Help File/Documentation Issues. (Discussion Only)

[trancexx]

Like it's said already, WinAPI defines TRUE and FALSE as 32 bit fields with mentioned values. AutoIt's internal bool Variant type isn't defined like that and requires conversion to corresponding interface's type during interaction.

Some other interfaces define matching types differently. For example COM interface defines its version as 16 bit value, all bits zero for false and all bits 1 for true. AutoIt therefore converts True to number that can be interpreted as -1 when interacting with this interface.

If Jon intends to document explicitly conversion then he should document it precisely for each case, each interface.

In my opinion, it's good idea to document it. The only problem is conversion that happens in opposite case, for example during callbacks. Then currently you don't get AutoIt's bool for WinAPI's BOOL type so somebody could ask, hey what's that now?

Edited August 24, 2013 by trancexx

[jpm]

I'm not quite sure what you are saying in regards to case sensitivity, as I've not mentioned that?  Windows BOOL is either TRUE (defined as 1 in the Windows SDK) or FALSE (0).  AutoIt currently follows this convention when translating True/False keywords.

I just mention the bool/BOOL just to warn that MSDN say BOOL but if you use bool in  DllCall() it do the same.

[Ascend4nt]

Some other interfaces define matching types differently. For example COM interface defines its version as 16 bit value, all bits zero for false and all bits 1 for true. AutoIt therefore converts True to number that can be interpreted as -1 when interacting with this interface.

 

This is a good point, and something I hadn't even realized about COM's VARIANT_TRUE being defined as -1.  But really that's always been my experience is that only one of 2 values are used to represent a boolean 'true' in pretty much every language I've messed with: 1 or -1. API's and misc. functions often just test for non-zero, but COM is special..

*Ah, while reviewing this I found this nice discussion on the reasoning behind that and more: BOOL vs. VARIANT_BOOL vs. BOOLEAN.

If Jon intends to document explicitly conversion then he should document it precisely for each case, each interface.

In my opinion, it's good idea to document it. The only problem is conversion that happens in opposite case, for example during callbacks. Then currently you don't get AutoIt's bool for WinAPI's BOOL type so somebody could ask, hey what's that now?

 

It does make sense that it should be documented how the conversions work, even if just for DLLCall and COM interactions.

Edited August 24, 2013 by Ascend4nt

[czardas]

To me this boolean stuff seems to be from the realms of fantasy. If specific numeric values are given to the logical data type 'TRUE' (depending on circumstance) then perhaps it ought to be mentioned. Then again I could be misinterpreting this.

Edited August 24, 2013 by czardas

[AZJIO]

_SQLite_Open
$sDatabase_Filename    |    [optional] Database filename, by default will opens an in memory database.

the path to the file

[FireFox]

It's the same thing for file open etc.

It's filename and not filepath, either we change everything or it stays like this.

[AZJIO]

Better to use the path to the file, because when you run a .bat file it will search in a different directory.

[FireFox]

I changed the description to:

Database filename, by default will open the memory database.

Because the original one means nothing for me.

[FireFox]

Better to use the path to the file, because when you run a .bat file it will search in a different directory.

The story of the Filename/Filepath is "fixed" in the next release

[guinness]

You're lucky FireFox saw this, because I didn't. Hence why I encourage people to use the proper channels.

[jchd]

I changed the description to:

Database filename, by default will open the memory database.

Because the original one means nothing for me.

Better say: "a memory database". You can open as many distinct memory DBs as you whish.

[FireFox]

Better say: "a memory database". You can open as many distinct memory DBs as you whish.

OK, fixed.

[Ascend4nt]

There's another thing I just realized lately - I've relied on AutoIt behavior of zeroing out memory used for DLLStruct's when using DLLStructCreate().  This should probably be documented if its going to remain in future versions, which I believe it should as there are certain scripts I know of that will fail or error-out if this is changed. This of course is only applied to new memory allocations, not to DLLStruct's created with pointers.

[AZJIO]

In the section SQLite , you can add training examples:

Table unique - http://forum.ru-board.com/topic.cgi?forum=5&topic=33902&start=2000#7

Get number of rows - http://forum.ru-board.com/topic.cgi?forum=5&topic=33902&start=2000#4

[FireFox]

There's another thing I just realized lately - I've relied on AutoIt behavior of zeroing out memory used for DLLStruct's when using DLLStructCreate(). This should probably be documented if its going to remain in future versions, which I believe it should as there are certain scripts I know of that will fail or error-out if this is changed. This of course is only applied to new memory allocations, not to DLLStruct's created with pointers.

I understand what you said whoever I can't see where could it change the behavior of some scripts.

Can you give an example? So I can document it clearly in the helpfile?

Br, FireFox.

[Mat]

I understand what you said whoever I can't see where could it change the behavior of some scripts.

Can you give an example? So I can document it clearly in the helpfile?

Br, FireFox.

 

If the behaviour were to change, so AutoIt no longer zeroed all the memory automatically, then this would break scripts.

Basically it just needs the sentence "When creating a structure, AutoIt zeroes all the data". This is always useful to know, as in a lot of languages you can't assume that will be the case.

[FireFox]

I know that I would have to add this sentence but it was for my own knowledge.

Do you mean that if it wasn't the case, if we create a struct we would get what's in the memory address allocated like in C for example?

[mLipok]

Many documents use different words to denote the same thing.

For example, in these functions:

FileGetTime, FileGetSize, FileGetShortName, FileOpen, Ini ....

used words are "filename" or "file"

and in any case you need to use the full path

I propose to standardize a system for recording the information.

[Mat]

Many documents use different words to denote the same thing.

For example, in these functions:

FileGetTime, FileGetSize, FileGetShortName, FileOpen, Ini ....

used words are "filename" or "file"

and in any case you need to use the full path

I propose to standardize a system for recording the information.

This has already been discussed by the MVPs.

In most cases you do not need to use the full path as a relative path will work as well, so the next release should have them all standardised to "file path" as well as explaining what that means.

[mLipok]

Thanks for the info