[Monetdb-developers] clients on windows 32...
Why the option \h - show the readline history mclient.exe -lsql --help Is not available on windows? Is an design feature or a bug? Regards, Romulo
On Fri, Oct 12, 2007 at 01:22:23PM +0200, Romulo Goncalves wrote:
======== $ Mapprove.py --help -------- [...] -x <ext> approve only output files *.<ext><sys> (<ext> = 'out' or 'err') (default: <ext> = 'out' & 'err') -S <sys> approve specific output *.<ext><sys> (<sys> = '[.(<OS>[<RELEASE>]|<DIST>[<VERSION>])][.(32|64)bit][.oid(32|64)][.Five][.STATIC]', (default: longest match for <sys> = '[.(Linux[2.6.20-1.2933.fc6]|Fedora[6])]') [...] ======== or in short: Mapprove.py ... -S.WindowsNT ... Stefan
-- | Dr. Stefan Manegold | mailto:Stefan.Manegold@cwi.nl | | CWI, P.O.Box 94079 | http://www.cwi.nl/~manegold/ | | 1090 GB Amsterdam | Tel.: +31 (20) 592-4212 | | The Netherlands | Fax : +31 (20) 592-4312 |
On Fri, Oct 12, 2007 at 02:33:37PM +0200, Romulo Goncalves wrote:
Yes I know I just tried the shortcut (ask by email), but once the answer took more than 10 minutes to arrive a decide to Jump again on the README.
Try it the other way, next time ;-)
PS: Windows it would be nicer because NT is gone ;)
"Officially" (internally), "NT" was "WindowsNT 4.*", while Windows2000 was "WindowsNT 5.0", and XT was/is "WindowsNT 5.2" --- at least that's what command `ver` used to say on a Windows command prompt (hence, not my/our decisions but that of the guys from Redmond). Just try yourself. (Don't (want to) know what Vista is ...) Stefan -- | Dr. Stefan Manegold | mailto:Stefan.Manegold@cwi.nl | | CWI, P.O.Box 94079 | http://www.cwi.nl/~manegold/ | | 1090 GB Amsterdam | Tel.: +31 (20) 592-4212 | | The Netherlands | Fax : +31 (20) 592-4312 |
Hello,
is \h supposed to work on Linux (Ubuntu feisty)? I get: $ mclient -l mal -d demo -H mal>\h MAPI = monetdb@localhost:50000 QUERY = \h ERROR = !SyntaxException:parseError:\h !SyntaxException:parseError:^<identifier> expected However, readline functionality is present: Ctrl-p, Ctrl-n, Ctrl-a, Ctrl-e all work correctly. Kambiz
darabi@web.de wrote:
Problem NOTconfirmed on Fedora. It works with both empty/non-empty history file. However, '\' should be the first character of the line, otherwise you indeed get the error being mentioned. Please check. If it fails then a deeper analysis of the Ubuntu version is needed. mk@gio::~> mclient -l mal -d demo -H mal>\h mal>io.print(1); [ 1 ] mal>\h 1 \h 2 io.print(1); mal>\h 1 \h 2 io.print(1); 3 \h mal> \h MAPI = monetdb@localhost:50000 QUERY = \h ERROR = !SyntaxException:parseError: \h !SyntaxException:parseError:^<identifier> expected mal>
Hi, in Kambiz' case, the error message is not complaining about an extra space, but about the '\' as first character on the line. Hence, the usual questions: Kambiz, which version of MonetDB are you using, and how did you install it --- if I recall correctly from previous mails, you did install using the monetdb_install.sh script, right, and hence use the super source tarball? But which version? Further, a simple mclient -l mal --help (with a running mserver5) will reveal whether readline support (and hence \h) is indeed enabled in your build of MonetDB. Finally, do other \* commands work? Stefan On Sat, Oct 13, 2007 at 08:21:01AM +0200, Martin Kersten wrote:
-- | Dr. Stefan Manegold | mailto:Stefan.Manegold@cwi.nl | | CWI, P.O.Box 94079 | http://www.cwi.nl/~manegold/ | | 1090 GB Amsterdam | Tel.: +31 (20) 592-4212 | | The Netherlands | Fax : +31 (20) 592-4312 |
Stefan Manegold wrote:
No, the error indicates that '\' is sent to the server, where the MAL parser does not recognize it as a legal start. '\' as first characters are already trapped in the basic readClient() function, before the parser takes control. That's why I am suspicious if there is a leading blanc character.
On Sat, Oct 13, 2007 at 09:09:59AM +0200, Martin Kersten wrote:
I just compared Kambiz v ERROR = !SyntaxException:parseError:\h !SyntaxException:parseError:^<identifier> expected ^ with Martin v ERROR = !SyntaxException:parseError: \h !SyntaxException:parseError:^<identifier> expected ^ In the second case (Martin) there's indeed a ' ' that the error message complains about, while in the first case (Kambiz) there is no ' ', but the error message complains about the '\' --- at least that's how I read the above error messages. Stefan
-- | Dr. Stefan Manegold | mailto:Stefan.Manegold@cwi.nl | | CWI, P.O.Box 94079 | http://www.cwi.nl/~manegold/ | | 1090 GB Amsterdam | Tel.: +31 (20) 592-4212 | | The Netherlands | Fax : +31 (20) 592-4312 |
MonetDB-Mars-SuperBall-SR3.tar.bz2. I have the impression that something is wrong with the source tarball.
Do you mean 'mclient -l mal -H' ? Readline works as I mentioned:
However, readline functionality is present: Ctrl-p, Ctrl-n, Ctrl-a, Ctrl-e all work correctly.
And the history is also saved correctly: I can get to the commands of the last session with Ctrl-p.
Finally, do other \* commands work?
\q does. \d doesn't: $ mclient -l sql -d demo -H sql>\d MAPI = monetdb@localhost:50000 QUERY = \d ERROR = !unexpected symbol (\) more> I would like to re-build. Which one would you recommend: --nightly=target download and install a nightly snapshot of the stable or current branch, target must be 'stable' or 'current' --cvs checkout a CVS snapshot of the current branch Thanks Kambiz
On 13-10-2007 10:15:20 +0200, darabi@web.de wrote:
Can you retrieve the installed readline version using Ubuntu's package manager somehow? Also whether you are using GNU readline, or BSD's libedit?
I think --nightly=stable will give you sources similar to the release candidates, so that would be most helpful.
On Sat, Oct 13, 2007 at 10:15:20AM +0200, darabi@web.de wrote: [...]
nope. I meant what I wrote: `mclient -l mal --help` to get the help message and see whether this lists '\h' or not. Stefan -- | Dr. Stefan Manegold | mailto:Stefan.Manegold@cwi.nl | | CWI, P.O.Box 94079 | http://www.cwi.nl/~manegold/ | | 1090 GB Amsterdam | Tel.: +31 (20) 592-4212 | | The Netherlands | Fax : +31 (20) 592-4312 |
On Sat, Oct 13, 2007 at 11:30:47AM +0200, darabi@web.de wrote:
That exactly was my intention: get confirmation that the "old" MapiClient used in the Mars-SR3 tarball (in that "MonetDB Clients 1.18" version, mclient is just a tiny script that allowas to call MapiClient as mclient) is significantly different from the new mclient from the latest "MonetDB Clients 1.20" Stable branch that will be distributed with the upcoming release. Next to changing the name (MapiClient -> mclient) the functionallity (in particular help messages and '\'-commands) has been extended, cleaned-up and (most importantly) *fixed* and thus made work (properly) at all. Once again, this shows how impartant it is that both users and developers state clearly which version they are using/talking about ... Stefan -- | Dr. Stefan Manegold | mailto:Stefan.Manegold@cwi.nl | | CWI, P.O.Box 94079 | http://www.cwi.nl/~manegold/ | | 1090 GB Amsterdam | Tel.: +31 (20) 592-4212 | | The Netherlands | Fax : +31 (20) 592-4312 |
participants (6)
-
darabi＠web.de -
Fabian Groffen -
Martin Kersten -
Romulo Goncalves -
Sjoerd Mullender -
Stefan Manegold