I wondered if anyone has had similar problems or knew of a fix. Every time I select a word file on my c drive (which is the partition that holds my old vista files) the memory leaks and my machine freezes up until I can kill the process.
chuckr your being sarcastic right? Memory leak - Wikipedia, the free encyclopedia
I thought they might have changed the definition of the word-pair.
...can only be diagnosed by a programmer with access to the program source code...
I think we're going to have fun here :)
The Wikipedia entry is a bit ambiguous (while accusing others of n00bery). It is entirely possible to diagnose a memory leak without access to app source or any type of code-level analysis. A process whose commit charge keeps on increasing in the absence of commensurate workload increases, to the point where it either runs into it's maximum size limit or artificially prevents itself from doing so, is absolutely leaking.
What they presumably mean is that one cannot fix the leak without source code, which is also technically incorrect. For example, perhaps the thing was written in assembly and there is no higher-level source to speak of.
Me too...
I guess that depends on what you mean by 'diagnose'.The Wikipedia entry is a bit ambiguous (while accusing others of n00bery). It is entirely possible to diagnose a memory leak without access to app source or any type of code-level analysis. A process whose commit charge keeps on increasing in the absence of commensurate workload increases, to the point where it either runs into it's maximum size limit or artificially prevents itself from doing so, is absolutely leaking.
In the OP, it seems to mean 'observe' the memory leak.
The word-pair used to mean c-programmers requesting free memory blocks for their code, then never releasing said memory back to the 'free-pool' in the OS.
The Memory Manager would then have those blocks marked as 'in use', when in fact they were not. But still being consumed.
That was considered 'memory-leak', along with other factors.
Eventually, the sum of these 'in-use' marked blocks, though not in use, would consume, etc.
Here, 'diagnose' would mean "repair".What they presumably mean is that one cannot fix the leak without source code, which is also technically incorrect. For example, perhaps the thing was written in assembly and there is no higher-level source to speak of.
I agree with "cannot fix the leak without source code", and I believe it is technically correct:
An 'assembler' requires source code (mnemonics) as input, to generate the computer's binary op-codes, etc. as an object file for linkage, etc. to generate the .exe executable...
You're not confusing 'assembly' with 'machine code', are you?
"Higher level" source is not required -- Assembly, of itself, is source code.
Viz:
This was for radar antenna 'Ellipticity Measurement' analysis on an MPS-36 radar at the Kwajalein Missile Range.
(See there, they changed the 'terminology' again...)
Ronald Reagan Ballistic Missile Defense Test Site - Wikipedia, the free encyclopedia
Had to find 'half-power' points at all quadrants, as well as determine the 'Full-power' point.
The code was called from a 100 pps 'Interrupt handler', which also gathered Elevation, Range, time-code, and other stuff.
This was my first run-in with Microsoft, regarding errors in their documentation:
The ordering and quantity of parameters on the stack.
I had determined the necessary corrections required for their documentation and received their 'prima donna' attitude.
It was also the last.
Agreed. They're using "diagnose" as in "diagnosis: there's a memory leak".
Yes. For what it's worth, I wasn't disagreeing with anything you said in your previous posts. My issue was with Wikipedia's wording :)
Assembly is only the "source" if the executable was written in assembly. Otherwise, if the app was written in C, then the "source" files will be text representing C code. Ditto C++, C#, Java...
Once we agree on those semantics, I think we'll both be on the same page regarding the feasability of rectifying memory leaks through analysis of the disassembled mnemonics - with no access to the original source code.
(As a minor off-topic point, "object files" are produced by a compiler, not by an "assembler".)
A calling convention mismatch of some sort? I'd be interested to hear about your respective positions on the mechanics of that issue. We should probably take it up elsewhere though, to give the OP their thread back :)
Quote