Me have it with the DBe page:
1) resize the palette so that it offers horizontal scrolling
2) click scrolling buttons until bounds are reached (and further if no AV
yet) then back, so on many times until AV happens.
-
Is it as designed ? If yes then in which cases ?
The data are set in AfterOpen .
--
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
wh
There's the need in such ones.
--
Got visibility?
Most devs has no idea what their production app looks like.
Find out how fast your code is with AppDynamics Lite.
http://ad.doubleclick.net/clk;262219671;13503038;y?
http:
> But we need more statistic before putting it to the README.
>
Me had 2 such hangs today with the screen deactivated. But it allows to
perform some work, at least :)
--
Got visibility?
Most devs has no idea what thei
>> >>> No coming back :( The pipeline to GDB doesn't recover by itself.
>> >>
>> >> Maybe different gdb versions? Mine says "GNU gdb (GDB) 7.4.1".
>> >
>> > Tried with mingw32 GDB 7.4.1. Hanged on 1-st save attempt.
>>
>> Also hangs with GDB killed before saving data.
>>
> If it is reproducible it
>>> No coming back :( The pipeline to GDB doesn't recover by itself.
>>
>> Maybe different gdb versions? Mine says "GNU gdb (GDB) 7.4.1".
>>
>
> Tried with mingw32 GDB 7.4.1. Hanged on 1-st save attempt.
Also hangs with GDB killed before saving data.
-
>> No coming back :( The pipeline to GDB doesn't recover by itself.
>
> Maybe different gdb versions? Mine says "GNU gdb (GDB) 7.4.1".
>
Tried with mingw32 GDB 7.4.1. Hanged on 1-st save attempt.
--
Got visibility?
Most
> There is 'Target'-'Restart Debugger'.
>
IDE already hangs (no access to the menu) when there's the need in this
operation.
--
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Downloa
> Deleting all GDB related stuff fixed the problem temporarily - now it
> happens again.
> This problem is bound to pipes to/from GDB (IDE-in-IDE run shows it). Is
> it possible to detect no response (for any reason - wrong params,
> died/zombie process,..) from GDB to restart it so on ?
> These ha
>> May it be because of a NIL address BP ? (see the attachment)
>
> Although it's impossible to delete this BP if it exists alone :)
>
>
Deleting all GDB related stuff fixed the problem temporarily - now it
happens again.
This problem is bound to pipes to/from GDB (IDE-in-IDE run shows it). Is
> May it be because of a NIL address BP ? (see the attachment)
Although it's impossible to delete this BP if it exists alone :)
--
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Downlo
May it be because of a NIL address BP ? (see the attachment)<>--
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://ad.doubleclick.net/clk;25
> Today me had it approx 50 times.
>
> Has someone else encountered that ?
Absolutely impossible to work today :( 100% sure for an eventual hang on
saving changed source files.
--
Everyone hates slow websites. So do we
> Today me had it approx 50 times.
>
> Has someone else encountered that ?
>
definitely it's affected by other apps and/or network attacks but which..
--
How fast is your code?
3 out of 4 devs don\\\'t know how their co
like Win-32 provides for its API based message boxes ( "Ctrl+C/Ins"
selects all text in the message box).
(very handy)
--
How fast is your code?
3 out of 4 devs don\\\'t know how their code performs in production.
Find
Today me had it approx 50 times.
Has someone else encountered that ?
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers
A bit cumbersome code, though:
//--
procedure tappmainfo.mt1ionewexec(const sender: TObject);
begin
with dmmainmo,lqrySessionData do begin
{returning if the needed submodule is active and works in needed mode }
if (tpoperationsinstancevarpo = @mt1inputfo) {текущий субмод
> You want to know the class of the current submodule? git master
> dcd07d48fca82ce794c2fd0595f70e95090657a7 has a ttabwidget.subform
> property
> for convenience.
>
1) it's always NIL for me in the below check:
if tpOperations.subform <> nil then begin
[..]
end
2) FreeAndNil(..) doesn't accep
>> //---
>> procedure tappmainfo.mt1ionewexec(const sender: TObject);
>> begin
>>twMain.activepageindex:= -1; // < does it remove the previous
>> instance ? <<<
>
> Yes (not tested).
It doesn't destroy.
-
>> in which cases should me set "instancevarpo" ? For embedding already
>> created submodules ?
>>
> To access them in your code. Note that it will be nil'd automatically
> after
> destruction of the subform.
Are they those instance vars which are defined in submodule units ?
> On Tuesday 25 September 2012 11:36:34 news.gmane.org wrote:
>> > Which where used by tab changing only. I think too you should use
>> //---
>> procedure tappmainfo.mt1ionewexec(const sender: TObject);
>> begin
>>twMain.activepageind
Which are the condition of NOT calling ?
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
wi
> Which where used by tab changing only. I think too you should use another
> container widget. As Sieghard writes, a tsimplewidget has a
> tcaptionframe for
> example.
>
Sure, but me still can't design the exact app interface (there'll be more
experimenting) and the tabbed (or mixed) design i
See the attached stack trace:
#0 0040C666 :0 fpc_raiseexception()
#1 005472E1 :0 VARIANTS_VARARRAYCREATEERROR()
#2 00553782 msevariants.pas:51 MSEVARARRAYCREATE(BOUNDS=0x0, DIMS=0,
AVARTYPE=8, result={VTYPE = 0, RES1 = 0, RES2 = 0, RES3 = 0, VSMALLINT =
0, VINTEGER = 0, VSINGLE = 0, VDOUBLE
> release() or free() the old and create() the new submodule?
>
Where..when.. how.. :)
Could You extend Your example so that it'll also shows the single tabpage
approach as well ?
--
Live Security Virtual Conference
Ex
>> Me need to change between different submodules over a single tabpage
>> of the main form on selection in the main menu. How can
>> "ttabpage.OnGetSubform" be used in this approach ?
>>
> Return the wanted submodule class.
>
Me don't change between tab pages so there won't be a chance to trigg
> Well - what's the mathematical / logical difference?
>
> (Do these generate different code, even?)
>
Me've just forgot to clean after inversion of the condition :)
--
Live Security Virtual Conference
Exclusive live eve
> if (tag >= low(menuenableccheckfuncar)) and (tag <=
> high(menuenableccheckfuncar)) then begin
>enabled:= menuenableccheckfuncar[tag]; // <<< HERE
>
Ha, it needs C-like syntax ( with parenthesis ) :
enabled:= menuenableccheckfuncar[tag](); // WORKS
Me'm laughing loudly :)
Functions are defined as :
function cansettings:boolean;
begin
with dmmainmo.qrySettings do begin
result:= canlocalrefs;
end;
end;
--
Got visibility?
Most devs has no idea what their production app looks like.
Find out h
>> So that to enable/disable them depending on TAG property.
>>
>> (me ask it because can't get how to do it for menu with different length
>> of branches).
>>
> Use a recursive procedure (not tested):
> "
> procedure dosomething(const aitems: tmenuitems);
> var
> int1: integer;
> begin
> with ai
So that to enable/disable them depending on TAG property.
(me ask it because can't get how to do it for menu with different length
of branches).
--
Got visibility?
Most devs has no idea what their production app looks
> Actions on a dedicated data module if bound to a parent form won't be
> called in descending (to the parent) forms and it's reasonable since at
> same time there may be 2+ forms descending from the parent form - for
> which one to call the actions ?
>
At least FieldFieldLink-s don't work in this
>> = no problems with visibility/actuality in run-time for descendants.
>>
> Why can't this been done with a datamodule?
>
Actions on a dedicated data module if bound to a parent form won't be
called in descending (to the parent) forms and it's reasonable since at
same time there may be 2+ fo
> parent/submodule forms). With the action list, we can have & define all
> needed actions at same level (name space) with components they address to
> = no problems with visibility/actuality in run-time for descendants.
>
>
Appzs for my being such a (crazy) lover of inheritance :)
>> The mission is to pack multiple actions into a single placeholder (helps
>> to help designed form more clear in case there're a lot of much action
>> components on the form).
>>
> Please use data modules. I don't like Delphi's TActionList. ;-)
>
Not always handy - me encountered it yesterday wi
The mission is to pack multiple actions into a single placeholder (helps
to help designed form more clear in case there're a lot of much action
components on the form).
--
Everyone hates slow websites. So do we.
Make
> FindWidget(name) seems to do the work :)
[..]
> FindWidget(self.name) returns NIL :)
Queried again :)
--
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for
>> with mnuMain.menu.items[1].submenu do begin
>> for i:= items.count - 1 dowto 0 do begin
>> if items[i].action.tag < 1 then begin
>> items[i].free;
>> end;
>> end;
>> end;
>>
>
The correct one is :
//--
with mnuMain.menu.submenu.items[i].
> with mnuMain.menu.items[1].submenu do begin
> for i:= items.count - 1 dowto 0 do begin
> if items[i].action.tag < 1 then begin
> items[i].free;
> end;
> end;
> end;
>
Sure, "downto" - to avoid shifting numeration.
But anyway it doesn't compile :
//---
C
by their tag. Me mean proper syntax like:
with mnuMain.menu.items[1].submenu do begin
for i:=0 to items.count - 1 do begin
if items[i].action.tag < 1 then begin
items[i].free;
end;
end;
end;
--
--
>> Or at least to block the wrong operations with a descriptive "to-do"
>> message.
>
> Not possible because there can be closed descendents.
>
Then the only way - to propagate the changes to the descendants regardless
of their state. Though it may break DB bindings etc for their descendants
w
> Correct. Maybe it is possible to implement another complicated hack to
> allow
> it...
Or at least to block the wrong operations with a descriptive "to-do"
message.
As You tell we need to handle all irregularities :)
-
> The type of inherited internal persistent fields can not be changed.
> Please
> use field components or delete them first.
>
But the changes are tried on PARENT2FORM that's in the place of their
(calc fields) creation/definition (no inheritance aboard).
Really the problem looks like the olde
> But in your code you call append(), post() and what not...
>
Then for such scenarios (me have to use EDIT/POST because values in the
dataset are used as params to other datasets which provide field values to
the dataset again - it's a way of implementing multiple dependent choices
whose de
>> "MAINFO.tdbintegeredit1" causes the "Value required" exception which
>> turns
>> into crash.
>>
> Why is this my problem? ;-)
Well..it happens on starting application before a user can access smth on
forms so crashing whole application on a DB edit without even a touch to
this edit..
What
Accepted for new field but fails for changing data type of existing
fields, with the attached IDE exceptions.<><><><>--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
t
> Workaround: add them at the end of the list.
>
Doesn't always help (in PARENT2FORM.MMSESQLQUERY1 they're shift up to stay
before the component fields - whose are alway placed in end). Instead of
"Privleged instruction" we may have range check etc errors
(not 100% reproducible). I may be even
With latest test case, try to add a field of such type to
PARENT2FORM.TMSESQLQUERY1.
ft*int, ftNumeric are OK.
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
thre
> and get (see attachment). Looks OK for me.
>
> Martin
Opps, it seems that my version of PrintQuery mixed numbering of persistent
fields & fielddefs, correct ?
The correct one will be smth like :
procedure printquery (const qry:tmsesqlquery; const title: msestring);
var
s,s1,s2: msestring;
> Did you read my code snippet carfully?
> "
> procedure dmmt1moacquire;
> begin
> if refcount = 0 then begin
> modinstance:= tdmmt1mo.create(nil); <
> // application.createdatamodule(tdmmt1mo,modinstance); <<--
> // module will be destroyed by applicatio
> modinstance.free;
> ==
> Me tried FreeAndNil(modinstance) and also got SF. FPC sometimes
> exposes surprises :)
>
FREE doesn't help. Still SF.
Me forger if included the RELEASE/FREE call in the destructor, in the sent
test case.
destructor tparent2fo.destroy;
begin
inherited;
dm
>> That's if we inserted a form as a submodule then meaning of these
>> handlers
>> of the form reverses, correct ? Messy :) If yes, then to the
>> "README.hints" :)
>>
> It is how Delphi streaming works.
>
Just checked & tested.
For parent code accessing props of loaded components on parent for
>> That's me can rely on it (OnLoaded then OnCreate ..) in my programs,
>> correct ?
>>
> Not in submodules. Submodules onloaded is called after on create.
>
That's if we inserted a form as a submodule then meaning of these handlers
of the form reverses, correct ? Messy :) If yes, then to the
Click within empty area of PARENTFORM1, select all components, Delete
them via RClick, click within the empty area, press ESC button, RClick
for Paste, call Paste (gets available here) - me have an error.
Which error?
Here attached.
It is caused by clipboard data MSEide can't understand (copi
>> With the latest test case:
>>
>> - create a new form inherited form MAINFO
>> - note that the tabwiget is unusable (doesn't accept tab pages) & the
>> button lost its caption
>> - to fix we should call "revert to inherited" for all affected componets
>>
> Please try again with git master
> 66f
> Normally yes, if begingloballoading() has been called onloaded fires in
> notifygloballoading().
>
That's me can rely on it (OnLoaded then OnCreate ..) in my programs,
correct ?
--
Live Security Virtual Conference
Ex
> Hi,
> The wrong names of tcustommseform and tmsedatamodule onloaded and
> oncreate
> have been fixed in git master 8da93ce835e3e97cf2d4edec71fe7e31dbe872f0,
> what
> was onloaded before is now oncreate, oncreate becomes onloaded.
>
>
Does it mean that OnCreate is executed later than OnLoaded
Converting file "inputform.mfm" class "tinputfo" unit "inputform"'.
Converting file "mainbaseform.mfm" class "tmainbasefo" unit
"mainbaseform"'.
Converting file "dmcommondata.mfm" class "tdmcommondatamo" unit
"dmcommondata"'.
Converting file "dminfoconn.mfm" class "tdminfoconnmo" unit "dminfo
What are the conditions of change of return of "asold*" ? Applying
updates ?
POST doesn't affect the "asold*" return values (checked in
"field[n].OnChange" ).
--
Got visibility?
Most devs has no idea what their produ
- sqlqry1 provides (via lookupbuffer) localquery.field1
- sqlqry2 provides (via lookupbuffer) localquery.field2, and sqlqry2
depends on a param bound to localquery.field1
Will this work as to proper values in localquery.field1 &
localquery.field2 ?
-
>> Hmm... doesn't expose with the current test case. Please a clue what to
>> add to the test case related to displayvalues.
>>
> No clue. :-)
>
It may even be a damage to MFM files caused by the recent problems.
In which scenarios can these props be used ?
-
>> Here attached. Happens on deleting persistent fields (of qryMaster) by
>> the
>> fields editor.
>
> That seems to be another problem. Steps?
>
Hmm... doesn't expose with the current test case. Please a clue what to
add to the test case related to displayvalues.
---
> Better use field components
Sure but after stabilizing :) Although it's usually my choice.
--
Got visibility?
Most devs has no idea what their production app looks like.
Find out how fast your code is with AppDynamics
Here is a piece of my today's code :
with qryMaster do begin
actFindIndex.enabled:=
not (fields[0]{company_code}.isnull or fields[1]{catalog_code}.isnull
or fields[2]{client_type_code}.isnull);
end;
--
Got visi
> It fetches the values before post by a fieldlink updatedata call AFAIK .
>
What should me do to see the values ?
>> PS:
>> Me have had to recreate almost from zero some forms with inheritance
>> because of the wrong fields order bug. It was a really painful bug.
>>
> You use the field number re
Me tried to use it for writting initial values from a just SELECTed query
to another (local dataset with appended single record) query but the
target query seems don't receive the values (it should display them in
bound DB boolean edits). Where to dig?
PS:
Me have had to recreate almost fro
leaves in PAS-file, sure.
--
Got visibility?
Most devs has no idea what their production app looks like.
Find out how fast your code is with AppDynamics Lite.
http://ad.doubleclick.net/clk;262219671;13503038;y?
http://inf
that's components dropped on the from later.
It means manual deleting their references in PAS-file before new inherting
operation.
As designed ?
--
Got visibility?
Most devs has no idea what their production app looks l
A screenshot of the not informative error message please so I can see
what you
mean.
Here attached. Happens on deleting persistent fields (of qryMaster) by the
fields editor.<>--
Got visibility?
Most devs has no ide
> Why do you think "If there's an exception then rollback the last change"
> can
> be done without a big development effort?
>
If to intercept the mistake as early as possible so that it won't
propagate & branch..
Though an exception (me offer to rely on) usually happens on a quite deed
stag
>> If a dbenumeditlb on another form references a lookupbuffer on another
>> data module,..
>>
> Screenshot?
>
The current project haven't yet grown to this stage :)
--
Got visibility?
Most devs has no idea what their pr
> Not a 100% reliable receipt.
> With the tets case, Trying to edit "frame.caption" of the page of the
> inherited tabwidget makes this tabpage wrongly sized.
>
Hmm, it seems to be a common (w/o inheritance etc) problem of TTabWidget
with single page.
--
>> Why ? If there's an exception then rollback the last change.
>>
> OK, so please try it and send a patch.
>
>
It seems that You're tired. Really, the latest bugs ate like a can of
worms shaken out :)
--
Got visibilit
>> fix. Important for complex projects with multiple inter-form references.
>>
> Example?
>
If a dbenumeditlb on another form references a lookupbuffer on another
data module,..
--
Got visibility?
Most devs has no idea
> - to fix we should call "revert to inherited" for all affected componets
>
Not a 100% reliable receipt.
With the tets case, Trying to edit "frame.caption" of the page of the
inherited tabwidget makes this tabpage wrongly sized.
With the latest test case:
- create a new form inherited form MAINFO
- note that the tabwiget is unusable (doesn't accept tab pages) & the
button lost its caption
- to fix we should call "revert to inherited" for all affected componets
PS:
We introduced some more bugs but anyway we'll behefit f
> AFAIK the messages are informative.
>
They don't point to where (lookup buffer component name etc,..) to go for
fix. Important for complex projects with multiple inter-form references.
>> To rollback EVERYTHING (incl the initiating change) on such exceptions.
>>
> Needs another man-year of de
>> Here we have an anonymous exception message giving no guess where & what
>> to fix.
>
> Example?
>
In the latest test case, try to change (in fielddefs) field name of
tmsesqlquery2.FieldDefs[0] then reset Active property of the query. We
will have 2 exception sources here : 1) persistent li
For instance, with a local mode query :
- we have a lookup buffer with defined bindings (or a set persistent
fields)
- we change a field name (in FieldDefs) of the query which is bound to
the lookup buffer (or the fields set)
Here we have an anonymous exception message giving no guess wh
for local queries wit manually defined FieldDefs.
--
Got visibility?
Most devs has no idea what their production app looks like.
Find out how fast your code is with AppDynamics Lite.
http://ad.doubleclick.net/clk;26221967
>> And during this (parent.query.active => child.query.active) activation
>> we should check that there's no such field so there's no need to
>> inform the user about that since one just deleted this field
>> intentionally :)
>>
> What next? To give one's shoes a shine?
>
Really, what user should
> 3) "Delete" then "Paste" causes an Exception
Me mean "Delete component(s)" on PARENT1FORM.
"Paste" is available not every time after deleting - it depends on if the
form was focused etc...
--
Live Security Virtual C
>> Copy (with CTRL key) all components on PARENT1FORM and instantly Paste
>> them on this (PARENT1FORM) form.
>>
> Please try again with git master
> 5aba759c3005ca7fda92c77bd4722a9544c8c03c.
>
>
1) Cut & Paste everything of PARENT1FORM looses selection state of the
field components after past
> Looks OK for me:
> "
> tdbenumeditlb3: tmsesqlquery1 : Field not found : "tmsesqlquery1Field4"
> "
> which is true.
>
But it's PARENT1FORM which is the parent of PARENT2FORM that knows nothing
about its (future) internals (which were defined in PARENT1FORM2 and then
deleted). As designed ?
>> >> Fields are still swapped in runtime. We should either have displayed
> This is correct. Internal persistent fields always are before the
> external.
>
The order looks fixed !
Another issue:
Make sure that "tmsesqlquery1" is active both on PARENT1FORM & PARENT2FORM
then delete the 2 inte
Fields are still swapped in runtime. We should either have displayed
":2xx:3xx:" & different field order in the persistent fields editor or
displayed ":3xx:2xx:" with the existing order.
Did you check the the existing order with the field editor? Probably the
wrong
order is still in *.mfm.
F
> Please try again with git master
> 63bede0756b2c1bf7e8838df6c0399b5411feeb8.
>
Fields are still swapped in runtime. We should either have displayed
":2xx:3xx:" & different field order in the persistent fields editor or
displayed ":3xx:2xx:" with the existing order.
--
>> Me tried to prepare a testcase (all with local datasets) [here
>> attached] but it gets a DB exception on starts and exits.
>> If You could manage it to start then please check whether combo boxes
>> set proper (non-swapped & non-shifted) values in tmsesqluery1's fields
>> (will be reported by a
>> Definitely, they are copied twice - 1) as internal parts of the dataset
>> &
>> 2) as components dropped onto the form. So the components should store
>> references to datasets they belong to & ordering in these datasets - to
>> check against possible doubling.
>
> Steps to reproduce with curr
> Fields are are copied with the dataset because they are streamed inside
> the
> dataset, same as child widgets of a widget.
> Maybe they are copied twice?
>
Definitely, they are copied twice - 1) as internal parts of the dataset &
2) as components dropped onto the form. So the components sho
> Steps, expected result, actual result?
>
>
After selecting in two comboboxes, ShowMessage shows swapped values
":2xx:3xx:" although should be ":3xx:2xx:" since tmsesqlquery1Field4 (with
3xx values) has field index [1] (in the persistent fields editor ) as long
as tmsesqlquery1Field1(=mselo
> Steps?
1) create a new form
2) holding CTRL key, select all DB-componets on parent1form the try
a) "Cut"
b) "Paste" to the new form
--
Live Security Virtual Conference
Exclusive live event will cover all the ways
Could You reproduce ?
May me wait fixes today ?
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussi
> Also with the test case:
>
> Try selecting all components on "parent1form" then "Cut" via RClick. Me
> have "component named tmselongintfield1 already exists" exception here.
>
Then try to close project refusing to save any changes - me have an AV on
2-nd attempt.
---
Also with the test case:
Try selecting all components on "parent1form" then "Cut" via RClick. Me
have "component named tmselongintfield1 already exists" exception here.
--
Live Security Virtual Conference
Exclusive liv
>> Having placed the testcase within a simplier named directory
>> (c:\mse-tstcases\..), "everything is OK".
>>
> What means "everything is OK"?
The proper project loads :)
"There are no bugs" or "the testcase runs
> and
> shows the bugs"?
>
No, it still exits on a DB exception. Me very hope th
Anyway these names should match ones in FieldDefs (there will be complains
on non-matches on exiting the editor), and also me had an AV when
occasionally deleted such a name & dis smth (can't reproduce anymore).
Also, me guess that row select (even with selecting by clicking in the
autonumb
Having tried to load the testcase project on my work machine, another
project (last worked) is loaded instead if the testcase is placed within a
project in "C:\Documents and Settings\User\My Documents\mse-testcases\..".
Having placed the testcase within a simplier named directory
(c:\mse-ts
> Later, no time at the moment.
>
Me need it today, if possible :)
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers c
With the testcase at
http://www.likan.uz/uploads/mse_testcases/htmlframe.zip, how to switch
between "zeroinputform" & "mt1inputform" on "mainfo.tpSwitch" using
actions actZero & actNewMT1 rspectively ? (bound to the mainmenu [1][0,1]):
procedure tmainfo.newmt1exec(const sender: TObject);
beg
1 - 100 of 151 matches
Mail list logo
