 
easily.

how about adding an option to either get the results 0 based or 1 
based

 ... best done in your own script as that's a fairly specific and 
 [probably] uncommon requirement. Same goes for other 'one-off' requirements 
 like "all uppercase object names", or "all lowercase object names".

Furthermore, I sometimes have the problem, that errors in sql statements 
are not caught by try, if I call it from a view gui. 
Is this a known problem?

Looks like your data is tripping the driver up. Please add the following 
line:

	print mold s


prior to the "either all [block? v ...] line and post the last result 
back here.


As for the second issue, haven't experienced this myself. Do you 
have a small code snippet that exhibits the problem?

I'll try to find a minimal code to show the non-catched errors ...

Hi Ashley, while trying to find a minimal code example ... I found 
the error ... ;-) 

That's the error message ...
** User Error: SQLite SQL logic error or missing database
** Near: make error! reform ["SQLite" error]

And it was caused by:

      if string? face/user-data [
         if error? set/any 'err try [

            set pAddress-disp first rule compose [pAddress get guid = (face/user-data) 
            *]
	; rule creates an sql string and starts calls 'sql with it
	; yadda yadda yadda ...
         ]
      ][probe disarm err]

Do you find the error??? 

Somehow the [probe disarm err] block moved to the wrong if ... 


I don't know how this could trigger _this_ error, but after I moved 
the block the error has not occurred again.

sqlite open command expects a fully qualified local file name ... 
the line in question prepends the supplied file name with current 
dir unless the file name is in fact a path.

The log-path issue is best resolved by adding a log-file word to 
the sqlite context that defaults to %sqlite.log. You can then do 
the following in your code:

	sqlite/log-file: %my-path/my-log-file.log

Ashley - I have problems with writing permissions under Linux. Till 
I resolve it, I wanted to disable logging. I looked at the driver 
and I can see, there is a word - log?: false, but when you try to 
write to log, you don't test for it. Is that ok, or is that variable 
for any other purpose of logging?

Hi, I know that some talk of encryption was held here some time ago, 
but currently I was asked to eventually protect sqlite data and I 
am not sure what is correct aproach. I would not go DB-as-a-file 
encryption, then "unpacking" into memory, or so. I prefer app level 
encryption, but I am not sure about searches, using LIKE directive. 
Would it work?

I haven't given this ADO stuff a try to see if it fits my requirements. 
But will do so.

Performance will be about 50% without encryption for the SQLite extension. 
But I don't think that it's that much. Reading/writing to disk is 
a lot slower compared to execution speed of processors. So, I expect 
about 25% performance loss.

It will become a replacement over time.

Anyway, I read it that .NET isn't required. So, it should be a drop-in-replacement. 
The only thing I don't know is, how to access the exteneded functions 
than.

There is a virtual machine, but the code you run is native code by 
the time you run it. All new APIs for Windows are built on .NET.

SQLite has a function extension mechanism, so it seems likely that 
you will access the extensions through that. We'll see.

I have made a simple change to the sqlite.r driver, so it now "supports" 
nested transactions in that it uses reference counting to determin 
when to send a BEGIN or END to SQLite. Works good. Ashley, should 
I mail you the changed version?

1.0.1 available at: http://www.dobeash.com/download.html


Bumped version to 1.0.x as code is no longer "beta" (I actually had 
a few people email me about that!)


Incorporates new IMPORT function (twice as fast as a tight INSERT 
loop) and Robert's nested transaction handling logic (so nested BEGIN, 
COMMIT, END and ROLLBACK statements are correctly handled). Also 
changed logging behavior so all logging is off by default (i.e. use 
/log to turn on *all* logging).

I have a problem when I import a CSV file. I read the file (1.5 -2 
MB), parse it and than write it out to SQLite.


For some records I get scrambeld (invalid strings) in the database. 
Importing the CSV file via other SQLite tools work without any problem.


It looks like the Rebol memory somehow gets messed up. Or could it 
be on the way to the DLL?

Thanks Brock. Robert, looks like you've hit the same error I RAMBOed 
here: http://www.rebol.net/cgi-bin/rambo.r?id=4063&


A real pain as it can be consistently reproduced (and hence avoided), 
but with no pattern [that I can see]. If Carl implements SQLite-like 
support (see latest blog entry) then hopefully this will become a 
moot point as we won't require the library call anymore.

From his blogs it appears that Carl is just extracting SQLite's btree 
and indexing engine, but leaving out the SQL stuff that duplicates 
functionality already in REBOL (think blocks and series functions). 
You may be able to access the data (a little unlikely), but it won't 
be SQLite support.

It also seems as if Carl is talking about a generic DB layer, such 
as is defined in python.  so any db, could just implement a derivative 
of the base class and be compatible to some extent.

This is probably a real dumb question, but what is the advantage 
of SQLite over say a REBOL object database?

is it possible to easily add a new column to a db table already populated 
with data without losing that data?

Thanks, Ashley.  But that website has a flaw; it can't be downloaded 
easily. It should be either one html page or else one pdf file.

Related: has anyone written a script to download a complete website?

Can't find any. i wrote a crude one once, but can't find it on my 
harddrive.

Louis: if you are on os x, this one work pretty neat http://www.hexcat.com/deepvacuum/index.html
, it will dl the whole site and rewrite urls to make them relative, 
or you coud try with acrobat pro if you want a pdf version of the 
site

if you can use non-rebol stuff, use wget. It's very effective at 
mirroring sites to a local drive.

and DeepVacuum is a GUI for wget ;-)

batch-download is not suited for downloading a whole website. It 
doesn't do any spidering.

Well, I suppose if you do the spidering, you can supply the links 
to batch-download. Still, it's not designed for that. Better to get 
a specific tool for that.

as long as you use wget in the same location as where you put your 
downloaded files temporarily, you don't need to set the environment 
variables.

Then it's just in a DOS prompt:

wget -r -l <depth> <some site>

And it will download the site to <some site>/ directory

That's right. It's very rare that wget fails to resume a download.

sqlite.com contains a general domain site.

A good entry point if you just want to grab the SQL syntax is: http://www.sqlite.org/lang.html

You have a few errors in there.


 - your values are comma seperated (e.g. use SQL reduce ["..." val1 
 val2 ...] instead

 - tgl_nikah/text appears in your INSERT but isn't present in your 
 display
	- your "DB Info" button can use ... print ROWS "base"

Hi, I have a problem wheren using the ? binding feature with integer! 
values. For example:
	["select * from a where mynum  = ?" 1]
won't give an results. But
	["select * from a where mynum = ''1']
gives results.

Does:

	["select * from a where mynum = ?" "1"]

work? If so, then your "integer" is in fact a string.

I have created a semicolon seperated file and imported it via the 
SQLite command line tool. All numbers where just plain included, 
not guarded by " chars.

Ashley, I notice that sqlite.r value binding is missing some records 
when using LIKE.


>> sql ["select * from base where alamat like ?" "%Grand%"]   <<=====<<< 
This fails.
== "No rows selected."

>> sql ["select * from base where alamat like '%Grand%'"]      <<====<<< 
This finds a record.

I read all thre using read/lines, choose info I want, append it to 
resulting block. Probing block shows no defects. I believe it is 
a REBOL low level bug with some hidden chars. It happened on me in 
the past already, in different situation ...

If I import one file at a time, clear the block, then data is OK 
in sqlite, but if I append first to one block, then insert into sql, 
data is corrupted on few random places ...

I am confused about what driver is doing ... the difference of Direct 
mode. Hmm, maybe I do - there is a difference between the string 
mode, and block mode. With string mode, the driver does not touch 
the expression, so I am responsible for putting VALUES('06-Dec-2006') 
or VALUES('"06-Dec-2006"') - simply put - if I want in db to have 
my date value being represented as a REBOL string, I have to put 
it into parenthesis myself. So actually parenthesing it twice, as 
sqlite itself already uses '06-Dec-2006' and considers it being a 
string ...

Maybe there is a bug with block mode?

Does block mode do datatype conversions? Perhaps there is a type 
mismatch.

Petr, I had the same problem. There is a RAMBO ticket and Ashley 
posted his findings about it.