Altap Salamander Support Forums

(
Not sure I'm following, but here's a small one that looks unpack-able for me:
http://rybkaforum.net/mwf/rybkaattach/4 ... Knights.7z
From: http://rybkaforum.net/cgi-bin/rybkaforu ... pid=234552
Suppose it's the :mem1536m that's causing the failure in my case?
)

This archive gives me a different error message, but indeed it fails. Even copying one file from it fails. So this seems to be a different problem to me.

therube wrote:(
Not sure I'm following, but here's a small one that looks unpack-able for me:)

You probably meant non-unpackable? Yeah, that's what I meant in my post - PPMD with 1.5GB dictionary. Looks like we can get different errors in such circumstances. My bad.

mdruiter wrote:As I'm not as convinced as you might be that the problem is understood, I'ld like to know whether I could help to debug it any further, if it occurs again...

Well, it is not understood because it is not believed the file you attached to this thread can cause such problems. Nobody can confirm it. What you could do, is:
1) have the error from post #1 displayed
2) start 2nd instance of AS
3) go to Help/Task List
4) break the 1st instance
5) Send me the bug report

Thanks! Now I hope it will happen again.
As soon as I noticed that other people could not reproduce it, I downloaded my own file and tried again to be sure. And I got the same error. Until I restarted Salamander, which 'solved' it...

Hi mdruiter,
I was able to reproduce the problem with the archive from your first post and after sending the bugreport to Jan Patera, he was able to reproduce it as well...

I'm crossing my fingers hoping the problem will be found and resolved quickly.

mdruiter wrote:With some 7z archives, Unpack archive does not seem to work.
Insufficient memory to allocate 1231194888 bytes.

Uff, it took me several hours to find out there was uninitialized variable preallocating randomly big piece of memory. It was affecting only Alt-F6/Alt-F9.
It will be solved in the next release of Salamander (AS2.53).
Thanks for reporting and to Zarevak for helping out!

@zarevak, what steps did you need to take to force the issue using mdruiter's test file?

therube wrote:@zarevak, what steps did you need to take to force the issue using mdruiter's test file?

He suggested to defragment address space of Salamander by loading many plugins and opening many documents (e.g. in EV and PDF). The key tool, that he told me about and I am grateful for, is VMMap (http://technet.microsoft.com/en-us/sysi ... 35533.aspx) that, among other things, shows the largest free block.
The bad thing on this particular bug was that it could never occurr in a debug build, that's why I initially saw 1MB as the largest block.

> loading many plugins

I tried that too.

> and opening many documents

Didn't think of that.

> VMMap

Was looking at that the other day, but (for me) will take more study before I have a basic understanding of what it's telling me.

Thanks.