Thanks a lot, Neil. Finally a "make sense" answer. Thus, know I can see
why the script was not failing within eval (due to PrintError being on). It
is quoted in the docs and my fault for overlooking it. Definitelly when we
answer questions on this list we mostly tell people to turn their RaiseError
and PrintError on, though not explaining a clear difference between
PrintError's involvement in the process. It should probaly be made more
clear that PrintError should not be turned on when using eval{} for
transaction processing.
Ilya Sterin
-----Original Message-----
From: Neil Lunn
To: Sterin, Ilya; '[EMAIL PROTECTED]'
Sent: 05/07/2001 9:38 PM
Subject: RE: executing atomic transactions in DBI
Should this really be discussed off the list? Abolutely not.
Ilya, your connect string in this script uses PrintError, as you
knowingly
mention yourself. I thought that this was made very clear in the docs.
>From the fine Manual:
""PrintError"" (boolean, inherited)
This attribute can be used to force errors to generate
warnings (using "warn") in addition to returning error
codes in the normal way. When set "on", any method
which results in an error occuring will cause the DBI
to effectively do a "warn("$class $method failed:
This of course overides your calls to die, as you seem to imply as an
intended behaivor in your message. Thus the calls to die are actually
not
needed. If you want different behaivour, remove PrintError => 1 from the
connection.
If RaiseError is on, the script will die. If it is off the script will
not
die unless told to. If PrintError is on an error is trapped as a warn,
otherise no warn is issued.
Back to the start of this message. You reported to the list that you
believed RaiseError was not always wroking. You posted your script, and
people can scrutinise and reply. That is the one great benefit of this
list
existing. We can all help each other.
Also as Tim mentioned himself recently, we should be quoting the docs
more
often. This does help re-enforce their worth. If there is something not
clear in the documentation, it should also be discussed here. If there
is a
better way to do it, it will almost certainly end up in the
documentation.
Neil
> -----Original Message-----
> From: Sterin, Ilya [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, May 08, 2001 12:32 AM
> To: '[EMAIL PROTECTED]'
> Subject: RE: executing atomic transactions in DBI
>
>
> Randal, we'll discuss off list since I totally agree that
> this (rumor??) can
> get out of hand. The script provided is supposed to fail
> when the query
> does not execute due to mispelled table name, but it rather
> prints the error
> message (as it suppost to) but then proceeds to print the
> rest of the form.
> The problem might also be in my script, server configuration,
> etc..., so
> before this gets out of hand, I'd rather this be investigated.
>
> Ilya Sterin
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
> To: Sterin, Ilya
> Cc: 'Stephen Clouse '; '[EMAIL PROTECTED] '; '[EMAIL PROTECTED] '
> Sent: 05/07/2001 8:05 AM
> Subject: Re: executing atomic transactions in DBI
>
> >>>>> "Ilya" == Sterin, Ilya <[EMAIL PROTECTED]> writes:
>
> Ilya> Not sure what you consider a rumor, I just simply stated having
> Ilya> a problem with it before and that in my experience I always
> Ilya> check for return value, due to either the query failing and/or
> Ilya> executing without failing, but still producing an unexpected
> Ilya> result. By implying that the trasaction does not fail without
> Ilya> checking for return, does not necessarily mean that there was an
> Ilya> SQL error, but rather the query produced an unexpected result.
>
> I think perhaps the problem was sloppy terminology. A RaiseError=1
> block should prevent you from ever needing "|| die" after each call.
> Let's call that a "DBI error".
>
> You seem to also be including the problem of a "domain error" where
> something doesn't make sense, even though DBI has performed perfectly
> in getting the query to and from the database. No amount of "|| die"
> will detect that, and therefore RaiseError=1 will be ineffective.
>
> So I think I was claiming that you were spreading FUD for claiming
> that RaiseError=1 still needs "|| die" to trigger DBI errors. I'm
> still waiting for a sample of that. The code you just posted does not
> have an eval {} block, so I doubt that you sent code that actually
> shows the problem.
>
> But it appears that you also want a way of detecting domain errors
> without having to check the length of the results. That's your code,
> your responsibility.
>
> Ilya> Sorry if I offended you in any way, since you decided to flame
> Ilya> me, without reading the whole thread.
>
> You're not offending me, and I'm not flaming you. I'm just in rumor
> control mode, since something like this can get out of hand.
>
> Please post some code that uses RaiseError=1, no "|| die" and an eval
> block, and yet still doesn't detect a DBI error, and that we can
> verify. I'll bet this month's rent that either you can't, or that
> we've found a bug in DBI that must be corrected.
>
> In any case, I'll stand by my claim that as I understand it, it's
> intended and implemented that RaiseError=1 completely replaces the
> need for adding "|| die" to detect DBI errors.
>
> --
> Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777
> 0095
> <[EMAIL PROTECTED]> <URL:http://www.stonehenge.com/merlyn/>
> Perl/Unix/security consulting, Technical writing, Comedy, etc. etc.
> See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl
> training!
>
________________________________________________________________________
__
Please Note :
Only the intended recipient is authorised to access or use this e-mail.
If
you are not the intended recipient,
please delete this e-mail and notify the sender immediately. The
contents
of this e-mail are the writer's
opinion and are not necessarily endorsed by the Gunz Companies unless
expressly stated.
We use virus scanning software but exclude all liability for viruses or
similar in any attachment.