Re: Source code changes un-recoverable after a crash

On Feb 18, 2017, at 4:35 PM, Ken Killian <kkillian@xxxxxxxxxxxx> wrote:

Jon,

Good Points! It is easy to forget that we are used to the rock solidness of IBM i running on the IBM Power System!

Maybe up to 99.5% of time RDI is pretty SOLID. But, when it does fail, it is at the worst time. No time is a good time for a failure. <LOL> And not getting helpful diagnostic messages causes much aggravation.

I was trying to do a fast open a member on another box, and I get a Java Null Pointer error. And then I think to myself, "Dummy, you didn't connect to that system yet..."
So, that is one example where the software should say: "That connections is not active, do you wish to connect now?" Wouldn't that be great! <Hint Hint>

When setting up newbie on RDI, the FIRST thing I do is set to classic mode, and set the Active window border to their Favorite color! And visually, the newbie has a clue to what window is Active! Plus they have ownership, because the outline is in their Favorite color!

After showing them how to compile. They get the Error listing gets the focus. And I have them go to the "compile" menu option and say try to compile it again. They notice it is grey-out. And I ask them with a smirk, "How do you compile an Error-Listing?" I then point out that their source member has to have the focus. And then have them try the compile again. And it is no longer Greyed-out. And they understand the "visual clue" of the active window highlighted in their favorite color!

So, yes, I was looking over the user's shoulder. And verified that the source member had the focus, and we got some sort of garbage worthless error messages.

Long-Story short: Save his member to a PC-FILE. We closed the source member WITHOUT saving. Since we got Java errors when trying to save it. <Big Frown>

We restarted RDI, and it "Magically" started to work again! That is why we call it "Restart-to-work-Software". When it fails, try Restarting it! I guess that is how client software is supposed to work.

I must admit this did NOT happen on my PC. And I haven't had that error in a very long on my laptop. Hence the 99.5% reliability of RDI. When it screws up, it screws up! So, just restart it, to "get-it-to-work-again". And if that fails, clear the cache and restart. And if that fails, then create a new workspace. Eventually, the restart will get it to work Again!

And another newbie with RDi 9.5.1.1 with a brand-new (First day on his new laptop) SCREAMING NEW Laptop i7 16-gig and SSD-HHD. Code Assist did not work, even after the outline was built. I said oh well, sometimes it does NOT work. This Developer was working on older Fixed format code. I said that happens lot with RDI. But, when code assist does work, it is pretty cool!

Note, I didn't have time to tell him about the restart, I had to rush off to meeting.

-Ken Killian-


-----Original Message-----
From: WDSCI-L [mailto:wdsci-l-bounces@xxxxxxxxxxxx] On Behalf Of Jon Paris
Sent: Friday, February 17, 2017 6:23 PM
To: Wdsci-L <wdsci-l@xxxxxxxxxxxx>
Subject: Re: [WDSCI-L] Source code changes un-recoverable after a crash

RDi is a client-server app. It can never be as 100% as SEU where you are directly interacting with the system all the time.

It is also difficult to completely bullet proof such as system due to the differences in settings etc. on the host. For example I had a lot of dropped connections and save problems on a client system. Turned out their firewall timed out a connection after a relatively short inactive period. When that happened there is nothing to notify the client or the server that the connection has gone. RDi will notice when you try to save, will create a new job but then fail to get a lock as the original job is still active. Normally it handles this reasonably well by offering the chance to kill the old job and away you go.

So saying to IBM - "make it as reliable as SEU" just ain't gonna happen - there are way too many variables, many of which (time outs for example) are beyond IBM's control as they are user controlled and/or result from VPNs and firewalls.

Point I am trying to make here is that we need to make practical suggestions to IBM as to how things could be handled better. That's how the design flaw in the debugger was eventually resolved. For example - after a dropped connection, rather than just create a new one, RDi could perhaps talk to a central job that "knows" what job this PC was last connected to and then try to re-make that connection. I'm not sufficiently comms savvy to know how/if that could be managed but I do know the current system is problematic.

By the way - I am assuming in all this that the user had made sure that the edit window was active when the problem arose? As I mentioned earlier this has not happened to me for a very long time but when it did (beyond the workspace recovery method discussed earlier) I did find one approach that worked consistently. Go to the connection and disconnect. Then from the same spot do a connect. If save is still not available, create a new temporary source member and copy/paste.


Jon Paris

www.partner400.com
www.SystemiDeveloper.com

On Feb 16, 2017, at 8:35 AM, Ken Killian <kkillian@xxxxxxxxxxxx> wrote:

Mark,

Just like me, this developer boots up his PC every day, no hibernation. His PC did not crash.

Sorry for the confusion, his changes in the original member were lost. But, we copied the source changes from the work-space.

We had to "re-apply" his changes. Which was aggravating and a time waster. So much for time productivity with RDI. Why wouldn't a "SAVE" work? It works in Notepad, notepad++, Visio, word, excel, Visual Studio, etc. This should be 100% ROCK SOLID PERIOD!

With my experience, I could have update the code myself in under 5-minutes. If we didn't care about the source change dates, we could have recovered in Under 30-seconds.

So, I guess I will have to give RDI Recovering instructions to all the developers, saying how to recover your source when RDI does NOT allow to SAVE your source changes.

And ultimately, I will get, we never have to worry about this with SEU. Why isn't RDI Stable/Reliable? When are we going to get a "stable" version of RDI?

And my typical reply will be, you will have to ask IBM, I didn't write this code. I only use it, it is good about 99% of time, but glitches pop-up.

Restart RDI, and it usually clears up the issue. If not, try clearing your cache and restart. And if that fails, then create a new workspace and try again. It is what I have to do.

-Ken Killian-

-----Original Message-----
From: WDSCI-L [mailto:wdsci-l-bounces@xxxxxxxxxxxx] On Behalf Of Mark Murphy/STAR BASE Consulting Inc.
Sent: Thursday, February 16, 2017 7:24 AM
To: Rational Developer for IBM i / Websphere Development Studio Client
for System i & iSeries <wdsci-l@xxxxxxxxxxxx>
Subject: Re: [WDSCI-L] Source code changes un-recoverable after a
crash

I was going to say something like that. Seems to me that PC crashes don't loose anything anymore, or very little because RDi notices the updated local copy and asks to use that. I can say no, don't use it, however, and then the changes are lost.

Mark Murphy
Atlas Data Systems
mmurphy@xxxxxxxxxxxxxxx


-----Jon Paris <jon.paris@xxxxxxxxxxxxxx> wrote: -----
To: Wdsci-L <wdsci-l@xxxxxxxxxxxx>
From: Jon Paris <jon.paris@xxxxxxxxxxxxxx>
Date: 02/14/2017 12:57PM
Subject: Re: [WDSCI-L] Source code changes un-recoverable after a
crash


"so all his changes were lost. "

Perhaps not. The workspace is constantly being saved on the PC. The actual interval being determined by the save interval which by default is 5 minutes.

In case like this in the past I have been able to recover the updated source BUT you have to be aware of the possibility _before_ opening the source from the host again. I don't know why RDi does not always "notice" that an updated local copy exists (but it hasn't happen to me in a very long time now) and prompt you. But in cases where you know that there is something bad going on it can be worthwhile to dig into the workspace and take a copy of the locally saved source file. Then you can use that as the source for updating once you re-open.

Also, I have found that sometimes Save As will work in cases where Save does not.


Jon Paris

www.partner400.com
www.SystemiDeveloper.com

On Feb 14, 2017, at 11:56 AM, Stephen M. West <stephen.west@xxxxxxxxxxxxxxxxxxxxxxx> wrote:

Ken,

I have experienced the issue where RDi will not save (Ctrl-S). I
manually copy/paste my source from RDi into Notepad, re-start RDi,
then copy the changes back from Notepad.

On another occasion, one of the developers here spent several hours
making changes to a source member, when his PC crashed un-expectedly.
Who knows why it crashed; these things happen. However, he did not
do a <ctrl-s> to save his source, so all his changes were lost. He
did not get the normal prompt to recover his changes at RDi startup.
Never would have happened in SEU, because the changes are always
retained, even when the green-screen app fails. I submitted an RFE
to IBM requesting that RDi be changed to save source to the IBM Power
i at regular intervals. It can be viewed by the public at:
http://www.ibm.com/developerworks/rfe/execute?use_case=viewRfe&CR_ID=
71090

The RFE was "Declined". Here is the help text behind "Declined":

"Declined: IBM has evaluated the request, and has determined that it
can not be implemented at this time or does not align with the
current multi-year strategy. This request may be resubmitted for
consideration after 12 months from the date of decline. The request
submitter will automatically be notified by email when the request
qualifies for reconsideration and resubmission."

Kinda felt like I was told "shut up and sit down" after requesting
simple functionality that should have been in RDi from the beginning.

Since I never received an email stating that it "qualifies for
reconsideration and resubmission", I presume it was a hard "NO!!!".

b.t.w. - Thanks for your persistence in pushing for improvements to RDi.
It has some awesome features that blow SEU away. You are not the
only programmer having issues with RDi. Please keep up your efforts
on behalf of us programmers who gave up after our efforts (it takes
time to craft an
RFE) were shut down with a simple "Declined". You are our
[un-official] advocate for RDi improvements.

Stephen West
Lead Software Engineer
Stillwater Insurance Group
12500 I Street, Suite 100 Omaha, NE 68137
W: (402)599-8325 | C(402)659-0452
Stephen.West@xxxxxxxxxxxxxxxxxxxxxxx

-----Original Message-----
On Tue, Feb 14, 2017 at 10:12 AM, Ken Killian <kkillian@xxxxxxxxxxxx>
wrote:

Mark,

I can understand that...

At work, I am called "Mr. RDI", since I promote RDI.

This morning, my co-worker just called me over to his Cube....

He was UNABLE to save his source changes in RDi 9.5.1.1, and he was
perplexed. It gave some cryptic Java Error, but unable to save his

source.

This is a BIG ISSUE!!!

He changed the code, and NEEDS to be able to save his CHANGES!

So, we manually copied his PC-File over to another file. Close his
source WITHOUT CHANGES; since we were UNABLE to save changes.

And manually re-open the member with NO CHANGES....

Now, he is forced to MANUALLY re-apply his changes, looking another
PC file. Where RDI was unable to allow him to save his changes!!!
WTF. It is NOT JUST ME!

Saving a change source file "should NEVER FAIL" period! Not even ONCE!

He wants to use RDI, but with issue, why not just use SEU, where you
KNOW it will SAVE YOUR CHANGES??????

PS. Once I get the error message from his Laptop, I will post it on
this message.



-Ken Killian-
Mobile Comm Back-End
Senior IBM i Developer using Rational Developer for I RDI 9.5.1.1
Running IBM I, that supports: Java, PHP, Ruby/Ruby on Rails,
Python, Node JS, RPGLE

--
This is the Rational Developer for IBM i / Websphere Development
Studio Client for System i & iSeries (WDSCI-L) mailing list To post
a message email: WDSCI-L@xxxxxxxxxxxx To subscribe, unsubscribe, or
change list options,
visit: http://lists.midrange.com/mailman/listinfo/wdsci-l
or email: WDSCI-L-request@xxxxxxxxxxxx Before posting, please take a
moment to review the archives at http://archive.midrange.com/wdsci-l.

--
This is the Rational Developer for IBM i / Websphere Development
Studio Client for System i & iSeries (WDSCI-L) mailing list To post a
message email: WDSCI-L@xxxxxxxxxxxx To subscribe, unsubscribe, or
change list options,
visit: http://lists.midrange.com/mailman/listinfo/wdsci-l
or email: WDSCI-L-request@xxxxxxxxxxxx Before posting, please take a
moment to review the archives at http://archive.midrange.com/wdsci-l.
--
This is the Rational Developer for IBM i / Websphere Development
Studio Client for System i & iSeries (WDSCI-L) mailing list To post a
message email: WDSCI-L@xxxxxxxxxxxx To subscribe, unsubscribe, or
change list options,
visit: http://lists.midrange.com/mailman/listinfo/wdsci-l
or email: WDSCI-L-request@xxxxxxxxxxxx Before posting, please take a
moment to review the archives at http://archive.midrange.com/wdsci-l.
--
This is the Rational Developer for IBM i / Websphere Development
Studio Client for System i & iSeries (WDSCI-L) mailing list To post a
message email: WDSCI-L@xxxxxxxxxxxx To subscribe, unsubscribe, or
change list options,
visit: http://lists.midrange.com/mailman/listinfo/wdsci-l
or email: WDSCI-L-request@xxxxxxxxxxxx Before posting, please take a
moment to review the archives at http://archive.midrange.com/wdsci-l.

--
This is the Rational Developer for IBM i / Websphere Development Studio Client for System i & iSeries (WDSCI-L) mailing list To post a message email: WDSCI-L@xxxxxxxxxxxx To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/wdsci-l
or email: WDSCI-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives at http://archive.midrange.com/wdsci-l.
--
This is the Rational Developer for IBM i / Websphere Development Studio Client for System i & iSeries (WDSCI-L) mailing list
To post a message email: WDSCI-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/wdsci-l
or email: WDSCI-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/wdsci-l.