Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.lang.clipper.visual-objects > #2119 > unrolled thread
| Started by | "Willie Moore" <williem@wmconsulting.com> |
|---|---|
| First post | 2011-09-30 14:01 -0500 |
| Last post | 2011-10-04 12:03 +0200 |
| Articles | 20 on this page of 45 — 13 participants |
Back to article view | Back to comp.lang.clipper.visual-objects
Vulcan build 208 "Willie Moore" <williem@wmconsulting.com> - 2011-09-30 14:01 -0500
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-04 09:59 +0200
Re: Vulcan build 208 "Willie Moore" <williem@wmconsulting.com> - 2011-10-05 19:34 -0500
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-06 09:09 +0200
Re: Vulcan build 208 Johan Nel <johan.nel@xsinet.co.za> - 2011-10-06 00:35 -0700
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-06 11:33 +0200
Re: Vulcan build 208 Johan Nel <johan.nel@xsinet.co.za> - 2011-10-06 03:51 -0700
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-06 13:28 +0200
Re: Vulcan build 208 E®!k \\/!$$E® <nospam@nospam.url> - 2011-10-06 15:28 +0200
Re: Vulcan build 208 "Ginny Caughey" <ginny.online@hotmail.com> - 2011-10-06 08:08 -0400
Re: Vulcan build 208 "Willie Moore" <williem@wmconsulting.com> - 2011-10-06 07:47 -0500
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-06 17:03 +0200
Re: Vulcan build 208 "Ginny Caughey" <ginny.online@hotmail.com> - 2011-10-07 08:27 -0400
Re: Vulcan build 208 "Willie Moore" <williem@wmconsulting.com> - 2011-10-07 08:51 -0500
Re: Vulcan build 208 Dieter <support@visualobjects.de> - 2011-10-08 16:01 -0700
Re: Vulcan build 208 "Willie Moore" <williem@wmconsulting.com> - 2011-10-09 10:11 -0500
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-07 16:47 +0200
Re: Vulcan build 208 "Dirk (Belgium)" <dirk.dot.herijgers@pbprojects.dot.be> - 2011-10-06 07:36 +0000
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-06 11:48 +0200
Re: Vulcan build 208 "Willie Moore" <williem@wmconsulting.com> - 2011-10-06 07:27 -0500
Re: Vulcan build 208 E®!k \\/!$$E® <nospam@nospam.url> - 2011-10-06 15:32 +0200
Re: Vulcan build 208 "Willie Moore" <williem@wmconsulting.com> - 2011-10-06 09:36 -0500
Re: Vulcan build 208 E®!k \\/!$$E® <nospam@nospam.url> - 2011-10-06 19:33 +0200
Re: Vulcan build 208 "Willie Moore" <williem@wmconsulting.com> - 2011-10-06 14:09 -0500
Re: Vulcan build 208 Massimo Bighelli <info@arcasistemi.it> - 2011-10-06 22:57 +0200
Re: Vulcan build 208 Karl Faller <k.faller_withoutthat_@onlinehome.de> - 2011-10-06 23:39 +0200
Re: Vulcan build 208 E®!k \\/!$$E® <nospam@nospam.url> - 2011-10-07 14:46 +0200
Re: Vulcan build 208 Massimo Bighelli <info@arcasistemi.it> - 2011-10-08 14:51 +0200
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-07 12:04 +0200
Re: Vulcan build 208 "Marc Verkade [Marti]" <marcatm@rti.nl> - 2011-10-07 15:52 +0200
Re: Vulcan build 208 E®!k \\/!$$E® <nospam@nospam.url> - 2011-10-08 14:16 +0200
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-06 17:15 +0200
Re: Vulcan build 208 Karl Faller <k.faller_withoutthat_@onlinehome.de> - 2011-10-06 19:07 +0200
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-07 12:03 +0200
Re: Vulcan build 208 Karl Faller <k.faller_withoutthat_@onlinehome.de> - 2011-10-07 14:41 +0200
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-07 17:03 +0200
Re: Vulcan build 208 Karl Faller <k.faller_withoutthat_@onlinehome.de> - 2011-10-07 17:26 +0200
Re: Vulcan build 208 Richard Wooters <Richard@rbj.com> - 2011-10-07 09:08 -0700
Re: Vulcan build 208 Dieter <support@visualobjects.de> - 2011-10-06 16:13 -0700
Re: Vulcan build 208 "Jamal" <vodotnet_nospam@yahoo.com> - 2011-10-08 17:51 -0400
Re: Vulcan build 208 Dieter <support@visualobjects.de> - 2011-10-08 15:42 -0700
Re: Vulcan build 208 Dieter <support@visualobjects.de> - 2011-10-06 16:41 -0700
Re: Vulcan build 208 D.J.W. van Kooten <public@ic2remove.this.com> - 2011-10-07 11:58 +0200
Re: Vulcan build 208 info@softwareobjectives.com.au - 2011-10-04 02:47 -0700
Re: Vulcan build 208 E®!k \\/!$$E® <nospam@nospam.url> - 2011-10-04 12:03 +0200
Page 2 of 3 — ← Prev page 1 [2] 3 Next page →
| From | E®!k \\/!$$E® <nospam@nospam.url> |
|---|---|
| Date | 2011-10-06 15:32 +0200 |
| Message-ID | <9f5p04Fj00U1@mid.individual.net> |
| In reply to | #2174 |
Willie,
In this scenario, your app will always close after an error, or am I
mistaken?
Not very usefull IMO.
Erik
"Willie Moore" schreef in bericht news:j6k6qo$kir$1@speranza.aioe.org...
Dick,
I am not sure where you got the idea that you cant have a general error
handler. I use the framework below in the start of all of my winforms app. I
have a slightly different one for my WPF apps but both use the same error
logging routine. Error handling is handled a bit differently from VO but in
my opinion, it is a better system. In my methods, I use local handlers (try
catch blocks) to catch and handle things that are specific to that method.
If an exception happens, dotNet first looks for a local errorblock. If it
does not find one it then looks at the next method up the call stack for a
handler that can handle the exception. If it reaches the root of the
application without finding a block then you get a error box that looks a
lot like VO's. By adding the try block in my start function I can catch and
log the errors.
As far as the other points. I structure my apps so all of the projects in my
solution output to the same directory. I rarely touch my base libraries so I
don’t compile them unless I make a change. I do the same with my GUI. I
separate all of my GUI forms into their own assembly. In my exe I subclass
each form and add my event handlers to the subclass just like I did in VO.
One of the nice things about Visual Studio is that you can have different
settings for different config tasks (ie debug, staging, release, etc). In
debug I don’t recompile everything but in release I do so all of the DLLs go
out with the same datetime stamp.
Regards,
Willie
GLOBAL goSysInfo As SysInfo
/// <summary>
/// The main entry point for the application.
/// </summary>
[STAThread] ;
FUNCTION Start() AS VOID
LOCAL oShell AS ShellWindow
LOCAL lRunApp AS Logic
TRY
SetupRightSLE()
Application.EnableVisualStyles()
Application.SetCompatibleTextRenderingDefault( false )
oShell := ShellWindow{}
lRunApp := oShell:Start()
IF lRunApp
Application.Run( oShell )
ENDIF
CATCH ex as System.Exception
LOCAL oLog as ErrorLog
oLog := Errorlog{}
oLog:DumpException(ex)
MessageBox.Show(ex:ToString())
END TRY
RETURN
FUNCTION SetupRightSLE() AS VOID
rightSLE.DefaultSelectionOnFocus := focusBehavior.All
RightSLE.DefaultProcessEnter := KeyHandling.Process
RightSLE.DefaultProcessArrow := KeyHandling.Process
RightSLE.DefaultOverwrite := OverwriteBehavior.Never
RETURN
[toc] | [prev] | [next] | [standalone]
| From | "Willie Moore" <williem@wmconsulting.com> |
|---|---|
| Date | 2011-10-06 09:36 -0500 |
| Message-ID | <j6kecb$9lh$1@speranza.aioe.org> |
| In reply to | #2178 |
Erik,
Yes this would exit and log the error. The error handling that I do at the
method level is designed to catch and recover when I can. But if something
comes up that I had not anticipated, I will at least get a log that I can
see when the problem is located.
Regards,
Willie
"E®!k \/!$$E®" wrote in message news:9f5p04Fj00U1@mid.individual.net...
Willie,
In this scenario, your app will always close after an error, or am I
mistaken?
Not very usefull IMO.
Erik
"Willie Moore" schreef in bericht news:j6k6qo$kir$1@speranza.aioe.org...
Dick,
I am not sure where you got the idea that you cant have a general error
handler. I use the framework below in the start of all of my winforms app. I
have a slightly different one for my WPF apps but both use the same error
logging routine. Error handling is handled a bit differently from VO but in
my opinion, it is a better system. In my methods, I use local handlers (try
catch blocks) to catch and handle things that are specific to that method.
If an exception happens, dotNet first looks for a local errorblock. If it
does not find one it then looks at the next method up the call stack for a
handler that can handle the exception. If it reaches the root of the
application without finding a block then you get a error box that looks a
lot like VO's. By adding the try block in my start function I can catch and
log the errors.
As far as the other points. I structure my apps so all of the projects in my
solution output to the same directory. I rarely touch my base libraries so I
don’t compile them unless I make a change. I do the same with my GUI. I
separate all of my GUI forms into their own assembly. In my exe I subclass
each form and add my event handlers to the subclass just like I did in VO.
One of the nice things about Visual Studio is that you can have different
settings for different config tasks (ie debug, staging, release, etc). In
debug I don’t recompile everything but in release I do so all of the DLLs go
out with the same datetime stamp.
Regards,
Willie
GLOBAL goSysInfo As SysInfo
/// <summary>
/// The main entry point for the application.
/// </summary>
[STAThread] ;
FUNCTION Start() AS VOID
LOCAL oShell AS ShellWindow
LOCAL lRunApp AS Logic
TRY
SetupRightSLE()
Application.EnableVisualStyles()
Application.SetCompatibleTextRenderingDefault( false )
oShell := ShellWindow{}
lRunApp := oShell:Start()
IF lRunApp
Application.Run( oShell )
ENDIF
CATCH ex as System.Exception
LOCAL oLog as ErrorLog
oLog := Errorlog{}
oLog:DumpException(ex)
MessageBox.Show(ex:ToString())
END TRY
RETURN
FUNCTION SetupRightSLE() AS VOID
rightSLE.DefaultSelectionOnFocus := focusBehavior.All
RightSLE.DefaultProcessEnter := KeyHandling.Process
RightSLE.DefaultProcessArrow := KeyHandling.Process
RightSLE.DefaultOverwrite := OverwriteBehavior.Never
RETURN
[toc] | [prev] | [next] | [standalone]
| From | E®!k \\/!$$E® <nospam@nospam.url> |
|---|---|
| Date | 2011-10-06 19:33 +0200 |
| Message-ID | <9f673eFddrU1@mid.individual.net> |
| In reply to | #2179 |
Willie,
> Yes this would exit and log the error. The error handling that I do at
> the method level is designed to catch and recover when I can. But if
> something comes up that I had not anticipated, I will at least get a log
> that I can see when the problem is located.
Sure, and IMO that is a good strategy.
But it does not help Dick, where he is able in VO to hook in his own
errorhandler.
I have taken my lesson from good 'old' Rod da Silva, i think it was on a
conference at the end of the previous century, where he showed us a coding
framework for code that could not crash. It was used for writing services,
but very usefull for ordinairy apps as well.
He gave us some very practical rules of thumb.
- the number of rows of code in a method should not exceed the max. visible
on a screen. If there are more rows, split up.
- every method should have a begin-recover-end sequence inside ( something
that can be implemented very easy using UDC's )
- every method should return a boolean, true if succeeded and false
otherwise.
- if you need another value, pass a parameter by reference.
I have used this for many years and am very happy with this.
But unlike me, writing several new (smaller) app's each year, Dick has a
major large application, born in VO 1.0 and extended over the years.
Not an easy task to rewirte it with new coding principles.
Erik
"E®!k \/!$$E®" wrote in message news:9f5p04Fj00U1@mid.individual.net...
Willie,
In this scenario, your app will always close after an error, or am I
mistaken?
Not very usefull IMO.
Erik
"Willie Moore" schreef in bericht news:j6k6qo$kir$1@speranza.aioe.org...
Dick,
I am not sure where you got the idea that you cant have a general error
handler. I use the framework below in the start of all of my winforms app. I
have a slightly different one for my WPF apps but both use the same error
logging routine. Error handling is handled a bit differently from VO but in
my opinion, it is a better system. In my methods, I use local handlers (try
catch blocks) to catch and handle things that are specific to that method.
If an exception happens, dotNet first looks for a local errorblock. If it
does not find one it then looks at the next method up the call stack for a
handler that can handle the exception. If it reaches the root of the
application without finding a block then you get a error box that looks a
lot like VO's. By adding the try block in my start function I can catch and
log the errors.
As far as the other points. I structure my apps so all of the projects in my
solution output to the same directory. I rarely touch my base libraries so I
don’t compile them unless I make a change. I do the same with my GUI. I
separate all of my GUI forms into their own assembly. In my exe I subclass
each form and add my event handlers to the subclass just like I did in VO.
One of the nice things about Visual Studio is that you can have different
settings for different config tasks (ie debug, staging, release, etc). In
debug I don’t recompile everything but in release I do so all of the DLLs go
out with the same datetime stamp.
Regards,
Willie
GLOBAL goSysInfo As SysInfo
/// <summary>
/// The main entry point for the application.
/// </summary>
[STAThread] ;
FUNCTION Start() AS VOID
LOCAL oShell AS ShellWindow
LOCAL lRunApp AS Logic
TRY
SetupRightSLE()
Application.EnableVisualStyles()
Application.SetCompatibleTextRenderingDefault( false )
oShell := ShellWindow{}
lRunApp := oShell:Start()
IF lRunApp
Application.Run( oShell )
ENDIF
CATCH ex as System.Exception
LOCAL oLog as ErrorLog
oLog := Errorlog{}
oLog:DumpException(ex)
MessageBox.Show(ex:ToString())
END TRY
RETURN
FUNCTION SetupRightSLE() AS VOID
rightSLE.DefaultSelectionOnFocus := focusBehavior.All
RightSLE.DefaultProcessEnter := KeyHandling.Process
RightSLE.DefaultProcessArrow := KeyHandling.Process
RightSLE.DefaultOverwrite := OverwriteBehavior.Never
RETURN
[toc] | [prev] | [next] | [standalone]
| From | "Willie Moore" <williem@wmconsulting.com> |
|---|---|
| Date | 2011-10-06 14:09 -0500 |
| Message-ID | <j6kud3$m6e$1@speranza.aioe.org> |
| In reply to | #2184 |
Erik, I think that we were both in Rod's session. I write very similar code. In dotNet I still write very similar code except I don’t use a UDC anymore. I just write try catch end try blocks. Instead of break when I encounter an error, I throw an error message. Very little ever makes it to the top but I have that protected with a logger just in case it does reach the top. Regards, Willie
[toc] | [prev] | [next] | [standalone]
| From | Massimo Bighelli <info@arcasistemi.it> |
|---|---|
| Date | 2011-10-06 22:57 +0200 |
| Message-ID | <j6l4nu$8eh$2@speranza.aioe.org> |
| In reply to | #2184 |
Erik, I agree (generally) on that principles, and other there are, general, and specific to the language / compiler, all coming from the C and ASM era. But I really don't understand the first. If I have a big monitor and small fonts, can I write longer methods? :-) Massimo > - the number of rows of code in a method should not exceed the max. > visible on a screen. If there are more rows, split up. > - every method should have a begin-recover-end sequence inside ( > something that can be implemented very easy using UDC's ) > - every method should return a boolean, true if succeeded and false > otherwise. > - if you need another value, pass a parameter by reference. > > I have used this for many years and am very happy with this. > But unlike me, writing several new (smaller) app's each year, Dick has a > major large application, born in VO 1.0 and extended over the years. > Not an easy task to rewirte it with new coding principles. > > > Erik >
[toc] | [prev] | [next] | [standalone]
| From | Karl Faller <k.faller_withoutthat_@onlinehome.de> |
|---|---|
| Date | 2011-10-06 23:39 +0200 |
| Message-ID | <8u7s87d8h7j47ausffgi86cc5g115li16g@4ax.com> |
| In reply to | #2186 |
Massimo, >If I have a big monitor and small fonts, can I write longer methods? >:-) you know, there are extremists who'd call any method longer than 7 lines to long and call to split ;-)) Good Night! Karl
[toc] | [prev] | [next] | [standalone]
| From | E®!k \\/!$$E® <nospam@nospam.url> |
|---|---|
| Date | 2011-10-07 14:46 +0200 |
| Message-ID | <9f8akbF608U1@mid.individual.net> |
| In reply to | #2186 |
Massimo, I think you understood the basic idea. I turn my monitor 90 degrees to have more code in my methods <g> Erik "Massimo Bighelli" schreef in bericht news:j6l4nu$8eh$2@speranza.aioe.org... Erik, I agree (generally) on that principles, and other there are, general, and specific to the language / compiler, all coming from the C and ASM era. But I really don't understand the first. If I have a big monitor and small fonts, can I write longer methods? :-) Massimo > - the number of rows of code in a method should not exceed the max. > visible on a screen. If there are more rows, split up. > - every method should have a begin-recover-end sequence inside ( > something that can be implemented very easy using UDC's ) > - every method should return a boolean, true if succeeded and false > otherwise. > - if you need another value, pass a parameter by reference. > > I have used this for many years and am very happy with this. > But unlike me, writing several new (smaller) app's each year, Dick has a > major large application, born in VO 1.0 and extended over the years. > Not an easy task to rewirte it with new coding principles. > > > Erik >
[toc] | [prev] | [next] | [standalone]
| From | Massimo Bighelli <info@arcasistemi.it> |
|---|---|
| Date | 2011-10-08 14:51 +0200 |
| Message-ID | <j6ph0e$m6r$1@speranza.aioe.org> |
| In reply to | #2195 |
Il 07/10/2011 14:46, E®!k \/!$$E® ha scritto: > Massimo, > > I think you understood the basic idea. yes I started programming when PC was an idea. > I turn my monitor 90 degrees to have more code in my methods <g> wow, it's better than mine. I only moved the window's bar on the left !!! :-) Massimo
[toc] | [prev] | [next] | [standalone]
| From | D.J.W. van Kooten <public@ic2remove.this.com> |
|---|---|
| Date | 2011-10-07 12:04 +0200 |
| Message-ID | <pjjt87h07mqnro56i9oibevauvl529ml9t@4ax.com> |
| In reply to | #2184 |
On Thu, 6 Oct 2011 19:33:33 +0200, E®!k \\/!$$E® <nospam@nospam.url> wrote: Hello Erik, >But unlike me, writing several new (smaller) app's each year, Dick has a >major large application, born in VO 1.0 and extended over the years. >Not an easy task to rewirte it with new coding principles. Exactly. We've made quite a few major changes to improve the program in some way. Most of the times, each required many 100's of adaptions in our programs. Dick
[toc] | [prev] | [next] | [standalone]
| From | "Marc Verkade [Marti]" <marcatm@rti.nl> |
|---|---|
| Date | 2011-10-07 15:52 +0200 |
| Message-ID | <4e8f0494$0$788$58c7af7e@news.kabelfoon.nl> |
| In reply to | #2184 |
I can remember this session too, but he said:
- every method should return an integer, 0 if succeeded and other
numbers/defies to identify errors
Gr, marc
"E®!k \/!$$E®" schreef in bericht news:9f673eFddrU1@mid.individual.net...
Willie,
> Yes this would exit and log the error. The error handling that I do at
> the method level is designed to catch and recover when I can. But if
> something comes up that I had not anticipated, I will at least get a log
> that I can see when the problem is located.
Sure, and IMO that is a good strategy.
But it does not help Dick, where he is able in VO to hook in his own
errorhandler.
I have taken my lesson from good 'old' Rod da Silva, i think it was on a
conference at the end of the previous century, where he showed us a coding
framework for code that could not crash. It was used for writing services,
but very usefull for ordinairy apps as well.
He gave us some very practical rules of thumb.
- the number of rows of code in a method should not exceed the max. visible
on a screen. If there are more rows, split up.
- every method should have a begin-recover-end sequence inside ( something
that can be implemented very easy using UDC's )
- every method should return a boolean, true if succeeded and false
otherwise.
- if you need another value, pass a parameter by reference.
I have used this for many years and am very happy with this.
But unlike me, writing several new (smaller) app's each year, Dick has a
major large application, born in VO 1.0 and extended over the years.
Not an easy task to rewirte it with new coding principles.
Erik
"E®!k \/!$$E®" wrote in message news:9f5p04Fj00U1@mid.individual.net...
Willie,
In this scenario, your app will always close after an error, or am I
mistaken?
Not very usefull IMO.
Erik
"Willie Moore" schreef in bericht news:j6k6qo$kir$1@speranza.aioe.org...
Dick,
I am not sure where you got the idea that you cant have a general error
handler. I use the framework below in the start of all of my winforms app. I
have a slightly different one for my WPF apps but both use the same error
logging routine. Error handling is handled a bit differently from VO but in
my opinion, it is a better system. In my methods, I use local handlers (try
catch blocks) to catch and handle things that are specific to that method.
If an exception happens, dotNet first looks for a local errorblock. If it
does not find one it then looks at the next method up the call stack for a
handler that can handle the exception. If it reaches the root of the
application without finding a block then you get a error box that looks a
lot like VO's. By adding the try block in my start function I can catch and
log the errors.
As far as the other points. I structure my apps so all of the projects in my
solution output to the same directory. I rarely touch my base libraries so I
don’t compile them unless I make a change. I do the same with my GUI. I
separate all of my GUI forms into their own assembly. In my exe I subclass
each form and add my event handlers to the subclass just like I did in VO.
One of the nice things about Visual Studio is that you can have different
settings for different config tasks (ie debug, staging, release, etc). In
debug I don’t recompile everything but in release I do so all of the DLLs go
out with the same datetime stamp.
Regards,
Willie
GLOBAL goSysInfo As SysInfo
/// <summary>
/// The main entry point for the application.
/// </summary>
[STAThread] ;
FUNCTION Start() AS VOID
LOCAL oShell AS ShellWindow
LOCAL lRunApp AS Logic
TRY
SetupRightSLE()
Application.EnableVisualStyles()
Application.SetCompatibleTextRenderingDefault( false )
oShell := ShellWindow{}
lRunApp := oShell:Start()
IF lRunApp
Application.Run( oShell )
ENDIF
CATCH ex as System.Exception
LOCAL oLog as ErrorLog
oLog := Errorlog{}
oLog:DumpException(ex)
MessageBox.Show(ex:ToString())
END TRY
RETURN
FUNCTION SetupRightSLE() AS VOID
rightSLE.DefaultSelectionOnFocus := focusBehavior.All
RightSLE.DefaultProcessEnter := KeyHandling.Process
RightSLE.DefaultProcessArrow := KeyHandling.Process
RightSLE.DefaultOverwrite := OverwriteBehavior.Never
RETURN
[toc] | [prev] | [next] | [standalone]
| From | E®!k \\/!$$E® <nospam@nospam.url> |
|---|---|
| Date | 2011-10-08 14:16 +0200 |
| Message-ID | <9fat9lFh9rU1@mid.individual.net> |
| In reply to | #2197 |
Hi Marc,
I can recall he mentioned this, as most win32 API functions work this way.
But if my memory is right, he was not to happy about this.
And of all the sesionmaterial I have collected, this one lacks.
Anyway, Rod for shure has influenced my programming in several ways!
Erik
"Marc Verkade [Marti]" schreef in bericht
news:4e8f0494$0$788$58c7af7e@news.kabelfoon.nl...
I can remember this session too, but he said:
- every method should return an integer, 0 if succeeded and other
numbers/defies to identify errors
Gr, marc
"E®!k \/!$$E®" schreef in bericht news:9f673eFddrU1@mid.individual.net...
Willie,
> Yes this would exit and log the error. The error handling that I do at
> the method level is designed to catch and recover when I can. But if
> something comes up that I had not anticipated, I will at least get a log
> that I can see when the problem is located.
Sure, and IMO that is a good strategy.
But it does not help Dick, where he is able in VO to hook in his own
errorhandler.
I have taken my lesson from good 'old' Rod da Silva, i think it was on a
conference at the end of the previous century, where he showed us a coding
framework for code that could not crash. It was used for writing services,
but very usefull for ordinairy apps as well.
He gave us some very practical rules of thumb.
- the number of rows of code in a method should not exceed the max. visible
on a screen. If there are more rows, split up.
- every method should have a begin-recover-end sequence inside ( something
that can be implemented very easy using UDC's )
- every method should return a boolean, true if succeeded and false
otherwise.
- if you need another value, pass a parameter by reference.
I have used this for many years and am very happy with this.
But unlike me, writing several new (smaller) app's each year, Dick has a
major large application, born in VO 1.0 and extended over the years.
Not an easy task to rewirte it with new coding principles.
Erik
"E®!k \/!$$E®" wrote in message news:9f5p04Fj00U1@mid.individual.net...
Willie,
In this scenario, your app will always close after an error, or am I
mistaken?
Not very usefull IMO.
Erik
"Willie Moore" schreef in bericht news:j6k6qo$kir$1@speranza.aioe.org...
Dick,
I am not sure where you got the idea that you cant have a general error
handler. I use the framework below in the start of all of my winforms app. I
have a slightly different one for my WPF apps but both use the same error
logging routine. Error handling is handled a bit differently from VO but in
my opinion, it is a better system. In my methods, I use local handlers (try
catch blocks) to catch and handle things that are specific to that method.
If an exception happens, dotNet first looks for a local errorblock. If it
does not find one it then looks at the next method up the call stack for a
handler that can handle the exception. If it reaches the root of the
application without finding a block then you get a error box that looks a
lot like VO's. By adding the try block in my start function I can catch and
log the errors.
As far as the other points. I structure my apps so all of the projects in my
solution output to the same directory. I rarely touch my base libraries so I
don’t compile them unless I make a change. I do the same with my GUI. I
separate all of my GUI forms into their own assembly. In my exe I subclass
each form and add my event handlers to the subclass just like I did in VO.
One of the nice things about Visual Studio is that you can have different
settings for different config tasks (ie debug, staging, release, etc). In
debug I don’t recompile everything but in release I do so all of the DLLs go
out with the same datetime stamp.
Regards,
Willie
GLOBAL goSysInfo As SysInfo
/// <summary>
/// The main entry point for the application.
/// </summary>
[STAThread] ;
FUNCTION Start() AS VOID
LOCAL oShell AS ShellWindow
LOCAL lRunApp AS Logic
TRY
SetupRightSLE()
Application.EnableVisualStyles()
Application.SetCompatibleTextRenderingDefault( false )
oShell := ShellWindow{}
lRunApp := oShell:Start()
IF lRunApp
Application.Run( oShell )
ENDIF
CATCH ex as System.Exception
LOCAL oLog as ErrorLog
oLog := Errorlog{}
oLog:DumpException(ex)
MessageBox.Show(ex:ToString())
END TRY
RETURN
FUNCTION SetupRightSLE() AS VOID
rightSLE.DefaultSelectionOnFocus := focusBehavior.All
RightSLE.DefaultProcessEnter := KeyHandling.Process
RightSLE.DefaultProcessArrow := KeyHandling.Process
RightSLE.DefaultOverwrite := OverwriteBehavior.Never
RETURN
[toc] | [prev] | [next] | [standalone]
| From | D.J.W. van Kooten <public@ic2remove.this.com> |
|---|---|
| Date | 2011-10-06 17:15 +0200 |
| Message-ID | <augr87p7a70faf30dc05atopt59eaisobc@4ax.com> |
| In reply to | #2174 |
On Thu, 6 Oct 2011 07:27:12 -0500, "Willie Moore" <williem@wmconsulting.com> wrote: >Dick, > >I am not sure where you got the idea that you cant have a general error >handler. As Erik alread wrote, the application quits. That's the same as VO in the pre-errorhandling period....I have now, for example, in my general error handling a check if the error handling concerns a missing field. In that case, the error is logged and the log is e-mailed to me at the next startup but the user doesn't notice the problem until I notify them to run a conversion (or even do that for them if I have access). And I catch more situations: everything is logged but the user won't notice the error. If that's still possible in Vulcan without writing a TRY-CATCH-END TRY on every five lines of code, then this problem is solved. If not, then error handling would still be a major concern in Vulcan. >As far as the other points. I structure my apps so all of the projects in my >solution output to the same directory. Even without all the time and effort involved with this kind of structuring, I think I a right in my observation that change-save-compile-re-running a simple Vulcan app takes at least 8 seconds more than the same app in VO. Right? That still will cost me half an hour waiting time per programming day... Again: I see the pro's as well and my plan is to spend time again in Vulcan this month to see the nice things. But the shortcomings weigh heavily.... Dick
[toc] | [prev] | [next] | [standalone]
| From | Karl Faller <k.faller_withoutthat_@onlinehome.de> |
|---|---|
| Date | 2011-10-06 19:07 +0200 |
| Message-ID | <pjnr879kvmh4dnstuvmh5bhsf88v467q3o@4ax.com> |
| In reply to | #2181 |
Dick, >That still will cost me half an hour waiting time per programming >day... i agree - but OTOH i think that you should restructure your way of programming a bit to avoid this penalty. I fully feel with you, as i'm culprit of the same bad ways - hacking quicker than i really think and hit F5 to let the compiler show me my stupidities ;-) But i try now to think more in advance, and i suspect this will in the end save more than half an hour a day (given, that wouldn't hurt in VO, too) - this morning i wrote a custom routine, (certainly nothing special) but about a page of code and was quite proud of me, when i finally let it run and it worked at the first try ;-) Re W8 - from what i have seen and read up to now, i can't take it seriously as a contender for desktop W7 - i see it as sort of a "fork" to accomodate the upcoming, better, meanwhile arrived devices like tablets, smartphones, etc. HAND Karl
[toc] | [prev] | [next] | [standalone]
| From | D.J.W. van Kooten <public@ic2remove.this.com> |
|---|---|
| Date | 2011-10-07 12:03 +0200 |
| Message-ID | <s9jt87tmgbteunmcqpt494cg9n563734oc@4ax.com> |
| In reply to | #2183 |
On Thu, 06 Oct 2011 19:07:42 +0200, Karl Faller <k.faller_withoutthat_@onlinehome.de> wrote: Hello Karl, >I fully feel with you, as i'm culprit of the same bad ways - hacking >quicker than i really think and hit F5 to let the compiler show me my >stupidities ;-) > >But i try now to think more in advance, Wise words and you're right too. However, plenty of situations just require you to change-recompile-restart serval times until the solution looks good. That will still take much more time in Vulcan. >Re W8 - from what i have seen and read up to now, i can't take it >seriously as a contender for desktop W7 - i see it as sort of a "fork" >to accomodate the upcoming, better, meanwhile arrived devices like >tablets, smartphones, etc. > I think that too. I just found more than a few messages all around forums and other sites suggestion M$ will eventually quit Silverlight first (understandable; even Adobe is going to HTML5 and leaves Flash) and then .Net. Interesting to see which way it is going. Dick
[toc] | [prev] | [next] | [standalone]
| From | Karl Faller <k.faller_withoutthat_@onlinehome.de> |
|---|---|
| Date | 2011-10-07 14:41 +0200 |
| Message-ID | <smst871ib9v6bpnkjp63askuqab93mcp7u@4ax.com> |
| In reply to | #2191 |
Dick, >M$ will eventually quit Silverlight >first (HTML5) >and then .Net. well, SL it might be, i don't know. But .net? After zillions of $ spent, to go a step ahead of functional C? What should it replace? Karl
[toc] | [prev] | [next] | [standalone]
| From | D.J.W. van Kooten <public@ic2remove.this.com> |
|---|---|
| Date | 2011-10-07 17:03 +0200 |
| Message-ID | <2p4u879dh7g8cjggplajqt1mhj6ul3el63@4ax.com> |
| In reply to | #2194 |
On Fri, 07 Oct 2011 14:41:08 +0200, Karl Faller <k.faller_withoutthat_@onlinehome.de> wrote: Hello Karl, >well, SL it might be, i don't know. >But .net? After zillions of $ spent, to go a step ahead of functional >C? What should it replace? > One of the articles I refer too: is from the Guardian. It litterally says: "Independent developers for Microsoft's platforms are increasingly concerned about the lack of clarity from the software giant over the future direction of its programming platforms such as Silverlight, Windows Presentation Foundation (WPF) and .Net compared to non-proprietary platforms such as HTML5, CSS3 and javascript." http://www.guardian.co.uk/technology/2011/jun/14/microsoft-windows-8-developers So it's not as Ginny seems to think that I try to find a reason to not go to .Net. I do however want to learn if there's a serious chance before I really start investing much time in it. This is one of the "concern" messages on a Silverlight forum: http://forums.silverlight.net/t/200185.aspx/1?Microsoft+Please+read+and+give+clear+guidance+I+don+t+want+to+be+developing+for+a+dead+platform+ Maybe just rumours (MS denied) but I've learned that zillions of $'s are easily flushed away if a better chance to become/remain market leader (or a chance to lose that leadership) comes up. Dick
[toc] | [prev] | [next] | [standalone]
| From | Karl Faller <k.faller_withoutthat_@onlinehome.de> |
|---|---|
| Date | 2011-10-07 17:26 +0200 |
| Message-ID | <qt5u87h1j7rjc1equ9om7va8qrfrf62nc3@4ax.com> |
| In reply to | #2199 |
Dick, >>well, SL it might be, i don't know. >>But .net? After zillions of $ spent, to go a step ahead of functional >>C? What should it replace? >> >One of the articles I refer too: is from the Guardian. It litterally >says: > >"Independent developers for Microsoft's platforms are increasingly >concerned about the lack of clarity from the software giant over the >future direction of its programming platforms such as Silverlight, >Windows Presentation Foundation (WPF) and .Net compared to >non-proprietary platforms such as HTML5, CSS3 and javascript." now, if the Guardian writes so, it clearly is not in any deeper way informed what programming is nor how people have to run their business;-) Naturally there is HTML5, CSS, javascript. Essential for browser based work. And i know "cloud" is THE buzzword out there. But while we here in Germany probably have one the "soundest" connections to the net worldwide, i wouldn't dare to rely any major business on the availability of the Inet connection. So, i wouldn't bet on the dissapearance of "ordinary" languages... Go ahead coding in VO, noone will stop 32bit Apps working on new windows for a long time to come. Karl
[toc] | [prev] | [next] | [standalone]
| From | Richard Wooters <Richard@rbj.com> |
|---|---|
| Date | 2011-10-07 09:08 -0700 |
| Message-ID | <79b00b4e-5e6a-4d21-a2ed-5c6ce28955b1@i33g2000yqm.googlegroups.com> |
| In reply to | #2199 |
I believe all of sits listed are just part a of the rumor mill. To better understand where Microsoft is coming from in terms of .NET and building Metro style apps for Windows 8, take a look at this from the MS Build conference here in SOCAL. http://channel9.msdn.com/Events/BUILD/BUILD2011/TOOL-531T
[toc] | [prev] | [next] | [standalone]
| From | Dieter <support@visualobjects.de> |
|---|---|
| Date | 2011-10-06 16:13 -0700 |
| Message-ID | <458939f2-dc8b-4c22-a36e-53a4302307eb@k10g2000vbn.googlegroups.com> |
| In reply to | #2164 |
On 6 Okt., 09:09, D.J.W. van Kooten <pub...@ic2remove.this.com> wrote: > On Wed, 5 Oct 2011 19:34:49 -0500, "Willie Moore" > > <will...@wmconsulting.com> wrote: > > Hello Willie, > > >Are you going to attend VODC this year? I hear rumors that sp4 may make an > >appearance. > > I found the websitehttp://www.vodc.visualobjects.de/en/ > > The cast of speakers is impressive. I guess there are no VO sessions, > or maybe those of Meinhard. Although Vulcan is still somewhere on my > long term planning the only thing I am interested in for the short > term is VO. And Brian wrote about this conference: > > "No more VOGUI no more VO runtime ." > > Except for the time I have to invest in transporting code (especially > 3rd party etc.) I have 4 reasons which keeps me from wanting to use > Vulcan: > > 1 A small change+recompile+re-run costs much more time than in VO. As > I do that 100's of time on a full day of programming this would > introduce half an hour of waiting time which I find unacceptable. > > 2 I can not sort anymore on changed entities as all entities are in 1 > program. Also I have lost program code occasionly in VO by editing > (and deleting) more than I intended. Chances that this happens are > far higher in Vulcan: it's like Edit all entities in VO.... > > 3 I've understood that it's not possible to set a general error > handler. If I would need to add error handling code on every few lines > which can go wrong I have to triple my amount of code. > > 4 The fact that the source code is visible without a lot of extra > effort > > On the sunny side: I do expect a more stable program and I think I > could benefit from some code more widely available in .Net. For now > these benefits do not outweigh the drawbacks for me. > > Apart from this: I have always wondered why VAT has to be paid on the > VODC conference fee. Everything else I order or sell abroad is 0% VAT > as long as I supply/have the VAT number... > > Dick Dick, > Apart from this: I have always wondered why VAT has to be paid on the > VODC conference fee. Everything else I order or sell abroad is 0% VAT > as long as I supply/have the VAT number... EU rules enforced this up to the beginning of this year. If you buy products you won't and would not have to pay VAT. If you use a service (like VODC or DevFest in the UK) VAT had to be paid (inside the EU) in the country where the service had been delivered meaning: VODC is a service deliverd in Germany, so you had to pay German VAT no matter if you represent a EU registered company or not. This has now been reversed in case both sides (the service user and the service provider) are EU registered companies (B2B relationship). In all other cases (if one side is not a EU registered company or one side is a private person) VAT still has to be paid. In this case the service using company can reclaim VAT from local authorities. So in the end you never had to pay VAT (in a B2B relationship). I'm sorry to say VODC homepage does not reflect this recent change of regulations, but all invoices sent out do so. Regards Dieter VODC Management
[toc] | [prev] | [next] | [standalone]
| From | "Jamal" <vodotnet_nospam@yahoo.com> |
|---|---|
| Date | 2011-10-08 17:51 -0400 |
| Message-ID | <j6qgk0$pmt$1@dont-email.me> |
| In reply to | #2188 |
My head is now spinning, feeling like I am in an endless loop! >> EU rules enforced this up to the beginning of this year. If you buy products you won't and would not have to pay VAT. If you use a service (like VODC or DevFest in the UK) VAT had to be paid (inside the EU) in the country where the service had been delivered meaning: VODC is a service deliverd in Germany, so you had to pay German VAT no matter if you represent a EU registered company or not. This has now been reversed in case both sides (the service user and the service provider) are EU registered companies (B2B relationship). In all other cases (if one side is not a EU registered company or one side is a private person) VAT still has to be paid. In this case the service using company can reclaim VAT from local authorities. So in the end you never had to pay VAT (in a B2B relationship). I'm sorry to say VODC homepage does not reflect this recent change of regulations, but all invoices sent out do so. << "Dieter" wrote in message news:458939f2-dc8b-4c22-a36e-53a4302307eb@k10g2000vbn.googlegroups.com... On 6 Okt., 09:09, D.J.W. van Kooten <pub...@ic2remove.this.com> wrote: > On Wed, 5 Oct 2011 19:34:49 -0500, "Willie Moore" > > <will...@wmconsulting.com> wrote: > > Hello Willie, > > >Are you going to attend VODC this year? I hear rumors that sp4 may make > >an > >appearance. > > I found the websitehttp://www.vodc.visualobjects.de/en/ > > The cast of speakers is impressive. I guess there are no VO sessions, > or maybe those of Meinhard. Although Vulcan is still somewhere on my > long term planning the only thing I am interested in for the short > term is VO. And Brian wrote about this conference: > > "No more VOGUI no more VO runtime ." > > Except for the time I have to invest in transporting code (especially > 3rd party etc.) I have 4 reasons which keeps me from wanting to use > Vulcan: > > 1 A small change+recompile+re-run costs much more time than in VO. As > I do that 100's of time on a full day of programming this would > introduce half an hour of waiting time which I find unacceptable. > > 2 I can not sort anymore on changed entities as all entities are in 1 > program. Also I have lost program code occasionly in VO by editing > (and deleting) more than I intended. Chances that this happens are > far higher in Vulcan: it's like Edit all entities in VO.... > > 3 I've understood that it's not possible to set a general error > handler. If I would need to add error handling code on every few lines > which can go wrong I have to triple my amount of code. > > 4 The fact that the source code is visible without a lot of extra > effort > > On the sunny side: I do expect a more stable program and I think I > could benefit from some code more widely available in .Net. For now > these benefits do not outweigh the drawbacks for me. > > Apart from this: I have always wondered why VAT has to be paid on the > VODC conference fee. Everything else I order or sell abroad is 0% VAT > as long as I supply/have the VAT number... > > Dick Dick, > Apart from this: I have always wondered why VAT has to be paid on the > VODC conference fee. Everything else I order or sell abroad is 0% VAT > as long as I supply/have the VAT number... EU rules enforced this up to the beginning of this year. If you buy products you won't and would not have to pay VAT. If you use a service (like VODC or DevFest in the UK) VAT had to be paid (inside the EU) in the country where the service had been delivered meaning: VODC is a service deliverd in Germany, so you had to pay German VAT no matter if you represent a EU registered company or not. This has now been reversed in case both sides (the service user and the service provider) are EU registered companies (B2B relationship). In all other cases (if one side is not a EU registered company or one side is a private person) VAT still has to be paid. In this case the service using company can reclaim VAT from local authorities. So in the end you never had to pay VAT (in a B2B relationship). I'm sorry to say VODC homepage does not reflect this recent change of regulations, but all invoices sent out do so. Regards Dieter VODC Management
[toc] | [prev] | [next] | [standalone]
Page 2 of 3 — ← Prev page 1 [2] 3 Next page →
Back to top | Article view | comp.lang.clipper.visual-objects
csiph-web