NoLeak - include to help you find memory leaks (PB5.20)

[mariosk8s]

This is a real cool add-in.
Now i only wonder if there's a way to get a call stack of some sort to be included.
That way when an allocate is in a nested function that's called from a lot of different places, we could find out who the evildoer was.

[sec]

@luis: Could you please update this to last version?

Thanks.

[luis]

@luis: Could you please update this to last version?

Doesn't it work already ? I tried with the last 5.20 beta and unless I'm missing something it does.

Only problem I see is in the demo the allocation of 99,999,999,999 bytes does not fail, I seem to recall in older version did so (don't remember which version).

Why does not fail BTW ? I understand the pointer returned is not to physical memory but nevertheless... but maybe what's changed is the underlying OS and not something inside PB ?

Practically allocatememory() never fails today until something horrible happen with the entire OS running out of resources ?


EDIT: now there is something else I don't understand. Currently PB should fill with zeros the allocated memory (by default).
So the OS should be forced not only to "virtually" reserve the memory, but to actually commit it since the memory is actually accessed.
How can the allocation above be successful then ?

EDIT 2: actually it does fail, phew, there was an overflow so the allocated memory was a lot less. I'll change it in the version updated for PB 5.20.

[luis]

@sec

You asked me to update the code to the latest PB version.
I took the time to check it, and I've reported back.
The question I made was not a rhetorical one.

Doesn't it work already ? I tried with the last 5.20 beta and unless I'm missing something it does.

[sec]

@luis: I am not expert at hunting the bugs.

I am trying code at your first post with 5.20, it says missing parameters at AllocateMemory, ReAllocateMemory, ...

Thanks

[luis]

Really ? I'll test it again. One of us is probably drunk, mybe I'm the one

[luis]

I just tried, I had the doubt of have used the wrong compiler.
I tried with 5.20 beta 16, x86, under Windows.

It works.

[sec]

Really ? I'll test it again. One of us is probably drunk, mybe I'm the one

Yes

My project has 'refresh' feature', it is increasing 4K after each 'refresh' . bUt i can't find where is root of cause problem.

[sec]

I just tried, I had the doubt of have used the wrong compiler.
I tried with 5.20 beta 16, x86, under Windows.

It works.

for example:

Code: Select all

  *Buffer = AllocateMemory(1024, #PB_Memory_NoClear)

[luis]

AAAAAH, I think I understand.

You are telling me it doesn't work wit YOUR code, not with MY EXAMPLE.

Right ?

Problem is I was testing what is posted HERE and thought you were doing the same, logically


OK, newer PB versions added a new param to allocatememory and in your code you are using it.

The code need to be updated to work with that, I'll do it as I have a moment, probably this evening.

I'll check the other commands too.

OK ? Bye!

[sec]

AAAAAH, I think I understand.

You are telling me it doesn't work wit YOUR code, not with MY EXAMPLE.

Right ?

Problem is I was testing what is posted HERE and thought you were doing the same, logically


OK, newer PB versions added a new param to allocatememory and in your code you are using it.

The code need to be updated to work with that, I'll do it as I have a moment, probably this evening.

OK ? Bye!

Yes, thanks you

[luis]

@sec

Should be ok, give it a try and let me know if I forgot something.
Thanks.

[sec]

@sec

Should be ok, give it a try and let me know if I forgot something.
Thanks.

I get macro error:

Code: Select all

Global remote_host_img = _CatchImage (#PB_Compiler_File, #PB_Compiler_Line, 9, ?my_png,  $7FFFFFFF)

[mariosk8s]

My project has 'refresh' feature', it is increasing 4K after each 'refresh' . bUt i can't find where is root of cause problem.

Try, checking that you call ClearStructure on all the Structs you free. NoLeak doesn't catch that and it can cause leaks.

[sec]

My project has 'refresh' feature', it is increasing 4K after each 'refresh' . bUt i can't find where is root of cause problem.

Try, checking that you call ClearStructure on all the Structs you free. NoLeak doesn't catch that and it can cause leaks.

@mariosk8s: Thanks,

Never hear about ClearStructure before.

Code: Select all

Threaded NewMap cserver.serverinfo()
ClearMap(cserver()) is equal ClearStructure(@cserver(), serverinfo) ?