dooshorama

i've clearly demonstrated my first instinct is to say something stupid then think when smarter people point things out to me.so, no. anywho, maybe we can close this thread.

from the error message, i would assume you are passing the controlID, not the handle -- did you read it? use ControlGetHandle.

ah, i see. thank you zedna and gary for the clarification. from reading the general help forum, i gathered that the port was only partially done for 3.2.10.0. now i've been corrected. sorry valik; you seem to have personal issues i'd rather not get involved with - take it easy. well then,

let's not make me into the bad guy. if your reasoning is that importing a limited set of an unsupported library into a stable release which breaks many existing "unsupported" scripts is OK, then i say your reasoning is immature and shows little understanding of an efficient project management. the choices are 2: 1) import a fraction of the library at a time into stable release 2) import functions into beta release until 100% complete, then bring into stable release go ahead and convince me why #1 is a better choice than #2, aside from being stubborn; i could be wrong here.

i'm not involved in the forum on a regular basis, especially on the development side, so i apologize in advance if the topic has been covered. i upgraded to 3.2.10.0 and found a large number of Const declaration errors involved with A3Lib. after using the search function on the forum, i found that Gary has ported many A3Lib functions over to UDFs. granted, it's a great effort, but i believe the project should have remained in the beta release until 100% complete and tested. all these collisions and the manual fixes involved on the user end seem unacceptable to me. is there a concept of project-management (or manager) here or do people just check in code without approval? i've downgraded to 3.2.8.1 in the meantime.

i'm attempting to find ways to manipulate a program with non standard controls: TClientGrid, TxRealGrid. the program seems to use a mix of Borland C++ and .NET; i'm trying to see which DLL the program is using to create these Grid controls, so that i can use DllCall as well. i'm using Anywhere PE viewer and Dependency Walker so far. i may have narrowed it down to Vcl60.bpl but i am not entirely sure. does anyone have any suggestions on what can be done to narrow down a specific control class to a dll?

resolution is set on max; window is maximized. as an alternative i was considering rotating the screen to portrait, instead of the default landscape. the problem is that the application has many-many "screens". that is just one of them. i've only examined a small sample so far. that approach could lessen the # of out-bound-screens, but i may still encounter one. so i'm thinking of a full-proof method. silk-test seems to be able to handle this situation. ex: you tell it to click on 50,850 & even though it's not visible, it seems to understand the extended client area & successfully click "B". so it seems technically feasible. it may require some DLL calls. this is a client app. its behaviour is strange however, in that controls dont exist until they are clicked on. so first i need to use coordinates, then the control "exists" at which point i can explicitly control the edit, list/combo box, etc.

i've been dealing with an issue where i am trying to click on a position which is technically part of the client, but is out of bounds in terms of the window size. i'll start with a picture of a sample application window: blue - title bar client area is seperated by a vertical tab into green & yellow. my aim is to click on A then B. problem is A & B do not have controlIDs and can only be accessed via coordinates. i know i can use controlclick to access the yellow area. this would make 0,0 as indicated regardless of the width of the green area. for example: controlclick("title","text","yellowArea","left",1,Ax,Ay) to click on A. problem is, there is a scroll bar involved (gray); portion of the client area (dark yellow) is not visible. B's coordinate (say 50,850) technicaly falls outside of the visible yellow area, which in this example bounds to 400,800. i can not seem to directly click on y>800. so to click on, B, using coordinates only, i would need to scroll down, put B in a known position in the visible yellow area, then issue a controlclick command. this is not an approach i would like to take unless necessary. my question is, is there a way to click on B's coordinate (y=850), when it is outside the visible control area (y=800)? thank you.

thanx larry. doing a [search] in autoit help on "listview" seems to show all GUIx functions and skip the Control command completely.

i'm attempting to automate a ListView32 control; not creating it. i searched the site and found old posts (in 2004) pointing out a series of "LV_xxx" for ControlCommand. it was being implemented in an unstable release of 3.0. as of 3.2.2 i do not see these commands in help. have these commands: - LV_GetItemCount - LV_GetSubItemCount - LV_GetText, index, subindex - LV_SelectAll - LV_SelectClear - LV_IsSelected, index - LV_Select, from index, [to index] - LV_Deselect, from index, [to index] - LV_GetSelected, [0 = single/first, 1 = multiple delimited by | ] been phased out or replaced by other commands or just not documented? thank you.

if you can do the remote registry read successfully, check the following location: HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion it holds several string values that may help: ProductName CurrentVersion CurrentBuildNumber

actually yes, it's more like sets. thanx man.i did not find this UDF in 3.1.0 help though. i guess i should have checked the UDF thread EDIT> now i'm confused.. are you referring to previously written functions? i can't find it. (i interpreted the wink as if they're already written)

i'm in need of a function that compares 2 arrays (1 dimentional) and returns the difference. for example: a1(1,2,3,4) vs a2(1,3) would return array (2,4). i could not find it via search. should i write my own or is there such a thing already?

this was exactly my question, thanx valik. PS> i'm not sure i like this behaviour. i think @error resets should be limited to the functions that actualy use it. i don't see the point in checking for @error if the function doesn't set it to 1. so why reset it to 0? if function_doesnt_use(@error) then dont_reset(@error) because, you don't see someone doing: function_that_doesnt_use_@error() if @error then ... no?

is @error set by commands that are not listed to use it in the help file? ..another words, does every command set @error when it is done? the reason i got suspicious was that i did a command like: [loop] $temp=stringstripws(filereadline($file),1) if @error = -1 then exitloop [endloop] but the loop never finished. even though the help file does not specify that "stringstripws" sets @error, it appears that it does. i'm not looking for help on the code, i've already taken care of it by breaking it down to seperate steps. main question still stands however.

ok. disregard the lv_gettext not working. i figured out what was wrong and it was an embaressingly stupid mistake.

thank you jon & trids. your efforts & examples made my life much easier. now questions.. for me LV_gettext & LV_getitem return 1 not the text. what gives? when should we expect updated documentation on the new command in the unstable build? EDIT> i got lv_gettext working in explorer. i just have to figure out what's going on elsewhere.

that worked great! thank you.

i am wondering if i can select a block and indent it (insert spaces), or remove the leading spaces? essentialy meaning.. moving the whole block left/rite. i recall in some editor this was done via ALT-left & ALT-rite keys.

since everyone seems to be indifferent about the actual implementation effort, perhaps we should clarify what exactly we are argueing about. ease of use? code ethics? cosmetics? ease of debugging? on grounds of cosmetics, i agree with Valik. the code will be easier to read. ease of use, i like the Holger idea, like the current ControlCommand() function. sure, some of the parameters may end up unused, but as far as referencing goes, it's all in 1 place. if i need to know what commands are available, there's 1 command and i can look on 1 page what it can do. with the Valik idea, i would need to remember/reference more commands. as an analogy, if you were in an office, and each person did only 1 thing, you would need to memorize every person's name for each individual task. the phonebook would be huge. where as the Holger idea, there's 1 person you can go to for everything you need (in a particular area). so actualy i prefer calling 1 function 4023 times rather than referencing 4023 functions. here's my take on the comparison: [Holger] + it's 1 function + help reference is located in 1 place - it looks odd - empty parameters may be confusing [Valik] + looks/reads better + maybe less confusing - many functions - each function will have different # of parameters (increases referencing the help file) - help references are located in different places for different functions i don't know about code ethics, ease of debugging, or else. >EDIT< is it possible to implement Holger's idea, but create function aliases to support Valik's idea? so there would be a CommandDoAll("this") and CommandDoAll("that") and also aliases .. CommandDoThis(), CommandDoThat()? or maybe the other way around?

i'll have to see what's new.but actualy i was waiting for the GUI stuff in particular.

any idea when the current beta (.102?) will be released as stable code? (has this been discussed before? )

i recommend using EXPECT for windows to do your telnet stuff. when it's done write to a file. in AutoIT check for existance of that file, then continue.

that is so vague, i'm not sure it makes sense. what statement exactly?

wow. i didn't realize i was that much of an idiot.