I was initializing my two nodes at roughly the same time, bringing up
a new release and I got this error from Transfersync. I reloaded and
it went away but it leads me to believe there might be some locking or
concurrency issues that need to be looked at?
Routines cannot be declared more than
Quick update on this - it's looking caching related. I have a page
that's dumping out my getRoleArray() and it just went from 151 to 149
while I was writing my post.
Brian
--~--~-~--~~~---~--~~
Before posting questions to the group please read:
On Jan 4, 1:46 pm, Mark Mandel mark.man...@gmail.com wrote:
This was a bug fix for the new Object meta data lifecycle, where there
were cases where the meta data's parents were sometimes not found, so
this wouldn't affect object caching.
Was there any other changes you made to the roll
Is there a way to fully disable the Transfer generation of numeric
keys? I normally use UUIDs so I haven't come across this before, but
it seems (on Postgres, Transfer 1.1) that:
id name=token type=numeric generate=false / - will expect a
table_seq to exist to fetch the ID from
id name=token
On Jan 26, 1:14 pm, Mark Mandel mark.man...@gmail.com wrote:
Just set the ID before running save() or insert(), then Transfer will
attempt to insert it with that value.
See:http://docs.transfer-orm.com/wiki/Primary_Key_Management.cfm#Manual
Weird, that's what I was doing. It looked like
In case anyone else finds this helpful, I created a printable cheat
sheet of the Transfer generated methods. Sometimes I model my objects
differently than my data model and I find this a helpful reminder of
which methods are available for each relationship type. It was
especially useful when I
A couple of weeks ago I released a big update to my app. When I roll
out new versions and restart ColdFusion on my two web servers, I'm
seeing service layer objects which are injected via Brian K's
TDOBeanInjectorObserver disappear in some of the first Transfer
objects that are loaded. My app
I just rolled out an update and hit this again which reminded me to
check in. Big props for the responses as well as the rapid fix. I
love it when I'm not the only one playing Mr. Edge Case (I think
that's my nick name??).
I'll add this into the mix and report back - thanks!
Brian
On Mar
On Mar 24, 5:45 pm, Mark Mandel mark.man...@gmail.com wrote:
How on earth is cf code causing a null pointer
This is 1.1 stable, or BER?
1.1 stable.
Signed,
Mr Edge Case
--~--~-~--~~~---~--~~
Before posting questions to the group please read:
On Jul 1, 10:40 pm, Elliott Sprehn espr...@gmail.com wrote:
I updated to the Trunk version and the errors have seemed to go away.
I'm going to load test it and low mem test it later this week and see
if it holds up.
I'm interested in your findings - I'm reasonably sure there is a mem
leak
On Sep 12, 3:53 am, Scott Brady dsbr...@gmail.com wrote:
This is what's in my onApplicationStart:
!--- Transfer initialization ---
cfset application.myTransferFactory =
createObject(component,transfer.TransferFactory).init(/config/datasour
Mark and I chatted tonight and discussed the possibilities of
restructuring Transfer's caching architecture to be pluggable as he
posted over the weekend. The idea is, out of the box, to provide a
Javaloader 1.0/eHcache implementation that would be more flexible and
would solve the memory issues
On Sep 21, 8:53 am, Clint clintmil...@gmail.com wrote:
We sustained our greatest # of unique visits ever last Friday and did
not require a restart of the JVM due to an OutOfMemoryException.
Previously, under heavy load (for us), JVM restarts would be required
every 45 to 90 minuteswhich
On Sep 20, 10:46 pm, Brian G brian-goo...@vfive.com wrote:
Mark and I chatted tonight and discussed the possibilities of
restructuring Transfer's caching architecture to be pluggable as he
posted over the weekend. The idea is, out of the box, to provide a
Javaloader 1.0/eHcache implementation
with Ghirardelli)
On Sep 29, 3:44 pm, Mark Mandel mark.man...@gmail.com wrote:
Wow!
What a way to wake up in the morning!
I will coordinate with Brian G (I can never spell your last name Brian ;o)
), and lock down some dates for getting this stuff done and out the door.
Which will also
I worked with Mark this afternoon to convert my system and get it
running on the new pluggable architecture. Pretty straightforward and
I've documented it for anyone else who wants to give it a quick go:
http://www.ghidinelli.com/2009/11/18/using-transfer-ehcache
Hope that helps! I've got it
This would impact unit tests that instantiate Transfer as well,
right? For example, I have coldspring instantiate my Transfer
instance and I instantiate CS to run unit tests over and over and
over. I can see how a suite of tests could exhaust the memory
quickly.
How would you suggest handling
On Nov 24, 9:54 pm, Matthew matthewbchamb...@gmail.com wrote:
cache. If I turn the cache back on it's lightning fast. The reason I
turned the cache off is because the CMS has a back end (which is a
seperate app so effectively has it's own Transfer cache). Perhaps I
need to work out a way to
On Nov 23, 2:38 pm, Mark Mandel mark.man...@gmail.com wrote:
How would you suggest handling this? Some kind of generic teardown?
Sounds good to me :D
Working like this (based on
http://www.mail-archive.com/transfer-dev@googlegroups.com/msg00730.html):
cffunction name=setUp returntype=void
My company is heavily invested in Transfer today. I'm doing some
future planning and it looks like Hibernate/CF9 ORM is the long-term
solution. I'd like to hear if anyone has migrated a Transfer
application to CF9 ORM/Hibernate and, if so, what was required? I'm
specifically interested in two
I have this try-delete-fall-back-to-soft-delete code:
cfset var clubmember = getClubMember(arguments.uidClubMember) /
cftry
cfset variables.transfer.delete(clubmember) /
cfreturn true /
cfcatch type=any
!--- there are FK dependencies, so mark as deleted
Mostly because there are many reasons why RI would prevent a delete so
the number of checks is many and keeping them up to date might prove
mistake-prone.
Brian
On Aug 11, 4:27 pm, Mark Mandel mark.man...@gmail.com wrote:
Stupid question, why not just do a check with some sql, rather than
I have members O2M clubmemberships in my app and I have a workflow
where members can edit clubmemberships on their account. The
clubmembership object is lazy loaded and proxied as an array. The
flow looks like:
* member.getClubMembership(clubid)
* set the values
* save it
Obviously sometimes
No, the (composed) object in the session is always in the Transfer
cache. This is why the values should be the same AFAICT. The
relationship looks like:
member o2m --- clubmembership --- m2o attendee
Before you can get to having a session.attendee, you must have a
clubmembership. However,
I personally have struggled with memory leak issues for a long time,
and some custom work by Mark to ehCache combined with frequent cache
discardAll()s has given me a stable app. For those of you still using
Transfer (or really any large OO CF app), I thought I'd share this
discussion on SO that
I'll share a few more odd errors just to add context. I restarted the app
(refreshed Transfer) and now several pages that use Transfer (but not all
pages) are failing with:
Expression: Unable to process the result of the XMLSearch for ''.
ColdFusion is unable to process the result of the
On Wed, Mar 20, 2013 at 6:01 PM, Brian G brian-...@vfive.comjavascript:
wrote:
I'll share a few more odd errors just to add context. I restarted the
app (refreshed Transfer) and now several pages that use Transfer (but not
all pages) are failing with:
Expression: Unable to process
)?
Should I switch back to trunk?
Brian
On Wednesday, March 20, 2013 5:02:18 PM UTC-7, Mark Mandel wrote:
On Thu, Mar 21, 2013 at 10:59 AM, Brian G brian-...@vfive.comjavascript:
wrote:
Mark - thanks, I've added some debugging but now I can't get it to break,
of course. I don't see unit
Ok, let me throw out an oddity... and see if anyone can help me think
through where to log to understand what is happening:
I have an object, event.event, it has a property, uidRegistrar. I'm trying
to create an event object and the decorate has a configure() that sets it
null:
In my searching I have uncovered people planning to do it or saying they
did it but no real specifics other than to use abstract classes.
On Wednesday, March 27, 2013 3:25:24 PM UTC-7, Mark Mandel wrote:
There have been some posts on list of people who have done exactly that.
Pretty sure
be a red herring.
Mark
On Tue, Apr 2, 2013 at 8:15 AM, Brian G brian-...@vfive.com javascript:
wrote:
Ok, so I let the app run for a few days and now I'm getting errors while
trying to basically instantiate Transfer objects. I've made some changes
to public/private packaging so I can
On Monday, April 1, 2013 6:38:18 PM UTC-7, Mark Mandel wrote:
In the stack overflow - is there any CF code being processed?
Mark
Stack trace is just filled with the repetitive XML processing error. It
never shows any CFML templates/codes/lines/etc. Makes it pretty hard to
diagnose??
On Tuesday, April 2, 2013 8:47:51 PM UTC-7, Mark Mandel wrote:
Try/catch inside the XMLFileReader - if it overflows, then load up the
path again. It knows where the XML file is, so it can always try it again.
http://www.2ddu.com/
That did it... on my wedged future production server, first
On Wednesday, April 3, 2013 7:05:27 PM UTC-6, Brian G wrote:
That is where I'm doing it - but I wanted to test in prod that reloading
the XML did indeed fix the problem, and it did.
Here's what I wound up with in XMLFileReader.cfc:
try
{
results.addAll(xmlSearch(collection[counter
Alan,
This is a very interesting discovery. I'm on CF10 and have posted about
similar errors before - where CF reports it can't find a method on a
Transfer object which clearly exists (because it was running before the
restart and runs again after an additional restart). I have added some
So did Sean take over the project? :)
On Wednesday, April 16, 2014 5:56:10 AM UTC-7, Sean Coyne wrote:
If it is any help, I've imported the SVN history from the Riaforge SVN
into a Github repo.
https://github.com/seancoyne/transfer
I'm happy to merge any pull requests I receive.
On
Dylan,
I have also had these errors. I'm not sure if they are related to the
xmlSearch() thread-safety issues or not, but this code solved the problem
for me:
in com/events/EventManager.cfc around line 257:
try {
// when we frequently run discardAll(), the page then tries to fire
Since I was deep in the guts of Transfer this week solving my xmlSearch()
insanity, I have created a patched GitHub repository here:
https://github.com/ghidinelli/transfer
Here's what I've updated:
* Merged the pluggable_cache to be the master
* Worked around the xmlSearch() issue
* Added
On Tuesday, July 22, 2014 9:47:11 AM UTC-7, Tom McNeer wrote:
Is there any documentation for the use of the new caching configuration?
And are there other known areas where an old config might cause errors?
Tom,
My apologies, this was so long ago for me I forgot it all changed. Here's
James,
Does that just disable caching for the class emc?
On Wednesday, July 23, 2014 10:18:58 AM UTC-7, James Allen wrote:
objectCache
cache class=emc
scope type=none /
/cache
/objectCache
--
--
Before posting questions to the
On Wednesday, July 23, 2014 10:36:43 AM UTC-7, Brian G wrote:
Does that just disable caching for the class emc?
If that is the case, you can replace it with:
objectCache
cache class=emc
provider=transfer.com.cache.provider.NoCacheProvider/
/objectCache
I think that's a valid config
James,
If you dump the argument in the working version of Transfer, what do you
see?
If you can dump the stack trace, we can look at the call before that and
see if it's not passing in a named argument? But generally I find that
these errors are unrelated to what they claim to be. I
Ok, I have it running. Somehow Sean's import from SVN did not include
JavaLoader:
https://github.com/seancoyne/transfer/tree/master/com/util/javaloader
The directory is empty and was missing the lib. I took the opportunity to
upgrade to JavaLoader 1.1 from 2012 and now have Transfer 1.3
I took a few minutes and copied the transfer docs into Github - thank god
they accept HTML so I didn't have to reformat them:
https://github.com/TransferORM/transfer/wiki
Brian
--
--
Before posting questions to the group please read:
I don't (yet) run Railo but this looks like a poor guess by Railo to
determine the datatype in the returned query when (I suspect) the first row
has a null. I see this behavior in QoQ where it guesses wrong and causes
problems. I'm surprised it would happen with an actual query.
If you want
John,
The above does not go into transfer.xsd but rather transfer.xml. If you
already had an objectCache stanza, you simply replace it with the link to
the ehcache file and associated caching directives instead.
Brian
On Friday, March 6, 2015 at 1:50:33 AM UTC-8, John Gerald wrote:
On Saturday, March 7, 2015 at 7:28:31 PM UTC-8, Mark Mandel wrote:
Awesome work mate!
I really need to write up a blog post on all the hard work you've been
doing.
http://www.2ddu.com/
Thanks Mark. On one hand, I appreciate it, but on the other hand, nobody
should use Transfer any
Brad,
No reason to wait - send those pull requests for the two you have fixes. :)
I presume that allows you to run the unit tests like I suggested in #13? I
don't have a ton of experience with the Transfer Unit Tests but Mark did a
good job of writing a boatload so I would start by seeing
You can pull the latest from Github: https://github.com/TransferORM/transfer
On Wednesday, January 13, 2016 at 6:32:49 AM UTC-8, Marc wrote:
>
> Hello,
>
> since my hosting provider (Hostek) changed it's shared host from CF10 to
> CF11 I receive an error in Transfer whenever I do a request:
>
Hi list,
I'm looking for individuals or consultants who have experience upgrading
Model-Glue/Coldspring/Transfer apps to ACF 2021 from older versions of CF.
I'd like to pay you for a bit of your time to understand how it went. I'm
especially curious if you're still running some/all of these
50 matches
Mail list logo
