editor corruption

Hi Julian,

Please call tech support so that this issue can be dealt with properly.

I logged a bug like this ages ago but to do with CTRL-C and CTRL-V causing it and then typing FOR would print ROF (backwards). The only problem is it did it on my system but no one else could replicate it

I still get it - normally when I do a cut / copy / paste. Hmm. may look into that - thanks for the info

Yep - its always do do with cut/copy/paste

Clearly a punishment for being unoriginal!

heh.

I now get this on a regular (and almost reproducible) basis. It always involves a cut/copy/paste, and seems to happen when the code you are editing does not compile properly.

If you close the file, and reopen it, you can edit for another little while until the problem reoccurs.

Once the file compiles ok, there is no problem

ok, a further update. This corruption thingy was driving me mad today, and my little trick above had no effect.

In desperation I updated to Java2 (jre 1.6), and removed the -vm from the OEA startup script (forcing it to use the system default jre) and since then I have had zero problems .... perhaps this may help someone further down the line.

I'm having the same problem. I already had 1.6 installed on my box so I removed the -vm "C:\dlc\101b\jre\bin\javaw.exe" reference from my startup, but the issue came back after a few minutes. I'm assuming it didn't do anything and I was just extra careful because I hoped it was fixed. And I was just getting more productive in the new environment then I was in the old one.

I logged it with tech support, so maybe if enough people encounter it Progress will find the problem. I might try researching some of the Eclipse message boards too.

Thanks god I'm not the only "user" here

Unfortunately, I am still getting the error, but I am starting to learn to live with it. Arrrgghh.

oh man. Just installed 10.1B SP1 and it's gotten a whole heap worse. It's impossible to type without getting the exceptions. Dammit.

Julian,

Are you using a multi-core CPU?

I think so ... pentium D 3.2GHz .. shows 2 cpus in task manager

this has been driving me absolutely mad today

I'm using two hyperthreaded processors (total of 4 in TM). Progress told me they are just starting their research on the subject. I've given them the link to this thread.

Don.

Ok. I've turned off the dual-cpu option in the bios.

Taadaaa no more corruption after more than an hour. Previously I could not edit a file for more than 1 minute before I started hitting these problems.

perhaps it is a multi-cpu problem ..

Engineering is looking into this issue. Please stay tuned and if we determine a fix for this issue we will let you know. Thanks and sorry for the problems this is causing.

Mark Hanley

that's great. FWIW, I'm running 10.1B with SP1.

Did you post this back on PEG? Coincidentaly having exactly the same problem. All starts with Ctrl-C.

I think that I did a while back (it's been happening for quite some time). I will again ..

I have seen this as well. Also, when the editor starts typing "backwards", hitting the delete key will result in an "out of bounds" exception.

Not related to dual CPU. I get it and I am running a single CPU machine at work.

damn. thought that we had solved that one. Mind you, still have not had the error since I went single CPU

Note that some single CPU machines are capable of hyperthreading, if supported by the OS and enabled in the bios, so they behave like dual CPU machines. If you open task manager, are you only seeing one CPU display?

1180 posts ? You have waaaaaaaaaaaaaaay too much time on your hands

I wonder what the number would be for PEG...

I wonder what the number would be for PEG...

Probably less than mine.

Maybe, maybe not ... there's a lot of years in there and I have been more or less active during various periods. It would be fun to know, though.

For that which is currently searchable ... which doesn't include all the lists, by any means, nor is the time depth go back anything like the full time period, I come up with 4822 for you on the main PEG list and 4585 for me, but this is the instance count of the PEG search mechanism, which doesn't necessarily translate to actual posts. E.g. if I search on timk instead of Kuehn, I get 944 while cintegrity.com yields 3788, but the latter might be counting each post twice, once from the e-mail address and once from the .sig.

Not to take this threadjack too much further, but I've posted from a number of different accounts, so timk@.... wouldn't be a complete inventory of my posts by any means.

It's interesting none-the-less.

Back to editor corruption.....

I have seen it too, and to me it is clearly related to speed:

OpenEdge Architect 10.1B is installed on my PC at the office. When I am in

the office working on my PC, there is no problem at all.

But when I am home and open a VPN tunnel and start "Remote Desktop

Conenction" to use OEA on my office PC, then I get the errors you have

reported. These errors occur very often during office hours when the

infrastructure is busy, and not often (or not at all) at night when nobody

else is using the office systems.

When I am using Remote Desktop Connection and go to Preferences|Editor to

disable all the fancy stuff, like the Assistance options, then I don't get

the problem as quickly (but will still get it after a while).

So, my conclusion is that as you are entering text, the editor will try to

analayze immediately what you are entering to see if it can assist with

anything like auto-completion or syntax-coloring or this tooltip-like help,

and gets frustrated when keyboard events still come in while it is

calculating.

Jurjen Dijkstra

Anything is possible I guess. However, I'm having the problem on a machine that is really fast, quick RAID10, dual processors (2 hyperthreaded processors). I've not had the problem at work on a dual processor machine (no hyperthreaded) that is really slow. I've disabled all the "fancy" lookups and I hope that will resolve the issue until Progress figures out what is going on. I've never had the problem using Eclipse in the past and my Java friends say it is rock solid, so I'm assuming its something Progress did.

Yep, last night I had it on my local machine too.

I was trying to insert new lines in a file (prolint.p) that was already 1060 lines long. Other files did not give the problem. Then I opened prolint.p in the "text editor" in Eclipse, the non-intelligent black and white editor, and could edit without problems.

still not had any problem since I turned the dual core off. /me now hopes he hasn't tempted fate ...

I guess I need to try changing my bios to use only one CPU, but with such a global setting it just kills me to not use the full potential of my box for all the other things I do. Has anyone with the corruption issue evaluated the CPU issue?

not verified, but I can confirm that I was also have the same problems on my dual-core portable. Turned it off, not a problem since. Exactly the same issue was solved by doing the same on the desktop.

I agree, it's a real *** if the solution is to turn off the extra cpus - but, perhaps it also means that there is a pointer to the problem so that it can get solved faster

oh, and methinks that this is of sufficient seriousness to warrant a hot fix, not to have to wait until Aug / Sep for SP02 ....

after changing my portable to remove the dual-core option I have had no errors at all.

I'd love to report the same thing. I made a change in the BIOS to turn off Hyperthreading, but I was not able to disable the second physical CPU in the BIOS. Instead of 4 virtual CPUs I now have two in TM. I used Architect for about 30 minutes or so before I got the same old error. They are not as frequent, but they happen when I do copies. I'm not going to physically remove the second CPU, but I might play with the OS to see if I can disable its use.

Message was edited by:

Don Beattie

From Progress Technical Support:

This is just an update.

I logged a new bug # OE00150179, it is high priority and Development is working on it.

I think you know there is no workaround to the issue, but it can be reduced by disabling secondary processors.

Don.

at least that's a start, and Tech support have now found it

If you get a hotfix, let us all know - it seems to be a widespread issue.

Again, fwiw, disabling the second procesor has stopped the errors on my desktop and laptop

If disabling the second core/processor fixes the problem, have you tried setting affinity on the Eclipse process to see if that helps? I don't have access to my desktop, but if I remember correctly you can do this using the task manager (right click the process and there is a menu option to set affinity) or hunt on the net for the affinity.exe program which you add to the shortcut of Architect..

works for me (so far!) Many thanks

I loaded SP1 for 10.1B yesterday, and it definitely made the editor corruption problem worse. Before SP1, it only seemed to happen occasionally. After SP1, nearly anything I edited would have the problem.

The 'set affinity' trick seems have resolved the problem on my machine too!

The only other odd problem I have now, is that some of the keystroke shortcuts don't work until you first access the menu item. For example, CTRL-SHIFT-C will not work until I do a syntax check from the context menu with the mouse. Thereafter, CTRL-SHIFT-C works just fine.

I had the shortcut issue before I had the editor issue. Probably not related. If I switched perspectives it usually started working again.

The hot patch is on its way. I'm going to beta test it tomorrow. I'll let you know how it goes.

Beta Testing Priority Unofficial Patch

I turned hyper-threading back on (4 virtual processors), installed 10.1B01, and applied the PUP supplied by Progress Technical Support. I've not had any problems after a few hours of testing. I've reported the results to support so I'm assuming the PUP will be available to everyone that needs it soon.

that's great news. let us know !

You will have to call support to get the hot patch, I don't think they are going to post in the downloads.

are you serious ? For a problem like this ? Goddam ! Do you have a hotfix reference number that I could quote them ?

Thanks for the help

AFAIK, hot fixes are never posted on ESD..

#OE00150179 is the bug number that they gave me. That might make it easier for the tech person to give you the code. I'm not sure I have the final hot patch, they just asked me to beta test it, which I did. I've not heard back from the tech person about closing the ticket...

The patch you have been provided is an Unofficial Patch (previously referred to as a PUP) specifically for you to test with to confirm whether it does indeed fix the problem.

The Unofficial Patch will not be made available to others unless there are some very special circumstances.

Once the UP has been confirmed to have fixed the problem, Progress creates a Hotfix (previously referred to as a TempFix).

Once a Hotfix has been created, customers hitting the problem are able to contact Tech Support to obtain the Hotfix (on a customer by customer basis). i.e. Hotfixes are usually provided to directly to those customers requesting the fix, not through a public access method like the download center.

What I can tell you is that the Hotfix has been requested by support but has not been created by development yet, according to the bug details. Once created, hotfixes then need to go through a QA process (unlike an Unofficial Patch).

Given that many customers are hitting this bug, Development / Product Management might decide to distribute it differently, however it should not be assumed that this will happen.

If you want to be certain that you will receive the fix, log a support call and say that you are hitting this bug and that you need the Hotfix once it becomes available.

I hope you've found this update useful.

Regards,

Dave Smith

PSC Tech Support

Hot Patch 101B0103 is available to fix this problem. You probably have to contact tech support to get it.

Wondering, is lack of the mouse-wheel working in the editor part of the corruption? I swear it was working fine once before, now the mousewheel will not scroll.

Also, sometimes referenced includes will open on double-click and sometimes not?

Anyone else have any experience with these problems?

FYI. I've spoken with the Product Manager via email, and he agrees that the hotfix should be made publicly available.

I'll update the thread again once that is done.

If anybody requires the hotfix related to this issue prior to the 10.1B02 service pack release, they will need to log a call with Progress Technical Support.

Note that the issue fixes a specific bug related to Architect Corruption, as per Knowledge Center solution P123696.

If you find that the hotfix does not fix your issue, please log another support request as it may be a different issue.

Regards,

Dave.