Corrupted queries in ColdFusion MX 7
ColdFusionI hate to think that there is possibly a flaw in ColdFusion...it pains me to say it. But I think there might just be a problem.
One of our latest projects was an object oriented application (lots of objects stored in memory) that sees a pretty decent load. For example, one day when we noticed a problem we had in the neighborhood of 20k connections.
After getting everything as smoothed out as possible, we are still seeing some pretty strange exceptions coming through. And for the most part they seem to deal with invalid SQL syntax of some sort of another.
Our system has quite a bit of dynamic querying going on, but I would say probably significantly less actual queries to the DB server on an average request.
It seems, that around a heavy traffic period we begin to see these SQL errors coming through. So, we feel that it could very possibly be related to memory issues...but, being pretty shorthanded we have not been able to really dig as deep into the issue as we would like.
Some of the more common exceptions:
Message: Error Executing Database Query
Detail: ... Syntax error at token 0, line 0, offset 0.
OR
Message: Error Executing Database Query
Detail: ... Invalid parameter binding(s).
And a handful of actual corrupted SQL statements that make its way through.
Something like when querying a Users table... it might say
Message: Error Executing Database Query
Detail: ... Table 'Use' not found.
...basically truncating (maybe?) the SQL statement itself.
Initially the code was suspect, but we cannot find the source of the problem and so we hit Google for some commiseration.
Notice the last few comments on this blog post
And this forum thread on Adobe.com
Is anybody else experiencing this on their applications? The rumor is this issue is not present in Scorpio, but it would be nice if we could get a fix for MX7 if it truly is a CF bug.
***EDIT*** Link to the HotFix : http://kb.adobe.com/selfservice/viewContent.do?externalId=kb402002&sliceId=1
Steven Erat said:
Have you enabled database logging? Check the database logs to confirm the exact SQL received.
Martin said:
In my previous job, we ran into exactly the same problem you describe. We also noticed that other "wrapper" tags with generated content such as CFOUTPUT and CFSAVECONTENT triggered IllegalStateException errors. After digging into the java code we found that the exception is thrown when the generated content is somehow corrupted (there is comment in the code that states that this is inconceivable...).
Unfortunately, I have no solution to this problem. The kind of system you describe is very similar to the one I was working on and the errors did seem to occur under heavy load.
Unfortunately, I have no solution to this problem. The kind of system you describe is very similar to the one I was working on and the errors did seem to occur under heavy load.
Aaron Lynch said:
Martin could that message been something like:
"Out is not a NeoBodyContent!! Inconceivable!"
This is not good my friend.
"Out is not a NeoBodyContent!! Inconceivable!"
This is not good my friend.
Ruud Hermans said:
We're experiencing the 'Syntax error at token 0, line 0, offset 0' and the 'Invalid parameter binding' a couple of times a week on an application that has a few 100 users a day, running on CFMX 7.0.2. Unfortunately those errors are only a minor nuisance compared to another error we're seeing 3 - 7 times an hour: java.lang.IllegalStateException at coldfusion.tagext.lang.ModuleTag.doAfterBody
That one is similar in that it happens in CF code that is perfectly valid and in that I don't have a clue to what to do against it. :(
So far I can only hope someone at Adobe has a bright moment and some spare time to fix some things...
That one is similar in that it happens in CF code that is perfectly valid and in that I don't have a clue to what to do against it. :(
So far I can only hope someone at Adobe has a bright moment and some spare time to fix some things...
Martin said:
Yep. That was it. Found it in the OutTag class or something which is used by cfquery, cfoutput and cfsavecontent (and probably others) I reckon.
Steve Bryant said:
Yep. We are seeing the same thing. Happens to hard-coded queries and ones that use cfqueryparam. It only seems to happen during high traffic and with somewhat length query (over, say, 200 characters).
An examination of the error data shows that the SQL gets mungled somehow from what we have written in cfquery.
An examination of the error data shows that the SQL gets mungled somehow from what we have written in cfquery.
Aaron Lynch said:
Well, in a system with consistent high traffic and complex object relationships (LIKE MY APP!!!) this seems like it might be a pretty significant problem.
Adrock said:
Well, it sounds like a problem with the underlying DataDirect drivers. If the rumor is true (that it's fixed in Scorpio) then that could mean good news for CFMX 7 since it's probably due to new DataDirect drivers.
However, there are reports that there is a 5-8x increase in component creation. Considering your application has a lot of objects, you probably want to move to Scorpio right away.
However, there are reports that there is a 5-8x increase in component creation. Considering your application has a lot of objects, you probably want to move to Scorpio right away.
Martin said:
Adobe started taking notice in Feb this year in the forum thread you mentioned but I think people are focusing on the cfqueryparam aspect of this issue when they really should be shouting about the NeoBodyContent thing. I'm convinced it is the key to this issue. The invalid parameter binding errors are just a by-product of corrupted generated content in cfqueries.
Aaron Lynch said:
From what I could gather from the forum thread, Adobe didn't seem to really be 'taking notice'. Other than just an offhand 'please let me see your code' kind of response. (I really hope I am wrong about that by the way.)
Aaron Lynch said:
@adrock The upgrading to 8 sounds like a great idea...if only I controlled the purse-strings!
TJ said:
(repost, 1st one listed as spam)
We are running several high traffic sites, one getting 2m page views a day. We use coldSpring and cfc object caching and tons of other memory caching. The only time I have seen the "binding" error when the server was low on memory due to some abnormal traffic. Even then I only saw it once. We run multiple instances with 1g of memory max heap for each and we cycle each instance every night. We are also running the newest CF hotfixes and Jrun 4 updater 5. I belive the drivers were updated in one of those.
We are running several high traffic sites, one getting 2m page views a day. We use coldSpring and cfc object caching and tons of other memory caching. The only time I have seen the "binding" error when the server was low on memory due to some abnormal traffic. Even then I only saw it once. We run multiple instances with 1g of memory max heap for each and we cycle each instance every night. We are also running the newest CF hotfixes and Jrun 4 updater 5. I belive the drivers were updated in one of those.
Aaron Longnion said:
Hi -
Just curious: has anyone tried installing a "fresh" version of CF 7.01 (not 7.02), and then running the same apps on it to see if the problem goes away?
The reason I ask is that I've been part of teams that had apps with tens of thousands of hits per day for over a year on 7.01 which used CFC in the application scope (some in other scopes), where every param had a cfqueryparam on it.... and we never got those errors. (currently I'm working on CF 6.1 apps with similar architecture, and we don't get those errors).
I'm very curious to know if this helps anyone...
Just curious: has anyone tried installing a "fresh" version of CF 7.01 (not 7.02), and then running the same apps on it to see if the problem goes away?
The reason I ask is that I've been part of teams that had apps with tens of thousands of hits per day for over a year on 7.01 which used CFC in the application scope (some in other scopes), where every param had a cfqueryparam on it.... and we never got those errors. (currently I'm working on CF 6.1 apps with similar architecture, and we don't get those errors).
I'm very curious to know if this helps anyone...
Adrock said:
What versions of DataDirect?
I'd want to rule out drivers first... because a) they cause 99% of db problems and b) easy to resolve.
I'd want to rule out drivers first... because a) they cause 99% of db problems and b) easy to resolve.
Aaron Longnion said:
it seemed from reading the various forums entries that folks had already tried a number of different drivers without solving the issue. One possible constant I thought might be the CF version... I could be wrong, but worth a try. ;}
Aaron Lynch said:
I'm double checking on the driver situation now. I have a feeling it is whichever DataDirect driver that it came packaged with. I know we have used the JTDS driver on some other MSSQL connections.
Worth a shot I guess, but I have read similar things to what Aaron mentioned.
Worth a shot I guess, but I have read similar things to what Aaron mentioned.
Bruce said:
I also get the "invalid parameter bindings" error. It occurs randomly. At least I know I'm not alone. We're using CF 7.02.
Ed Ralph said:
Hello folks.
I've been pretty active on the adobe thread you've mentioned, and I've posted elsewhere, including on comments in Ben Forta's blog (your link above) and on the DataDirect forums. We've been trying to resolve these issues for about 6 months now.
I'll try to answer some of the questions posed above, but first I'll say what issues we are faced with at the moment...
All the issues started as soon as we started testing a refactored codebase on production. The new codebase uses application.cfc and generally made more use of cfc's (including those persisted in memory - session, server and application).
We use CFMX 7.02 on Windows 2003 Server with SQL Server 2005 (on separate boxes).
Issues:
1. Invalid parameter bindings.
First we suspected drivers - so tried all the datadirect jdbc drivers, and the Microsoft 2005 JDBC driver. This made no difference. We tried a few other things, like changing cfsql_types thinking we've got some datatype mismatch going on but no joy. Thinking it was a memory related thing we've played about with the JVM heap sizes to see if that alleviated the problem. We then decided to remove the cfqueryparam, thinking that it would eliminate the errors, albeit causing the performance to take a hit. Well, the invalid parameter binding errors didn't happen, but the SQL statement started to come through truncated - as various other people have noticed. Doh. The best solution we've found is that calling stored procs instead of using inline sql statements gets around both the invalid parameter bindings and the sql truncation issue. That's where we are with that one at the moment.
2. Illegal State Exception
This happens with code that calls custom tags or the cfsavecontent tag (which is just a custom tag) - and as somebody mentioned above these call the neobodycontent class. We've tried hard to work around this bug (we are certain both of these are bugs) by replacing custom tag calls with cfc functions. This works in the most part - but there are some uses of cfsavecontent that are just hard to work around.
Other things of note
- We've tried CFMX 7, CFMX 7.01 and CFMX 7.02
- Adobe did not manage to help us because we could not reproduce the error in an isolated test case.
- I've heard rumours also that CF8 doesn't suffer from these bugs
- I'm presently trying BlueDragon 7 to see if our code works on this server (scary)
Fun fun fun...
I've been pretty active on the adobe thread you've mentioned, and I've posted elsewhere, including on comments in Ben Forta's blog (your link above) and on the DataDirect forums. We've been trying to resolve these issues for about 6 months now.
I'll try to answer some of the questions posed above, but first I'll say what issues we are faced with at the moment...
All the issues started as soon as we started testing a refactored codebase on production. The new codebase uses application.cfc and generally made more use of cfc's (including those persisted in memory - session, server and application).
We use CFMX 7.02 on Windows 2003 Server with SQL Server 2005 (on separate boxes).
Issues:
1. Invalid parameter bindings.
First we suspected drivers - so tried all the datadirect jdbc drivers, and the Microsoft 2005 JDBC driver. This made no difference. We tried a few other things, like changing cfsql_types thinking we've got some datatype mismatch going on but no joy. Thinking it was a memory related thing we've played about with the JVM heap sizes to see if that alleviated the problem. We then decided to remove the cfqueryparam, thinking that it would eliminate the errors, albeit causing the performance to take a hit. Well, the invalid parameter binding errors didn't happen, but the SQL statement started to come through truncated - as various other people have noticed. Doh. The best solution we've found is that calling stored procs instead of using inline sql statements gets around both the invalid parameter bindings and the sql truncation issue. That's where we are with that one at the moment.
2. Illegal State Exception
This happens with code that calls custom tags or the cfsavecontent tag (which is just a custom tag) - and as somebody mentioned above these call the neobodycontent class. We've tried hard to work around this bug (we are certain both of these are bugs) by replacing custom tag calls with cfc functions. This works in the most part - but there are some uses of cfsavecontent that are just hard to work around.
Other things of note
- We've tried CFMX 7, CFMX 7.01 and CFMX 7.02
- Adobe did not manage to help us because we could not reproduce the error in an isolated test case.
- I've heard rumours also that CF8 doesn't suffer from these bugs
- I'm presently trying BlueDragon 7 to see if our code works on this server (scary)
Fun fun fun...
Ed Ralph said:
Hello folks.
I've been pretty active on the adobe thread you've mentioned, and I've posted elsewhere, including on comments in Ben Forta's blog (your link above) and on the DataDirect forums. We've been trying to resolve these issues for about 6 months now...
I've been pretty active on the adobe thread you've mentioned, and I've posted elsewhere, including on comments in Ben Forta's blog (your link above) and on the DataDirect forums. We've been trying to resolve these issues for about 6 months now...
Ben Nadel said:
I have seen the error "Syntax error at token 0, line 0, offset 0" before in queries that have a lot of logic in them (which sounds like your highly dynamic ones do). the above error means that there is NO sql statement within the CFQuery tag. This usually happens if you are building your SQL statement with a bunch of CFIF statements or something and NONE of then for one reason or another are true on a particular page request. In that rare case, you end up trying to execute a blank SQl statement.
Is it possible that that is what is going on? I am not sure why somethingl like this would happen in high traffic times?
of course, it might be like playing the numbers in dating... you ask out every girl that you see in hopes that a percentage of them, no matter how small, will say "Yes". this has nothing to do with your technique, but is merely a reflection of chance.
Now, to take that to your sitution, high traffic might not affect the CF code in any way at all. It might just be upping the chance or some error occuring because you are "playing the numbers".
Is it possible that that is what is going on? I am not sure why somethingl like this would happen in high traffic times?
of course, it might be like playing the numbers in dating... you ask out every girl that you see in hopes that a percentage of them, no matter how small, will say "Yes". this has nothing to do with your technique, but is merely a reflection of chance.
Now, to take that to your sitution, high traffic might not affect the CF code in any way at all. It might just be upping the chance or some error occuring because you are "playing the numbers".
Christopher Wigginton said:
I preach the cfqueryparam sermon and had something similar occur. On this particular application, which also had lots of objects in memory, I turned off maintain connections in the CF datasource. While this slowed down the app a wee bit, the problem did go away.
Steve Bryant said:
Ben,
I can't speak to Aaron's situation, but I can tell you that isn't possible in mine. I have gotten the error on SQL statements that, while somewhat long, are not very dynamic (no cfifs, for example).
Christopher,
Good suggestion. I will have to look into that. How much did it slow it down?
I can't speak to Aaron's situation, but I can tell you that isn't possible in mine. I have gotten the error on SQL statements that, while somewhat long, are not very dynamic (no cfifs, for example).
Christopher,
Good suggestion. I will have to look into that. How much did it slow it down?
Aaron Lynch said:
@Ben, I think that error you refer to is happening for me when ColdFusion is failing to pass the SQL correctly. But it is interesting to note that it can happen in other ways too.
patrick said:
I have been experiencing the same issue where queries are undergoing devastating corruption.
Was wondering what could stop:
UPDATE Users
SET Username="patsy"
WHERE id=433
from becoming: UPDATE Users SET Username="patsy"
on the SQL end?
Can I sue Macrobe?
Anyone seeking a recently unemployed CF developer?
Was wondering what could stop:
UPDATE Users
SET Username="patsy"
WHERE id=433
from becoming: UPDATE Users SET Username="patsy"
on the SQL end?
Can I sue Macrobe?
Anyone seeking a recently unemployed CF developer?
Adrian J. Moreno said:
My coworker Andrew ran into this just last night. His query has three cfqueryparams. He found that the value passed to the first was also being passed to the other two, even though they were each given a distinct getter from a bean.
Dumping the value of the getters before the query was run confirmed that each getter was retuning the correct value.
We have CFMX 7.0.2 multi-instance installed in all of our environments. We also recently updated everything to JVM 1.4.2_13 with the exception of our production servers. They still run version 1.4.2_12.
We downgraded the QA server to that version and the problem went away.
Has anyone checked their JVM?
Dumping the value of the getters before the query was run confirmed that each getter was retuning the correct value.
We have CFMX 7.0.2 multi-instance installed in all of our environments. We also recently updated everything to JVM 1.4.2_13 with the exception of our production servers. They still run version 1.4.2_12.
We downgraded the QA server to that version and the problem went away.
Has anyone checked their JVM?
Hemant said:
Recently we were able to reproduce IllegalStateException with help of one of our users. We are trying to fix this issue.
Thanks for your feedback.
Thanks for your feedback.
Aaron Lynch said:
Interestingly we had been experiencing some other behavior that had us a little baffled. What we found was that the group hired to manage our hardware, specifically the load balancing configuration, had failed to do so correctly. We basically were round-robin'ing people right out of their sessions (even though we requested a sticky session config).
Once we finally got the load balancing configured correctly, we have seen a drastic reduction in the number of these exceptions. This obviously would have effect the application(s) in multiple ways.
Basically we now have half the object instantiations, sessions, etc that we were seeing. Which leads to less memory consumption under the same load. Could this error start occurring under heavy(er) memory use? Does it have to do with objects down in the Java level? I wonder if by increasing maz heap size, tweaking the garbage collection settings, etc (jvm tuning) this "bug" could be fixed for those of us seeing the same exceptions.
Once we finally got the load balancing configured correctly, we have seen a drastic reduction in the number of these exceptions. This obviously would have effect the application(s) in multiple ways.
Basically we now have half the object instantiations, sessions, etc that we were seeing. Which leads to less memory consumption under the same load. Could this error start occurring under heavy(er) memory use? Does it have to do with objects down in the Java level? I wonder if by increasing maz heap size, tweaking the garbage collection settings, etc (jvm tuning) this "bug" could be fixed for those of us seeing the same exceptions.
Hemant said:
What we have found is that under some condition the session cleanup is doing bad things while module tag is running.
So your observation wrt sessions seems to match out intial findings. We see these exceptions coming up after and during session timeouts. We have not observed any corelation with high memory.
One of senior engineers, Ram noted following about this issue:
This is caused by following combination ...
1. application.cfc
2. App event methods in application.cfc like onSessionEnd etc.
3. Calling a method on a CFC from app event method that is stored in the app scope
4. CFML request executing at the same time when app event method is being called. And the cfml page has a custom tag
However, this has been a very difficult issue to reproduce. I will post a comment here if the fix becomes available so that you can verify and let us know if it works for you.
The HF are posted on the adobe site and can be searched via http://www.adobe.com/cfusion/knowledgebase/index.c...
Thanks,
Hemant
So your observation wrt sessions seems to match out intial findings. We see these exceptions coming up after and during session timeouts. We have not observed any corelation with high memory.
One of senior engineers, Ram noted following about this issue:
This is caused by following combination ...
1. application.cfc
2. App event methods in application.cfc like onSessionEnd etc.
3. Calling a method on a CFC from app event method that is stored in the app scope
4. CFML request executing at the same time when app event method is being called. And the cfml page has a custom tag
However, this has been a very difficult issue to reproduce. I will post a comment here if the fix becomes available so that you can verify and let us know if it works for you.
The HF are posted on the adobe site and can be searched via http://www.adobe.com/cfusion/knowledgebase/index.c...
Thanks,
Hemant
Ed Ralph said:
Interesting Hemant.
The combination you list is exactly consistent with our application - newly re-written with application.cfc, plenty of app persisted cfc's, cfm pages with cfmodule/custom tag calls all happening in the same request. We haven't been able to reproduce it, other than putting the codebase into production.
In reply to another comment about the JVM, we've tried many flavours of JVM and all suffer the same problem. Including 1.4.2_12.
The combination you list is exactly consistent with our application - newly re-written with application.cfc, plenty of app persisted cfc's, cfm pages with cfmodule/custom tag calls all happening in the same request. We haven't been able to reproduce it, other than putting the codebase into production.
In reply to another comment about the JVM, we've tried many flavours of JVM and all suffer the same problem. Including 1.4.2_12.
Mark Stanton said:
We've had a support call open with Adobe regarding the IllegalStateException issue for some time now (I might be the customer Hemant is talking about).
It seems that in the last couple of weeks they have now started making real progress. So I'd keep an eye out for a patch coming soon to adobe.com.
- Mark
It seems that in the last couple of weeks they have now started making real progress. So I'd keep an eye out for a patch coming soon to adobe.com.
- Mark
KenS said:
We have not published a hot fix for this yet. We will publish one shortly. However, if you email me direct I can send the hot fix for IllegalStateException.
Ed Ralph said:
Yes - I'd like the hotfix please! My email is ed[dot]ralph[at]abcam[dot]com... (I can tell you if it really fixes it :o)
Aaron Lynch said:
Ed, could you please let me know if that fixes your illegalStateException issues?
Thanks,
Aaron
Thanks,
Aaron
Ed Ralph said:
Sure. Would you mind obfuscating my email address to protect it from the email harvesters?
Thanks
Ed
Thanks
Ed
Ed Ralph said:
Issue solved.
Thanks Ken for the hotfix, I applied it to a production server last Friday and haven't seen an illegalStateException or Invalid Parameter Binding(s) error since then.
We've been bashing away at this for the best part of a year - coming across many folks that are having the same issue. I notice that the hotfix hasn't been officially released on the Adobe website - but I'm happy to keep this one in production since the servers are significantly more stable now.
Thanks
Ed
Thanks Ken for the hotfix, I applied it to a production server last Friday and haven't seen an illegalStateException or Invalid Parameter Binding(s) error since then.
We've been bashing away at this for the best part of a year - coming across many folks that are having the same issue. I notice that the hotfix hasn't been officially released on the Adobe website - but I'm happy to keep this one in production since the servers are significantly more stable now.
Thanks
Ed
Aaron Lynch said:
Wow, ok...time to try the fix on our stuff. Thanks for letting us know your results Ed.
Ed Ralph said:
No problem. It is a huge relief here that we can now progress with a server platform that actually works as it is supposed to (until we find the next bug?).
It is worth noting here that during this time I had the opportunity to evaluate both BlueDragon 7 and CFMX8 (8 only in the last week or so). MX8 is still too buggy in beta 2 to do any good comparisons, but I was impressed with BD 7 and the new features, and also in how much support I got from New Atlanta. Ok so they clearly want us to migrate from CFMX to BD, but if it is an indication of the level of support you can expect from them then it is a ton more than you get from Adobe. I'll definately be evaluating BD7.NET when we upgrade from CFMX7...
It is worth noting here that during this time I had the opportunity to evaluate both BlueDragon 7 and CFMX8 (8 only in the last week or so). MX8 is still too buggy in beta 2 to do any good comparisons, but I was impressed with BD 7 and the new features, and also in how much support I got from New Atlanta. Ok so they clearly want us to migrate from CFMX to BD, but if it is an indication of the level of support you can expect from them then it is a ton more than you get from Adobe. I'll definately be evaluating BD7.NET when we upgrade from CFMX7...
Ken Smith said:
Nice to hear that our good work has paid off???
Ken Smith
CF/JRun Server Escalation Team
Adobe Systems
Ken Smith
CF/JRun Server Escalation Team
Adobe Systems
Ruud Hermans said:
Many thanks from me as well. Haven't seen a single error since I applied the hotfix and rebooted the server.
Btw, doesn't 'hotfix' mean you don't have to restart anything? Because I did got the errors after applying the hotfix at first. Only after the reboot the errors went away. (I wanted to just restart CF, but a well timed 2 hour power failure in our area did effectively reboot everything :) )
Btw, doesn't 'hotfix' mean you don't have to restart anything? Because I did got the errors after applying the hotfix at first. Only after the reboot the errors went away. (I wanted to just restart CF, but a well timed 2 hour power failure in our area did effectively reboot everything :) )
Ed Ralph said:
You have to restart the CF service so that it picks up the .jar file. You don't have to run an installer or anything so I guess really it's a luke-warm fix.
adrocknaphobia said:
@Ed: "MX8 is still too buggy"
That is surprising to hear. Early alphas of the release have been powering parts of Adobe.com without error (CFMX7 code) and the beta testers are saying the same.
How do you compare BD7 vs CF8? Doesn't really seem fair to New Atlanta. You should probably compare BD7 to CFMX7 (the one that came out 2.5 years ago). Functionally those version are somewhat the same (although migrating to BD usually requires a significant re-write of the application).
Either way, kudos to Adobe support for releasing a hot fix and responding the community.
That is surprising to hear. Early alphas of the release have been powering parts of Adobe.com without error (CFMX7 code) and the beta testers are saying the same.
How do you compare BD7 vs CF8? Doesn't really seem fair to New Atlanta. You should probably compare BD7 to CFMX7 (the one that came out 2.5 years ago). Functionally those version are somewhat the same (although migrating to BD usually requires a significant re-write of the application).
Either way, kudos to Adobe support for releasing a hot fix and responding the community.
Ed Ralph said:
- my buggy comment - There is one particular bug that has been resolved in MX8 RC1 (not out yet) that is a show-stopper for our applications - so I probably should have said 'too buggy for us'.
- comparing MX8 and BD7. You are right, comparing the two doesn't seem fair to New Atlanta who are usually playing catch up. However, BD7 were first with CFTHREAD. From a customers point of view (me), if I upgrate from CFMX7 it will either be to MX8 or to BD7. So I will have to evaluate the two side-by-side. You could argue that it is a sideways move to BD7 from CFMX7, but I'd be spending money that could otherwise be spent on MX8 - so it's an upgrade in my eyes. We'd only do MX7 to BD7 if it offered more - which from what I've seen it probably does...
- comparing MX8 and BD7. You are right, comparing the two doesn't seem fair to New Atlanta who are usually playing catch up. However, BD7 were first with CFTHREAD. From a customers point of view (me), if I upgrate from CFMX7 it will either be to MX8 or to BD7. So I will have to evaluate the two side-by-side. You could argue that it is a sideways move to BD7 from CFMX7, but I'd be spending money that could otherwise be spent on MX8 - so it's an upgrade in my eyes. We'd only do MX7 to BD7 if it offered more - which from what I've seen it probably does...
adrocknaphobia said:
@Ed
I'm emailing you offline. I'd like to hear more about the bug ur seeing in CF8 and see what we can do to fix it.
I'm emailing you offline. I'd like to hear more about the bug ur seeing in CF8 and see what we can do to fix it.
Ed Ralph said:
adrocknaphobia - I entered the bug in the bugtracker and it is fixed in RC1 - so there is no need to do anything on that front. But as we are only on public beta 2, I cannot do any more testing on CF8...
kens said:
The technote, http://www.adobe.com/go/kb401830">ColdFusion MX 7.0.2: Hot fix for IllegalStateExceptions in ModuleTag is now published on adobe.com and includes the hot fix.
Ed Ralph said:
Thanks Ken. Is the published hf identical to the one you emailed to me? i.e. should I replace my current hf with this one?
Thanks
Thanks
Ken Smith said:
The technote contains the same hot fix jar I have distributed directly the past week or 2. No need to update if you already got it from me.
Aaron Lynch said:
Ken, I this comment on another post:
"Where do we get info on new technotes/hotfixes these days? It's not listed here http://www.adobe.com/go/tn_17883 and I can't find a "recent technote" list anywhere in the CF support site either."
"Where do we get info on new technotes/hotfixes these days? It's not listed here http://www.adobe.com/go/tn_17883 and I can't find a "recent technote" list anywhere in the CF support site either."
Aaron Lynch said:
Rich, I just last week discovered this same issue. I haven't had time to chase down where it went or why it disappeared.
Rich Rein said:
It looks like it now lives at http://kb.adobe.com/selfservice/viewContent.do?ext...
Unfortunately, it still bugs me that google searches turned up the dead link, and that the CF hot fix page has no mention of it.
Unfortunately, it still bugs me that google searches turned up the dead link, and that the CF hot fix page has no mention of it.
Aaron Lynch said:
Well, thanks for the updated link Rich. I will update this in the content of the blog entry.
Ken Smith said:
Sorry for the confusion. Adobe just updated its CMS for publishing technotes. So, some recent technotes are published under new IDs. This only effects the few technotes first published during the change over but before the new system went live. The new shortcut URL is http://www.adobe.com/go/kb402002
Vince Bonfanti said:
@Ed, re: "comparing MX8 and BD7. You are right, comparing the two doesn't seem fair to New Atlanta who are usually playing catch up. However, BD7 were first with CFTHREAD."
Actually, not only CFTHREAD, but nearly all of the "new" features in CF8 are already in BD 7.0 or earlier releases, some as long as 5 years ago. See this blog entry for more info:
http://blog.newatlanta.com/index.cfm?mode=entry&en...
Also, note that we offer discounted upgrades from CF to BD; contact me privately if you need more details.
Actually, not only CFTHREAD, but nearly all of the "new" features in CF8 are already in BD 7.0 or earlier releases, some as long as 5 years ago. See this blog entry for more info:
http://blog.newatlanta.com/index.cfm?mode=entry&en...
Also, note that we offer discounted upgrades from CF to BD; contact me privately if you need more details.
Adrock said:
He we go again.... this time with comment spam.
Sadly, that list isn't even a small fraction of the new features in CF8. I wonder if that secret 'discounted upgrade' helps to offset the cost of redeveloping a CF application to run on the knock-off platform.
But I digress.
Don't settle for cheap imitations.
Viva ColdFusion 8!
http://www.adobe.com/products/coldfusion/
Sadly, that list isn't even a small fraction of the new features in CF8. I wonder if that secret 'discounted upgrade' helps to offset the cost of redeveloping a CF application to run on the knock-off platform.
But I digress.
Don't settle for cheap imitations.
Viva ColdFusion 8!
http://www.adobe.com/products/coldfusion/
Vince Bonfanti said:
The discounted upgrade program isn't a secret, it's posted on our web site. And Ed's already discovered--as many others have--how easy it is to port existing CFML applications to BlueDragon. BlueDragon costs less, has all the features you want, is more stable and reliable, performs better, and you get better tech support--what more could you ask for?!
One other thing I should mention for Ed--and other UK customers--watch for the launch of the New Atlanta UK web site in the next few weeks. I think you'll be pleasantly surprised at our UK pricing. Also note that we'll soon begin offering pre- and post-sales technical support from our London office.
Cheers.
One other thing I should mention for Ed--and other UK customers--watch for the launch of the New Atlanta UK web site in the next few weeks. I think you'll be pleasantly surprised at our UK pricing. Also note that we'll soon begin offering pre- and post-sales technical support from our London office.
Cheers.
Ed Ralph said:
Hello folks - I didn't mean to kick off a BlueDragon VS CFMX debate here - but I'd like to add my view here.
We've been using CF since at least version 4.5. We upgrade a short while after a new version comes out - we're on 7.02 right now. There are a dozen or so servers running it. In short, we've a lot of experience with CF. When CFMX became unstable for us we had to appraise BD. I was pleasantly surprised, and found their platform to be a decent alternative. As Vince said, it wasn't that hard to port to BD - and we have over a million lines of CF code.
There is definately a market for CFMX competitors, if these guys didn't exist then there would be far less pressure on Adobe to bring out the features in MX that all the developers want. I'm all for a healthy market with competitors pushing each other to produce top server applications.
We've been using CF since at least version 4.5. We upgrade a short while after a new version comes out - we're on 7.02 right now. There are a dozen or so servers running it. In short, we've a lot of experience with CF. When CFMX became unstable for us we had to appraise BD. I was pleasantly surprised, and found their platform to be a decent alternative. As Vince said, it wasn't that hard to port to BD - and we have over a million lines of CF code.
There is definately a market for CFMX competitors, if these guys didn't exist then there would be far less pressure on Adobe to bring out the features in MX that all the developers want. I'm all for a healthy market with competitors pushing each other to produce top server applications.
Loading....