I've got a problem with emplode on beta11pre (the player software is working fine though).

When I try and run emplode I get a message saying

"Microsoft Visual C++ Runtime Library:
Program: c:\program files\empeg-car\emplode.exe
abnormal program termination"

This is on Win2k Professional (release version), I have got Visual Studio 6 (including VC++ 6) installed.

(reported to bugs@empeg.com)

__
Unit serial number 47 (was 330 in the queue)...

Does the same under NT4 (SP6a)

(list 6284, S/N 00299, 4 gig blue)

Sounds like a missing Dynamic Link Library (DLL)

does it work under win98?

____________________________
Murray

Mine too on W2k.

I had it up and running once, synced some files. Started over again ... -> KO.

TommyE

Sounds like a missing Dynamic Link Library (DLL)

Sounds more like someone calling abort(), probably an uncaught exception. I reproduced it myself on Windows 2000 yesterday but NT4 seems fine (so I expect Windows 98 will also be fine).

*dons virtual deer stalker...*

--
Mike Crowe
I may not be speaking on behalf of empeg above :-)

In fact it was an uncaught exception (which calls terminate() which calls abort()) due to Windows 2000 not complying with the documentation for RegQueryValueEx.

Anyway, there's a fixed version up there now.

--
Mike Crowe
I may not be speaking on behalf of empeg above :-)

I would love to know what the problem was with that particular function, so that I may avoid it in the future. It's such a commonly-used function that I'm sure I'd run into the same problem in one of my programs eventually. What was it, exactly?

___________
Tony Fabris

The size passed back for a REG_SZ (or similar) is documented as including the space taken by the '\0' terminator. If the entry is an empty string it appears that Windows 2000 incorrectly returns 0 rather than 1.

--
Mike Crowe
I may not be speaking on behalf of empeg above :-)

So you're saying that, in the case of an empty string, lpcbData returns "1" on Win98, and "0" on Win2k?

That sucks.

The docs aren't clear on what the function should return in the case of an empty string. Personally, I would expect it to treat an empty string as a null string and always return 0.

___________
Tony Fabris

The docs aren't clear on what the function should return in the case of an empty string.

What is unclear about:

"If the data has the REG_SZ, REG_MULTI_SZ or REG_EXPAND_SZ type, then lpcbData will also include the size of the terminating null character."

?


--
Mike Crowe
I may not be speaking on behalf of empeg above :-)

What's unclear is that the terminating null only counts when there's actually a string to return. I'm pretty sure that, as far as Windows is concerned, a valid string which happens to be zero length is not the same as a null pointer to a string which doesn't exist. But they don't bother to tell you that or anything. The docs should specify exactly what that value should be if a null string is returned.

Difficult string handling is the only thing which pisses me off about C. I'll bet that if you totaled up all of the bugs and crashes in Windows software throughout history, the majority of them could be traced back to string-related things like trashed pointers and strings which aren't bounds-checked. If C had been built with proper native string handling from the beginning, all our jobs would be a lot easier.

___________
Tony Fabris

I don't know the first thing about programmong, so there's plenty about that statement that is confusing. :)

I'm glad there's guys (and/or gals) like you that understand this stuff.

"I drank what?" - Socrates

But if there's nothing to return then it fails. A null string is "", i.e. vaguely like &'\0', therefore the length returned should be one. A null _pointer_ is a completely different vessel of fish.

--
Mike Crowe
I may not be speaking on behalf of empeg above :-)

Difficult string handling is the only thing which pisses me off about C. I'll bet that if you totaled up all of the bugs and crashes in Windows software throughout history, the majority of them could be traced back to string-related things like trashed pointers and strings which aren't bounds-checked. If C had been built with proper native string handling from the beginning, all our jobs would be a lot easier.

Yeah, i have always maintained that C was but a mildly structured variant of assembly :)

Cheers!

Dragi "Bonzi" Raos
Zagreb, Croatia
#5196