headers
Michael Platzer | 2 Aug 12:48
Picon

Re: uploadLimit - catching errors

hi hannes,
it works now, thanks! but all other request-parameters are lost, resp. never set, since the error is thrown in helma [upload.parseRequest(request);] before the passed parameters can be set. i took a look at the code, and couldn't think of an easy way of how to circumvent this problem :-(
this implies that handling such an error in an application has to happen in the action itself (and not further down in the processing chain), which is not too nice, but sthg that i can live with.
i guess you don't see a quick solution to this problem either, do you?
thanks again,
  michi


Michael Platzer schrieb:
Hannes Wallnoefer schrieb:

Ok,

I added a new switch to apps.properties to catch upload errors in
application code. By setting

   appname.uploadSoftfail = true

Helma doesn't generate an immedate error response if the file upload
fails. Instead, req.data.helma_upload_error is set and contains a
string describing the error.

Apps need to be changed to work with this feature, so the default
behaviour is still to generate the error page.

Please send me feedback if you test this new feature, I haven't tried it myself.
 

it doesn't work yet, but i think you just forgot to read uploadSoftfail in the AppDescriptor-function.

greets,
 michi


_______________________________________________
Helma-user mailing list
Helma-user <at> helma.org
http://helma.org/mailman/listinfo/helma-user



-- DI Michael Platzer technical director -------------------------------------- knallgrau new media solutions gmbh pezzlgasse 7|1, A-1170 wien phone: +43 - 1 - 522 76 37 fax: +43 - 1 - 522 76 38 http://www.knallgrau.at weitere News im Knallgrau Weblog: http://www.knallgrau.at/company/weblog 
_______________________________________________
Helma-user mailing list
Helma-user <at> helma.org
http://helma.org/mailman/listinfo/helma-user
Permalink | Reply | 0 comments |
headers
smi | 2 Aug 17:34

Oracle / Set transaction / Error

Hi *!

As you may have noticed on this mailing list, we have some installations
of twoday running with Oracle databases using jdbc/odbc. Recently we
encountered some problems, namely the following error message appears
randomly when requesting pages through the web browser:

[Oracle]ORA-01453: SET TRANSACTION must be the first statement of
transaction

Using Google to demystify this error didn't help much. Any hints on how
to solve this issue?

Yours
smi
Permalink | Reply | 0 comments |
headers
thomas.foerster | 2 Aug 17:59
Picon
Favicon

RE: Oracle / Set transaction / Error

Hi!
Error ORA-01453 is generally raised if you want to set a rollback-segment (SET TRANSACTION USE ROLLBACK
SEGMENT ...) for the next transaction without commiting the transaction before.

Yours
Tom

> -----Original Message-----
> From: helma-user-bounces <at> helma.org [mailto:helma-user-bounces <at> helma.org]
> On Behalf Of smi
> Sent: 02 August 2005 17:35
> To: helma-user <at> helma.org
> Subject: [Helma-user] Oracle / Set transaction / Error
> 
> Hi *!
> 
> As you may have noticed on this mailing list, we have some installations
> of twoday running with Oracle databases using jdbc/odbc. Recently we
> encountered some problems, namely the following error message appears
> randomly when requesting pages through the web browser:
> 
> [Oracle]ORA-01453: SET TRANSACTION must be the first statement of
> transaction
> 
> Using Google to demystify this error didn't help much. Any hints on how
> to solve this issue?
> 
> Yours
> smi
> 
> _______________________________________________
> Helma-user mailing list
> Helma-user <at> helma.org
> http://helma.org/mailman/listinfo/helma-user
Permalink | Reply | 501 comments |
headers
smi | 2 Aug 18:05

Re: Oracle / Set transaction / Error

Hi Tom!

Thanks for the quick answer, but this doesn't help either. I suppose the 
way database queries are issued is completely up to helma, therefor it 
is quite hard to identify whether a commit is done or not.

Yours
Michael

thomas.foerster <at> orf.at schrieb:
> Hi!
> Error ORA-01453 is generally raised if you want to set a rollback-segment (SET TRANSACTION USE ROLLBACK
SEGMENT ...) for the next transaction without commiting the transaction before.
> 
> Yours
> Tom
> 
> 
>>-----Original Message-----
>>From: helma-user-bounces <at> helma.org [mailto:helma-user-bounces <at> helma.org]
>>On Behalf Of smi
>>Sent: 02 August 2005 17:35
>>To: helma-user <at> helma.org
>>Subject: [Helma-user] Oracle / Set transaction / Error
>>
>>Hi *!
>>
>>As you may have noticed on this mailing list, we have some installations
>>of twoday running with Oracle databases using jdbc/odbc. Recently we
>>encountered some problems, namely the following error message appears
>>randomly when requesting pages through the web browser:
>>
>>[Oracle]ORA-01453: SET TRANSACTION must be the first statement of
>>transaction
>>
>>Using Google to demystify this error didn't help much. Any hints on how
>>to solve this issue?
>>
>>Yours
>>smi
>>
>>_______________________________________________
>>Helma-user mailing list
>>Helma-user <at> helma.org
>>http://helma.org/mailman/listinfo/helma-user
> 
> 
> 
> _______________________________________________
> Helma-user mailing list
> Helma-user <at> helma.org
> http://helma.org/mailman/listinfo/helma-user
>
Permalink | Reply | 501 comments |
headers
Hannes Wallnoefer | 3 Aug 10:14
Picon
Gravatar

Re: Helma and fast-cgi

Hi Chris,

What I did was to to serve fastcgi pages with the C++ fastcgi library
and natively compiled Java with gcj. It works well, but I doubt it's
what you need for your client. What you could do is use the Java
fastcgi library from fastcgi.com to build a Helma invoker similar in
functionality to helma.servlet.StandaloneServletClient. That should be
doable but not necesssarily easy, since the fastcgi interface looks
pretty different from the Servlet API. Plus, I doubt that you'll be
able to do this without touching the Apache configuration at all. If
you try, please let me know how it goes.

Hannes

2005/8/3, Chris Zumbrunn <chris <at> czv.com>:
> Hannes,
> 
> I know you recently looked into using the fast-cgi protocol for some
> reason. How far did you take this? Is there now a fast-cgi gateway that
> can be used with Helma? A client is currently running a fast-cgi based
> solution that I implemented some years ago and I would like to change
> that application to use Helma. The customer would prefer not to touch
> the apache configuration on that machine, so I'll probably only be able
> to make the switch if I can run Helma as a fast-cgi application.
> Realistic?
> 
> chris <at> czv.com  +41 329 41 41 41
> Chris Zumbrunn Ventures - http://www.czv.com/
> Internet Application Technology - Reduced to the Maximum
> 
>
Permalink | Reply | 1 comment |
headers
Chris Zumbrunn | 3 Aug 10:48
Gravatar

Re: Re: Helma and fast-cgi

Thanks!  I'm looking into it right now.

Chris

On Aug 3, 2005, at 10:14 AM, Hannes Wallnoefer wrote:

> Hi Chris,
>
> What I did was to to serve fastcgi pages with the C++ fastcgi library
> and natively compiled Java with gcj. It works well, but I doubt it's
> what you need for your client. What you could do is use the Java
> fastcgi library from fastcgi.com to build a Helma invoker similar in
> functionality to helma.servlet.StandaloneServletClient. That should be
> doable but not necesssarily easy, since the fastcgi interface looks
> pretty different from the Servlet API. Plus, I doubt that you'll be
> able to do this without touching the Apache configuration at all. If
> you try, please let me know how it goes.
>
> Hannes
>
> 2005/8/3, Chris Zumbrunn <chris <at> czv.com>:
>> Hannes,
>>
>> I know you recently looked into using the fast-cgi protocol for some
>> reason. How far did you take this? Is there now a fast-cgi gateway 
>> that
>> can be used with Helma? A client is currently running a fast-cgi based
>> solution that I implemented some years ago and I would like to change
>> that application to use Helma. The customer would prefer not to touch
>> the apache configuration on that machine, so I'll probably only be 
>> able
>> to make the switch if I can run Helma as a fast-cgi application.
>> Realistic?
>>
>> chris <at> czv.com  +41 329 41 41 41
>> Chris Zumbrunn Ventures - http://www.czv.com/
>> Internet Application Technology - Reduced to the Maximum
Permalink | Reply | 0 comments |
headers
tobias.schaefer | 9 Aug 12:05
Picon
Favicon

commons-fileupload.jar

hallo hannes

seit der umstellung auf das commons-fileupload.jar werden die pfade von uploades (die nur win/ie
dummerweise immer komplett als filename mitsendet) nicht mehr durch slashes, sondern durch
backslashes getrennt (d.h. so, wie es unter windows abgelegt wurde).

das war zuvor noch anders, vermutlich hat ein codeschnipsel dafür gesorgt, dass nur slashes im pfad auftauchen.

abgesehen davon, dass es etwas überraschend aufgetaucht ist, ist es kein schwerwiegendes problem; es
lässt sich leicht in unserm code beheben.

andererseits aber könnte es ja sein, dass bei der umstellung dieses codeschnipsel abhanden gekommen ist
(oder absichtlich entfernt wurde, weil es obsolet erschien). dann würde ich es begrüßen, wenn es
wieder eingebaut werden würde.

auf jeden fall wollte ich es kundtun, falls andere nach diesem fehler suchen sollten.

andere browser (firefox, safari) verursachen dieses problem nicht, sie senden eh nur den namen des files
mit, nicht den ganzen pfad.

ciao,
tobi
Permalink | Reply | 1 comment |
headers
Hannes Wallnoefer | 9 Aug 12:23
Picon
Gravatar

Re: commons-fileupload.jar

Hiho,

> andere browser (firefox, safari) verursachen dieses problem nicht, sie senden eh nur den namen des files
mit, nicht den ganzen pfad.

Wenn ich das richtig verstehe, waere das ideale Verhalten das,
einheitlich immer nur den eigentlichen Filenamen zu verwenden, oder?

Hannes

> 
> ciao,
> tobi
> 
>
Permalink | Reply | 0 comments |
headers
Hannes Wallnoefer | 9 Aug 15:40
Picon
Gravatar

Re: commons-fileupload.jar

Am 09.08.05 schrieb tobias.schaefer <at> orf.at <tobias.schaefer <at> orf.at>:
> > Wenn ich das richtig verstehe, waere das ideale Verhalten das,
> > einheitlich immer nur den eigentlichen Filenamen zu verwenden, oder?
> 
> ja, wenn das möglich ist, wäre es m.e. tatsächlich am besten.

Hab die Aenderung gerade eingecheckt, eventuell vorhandene Pfade
werden ab jetzt entfernt (sowohl mit /" als auch "\" als Separator).

hannes

> 
> ciao,
> tobi
> 
>
Permalink | Reply | 0 comments |
headers
Michael Platzer | 12 Aug 10:01
Picon

defined functions for a prototype

Hi,

Just a short question with a hopefully easy answer:
How can i loop through all defined functions for a certain prototype?

thanks,
  michi
Permalink | Reply | 7 comments |

Gmane