Nokia N97 SDK - beta - released . Feedback welcomed.

[jan_ole]

Here are the expressions for comparison
- 4*n+1
- p*n+1, 1<=p<=4

For that comparison you need to know the internals of ConvertFromUnicodeToUtf8L() and here, you're only comparing the memory footprint, not the computational footprint of necessary reallocs() inside ConvertFromUnicodeToUtf8L() ( or a pre-scan, which is pretty unlikely. )

The "hole" theory is nice, but fails as soon as your p > 2.

Have you noticed that my reasoning is getting better over time, while you can not show anything except "waaa, ReAllocL sucks"?

No, sorry, didn't notice that. But I might just repeat what I've written before, "waaa, ReAllocL sucks when you use it for making space for a trailing zero which you could easily copy by other means and also need to use the cleanup stack".

I never ever said "waaa, ReAllocL sucks". Also, I've written that I wouldn't mind the example using ConvertFromUnicodeToUtf8L() and then a simple strncpy like mechanism. I'd actually preferred that.

But okay, context doesn't matter and that might include the text written here!?

Cheers
Ole

[wizard_hu_]

For that comparison you need to know the internals of ConvertFromUnicodeToUtf8L() and here, you're only comparing the memory footprint, not the computational footprint of necessary reallocs() inside ConvertFromUnicodeToUtf8L() ( or a pre-scan, which is pretty unlikely. )

HookLogger could reveal allocations inside, if you are that interested. Anyway, so far you were complaining about memory usage. And yes, your code is wasting memory, my proposal may waste CPU cycles. There is always a tradeoff.

The "hole" theory is nice, but fails as soon as your p > 2.

Check it once more. The "hole theory" was explained using the worst case, p=4. 4 > 2 as far as I can remember.

No, sorry, didn't notice that.

I know.

I wouldn't mind the example using ConvertFromUnicodeToUtf8L() and then a simple strncpy like mechanism. I'd actually preferred that.

Something like

HBufC8 *text=CnvUtfConverter::ConvertFromUnicodeToUtf8L(aText);
str.assign(static_cast<const char*>(text->Ptr()),text->Length());

could do that (or using the similar constructor of std::string).
It will not necessarily work, I do not speak STL in fact. And it can be a problem if the string operation runs out of memory.

[rnair510]

I've got the N97 emulator installed on XP pro and while the emulator itself seems to be running fine I cant install any new sis files on it. I'm trying to install the Platform services 2.0 sis and the Ovi store sis on the emulator and in both cases it says that "Certificate Error. Check with application provider". Any idea how I can get around this issue?

I'm trying to install the platform services widget to play with the WRT apis. Thanks

[truf]

I've got the N97 emulator installed on XP pro and while the emulator itself seems to be running fine I cant install any new sis files on it. I'm trying to install the Platform services 2.0 sis and the Ovi store sis on the emulator and in both cases it says that "Certificate Error. Check with application provider". Any idea how I can get around this issue?

I'm trying to install the platform services widget to play with the WRT apis. Thanks

If you google a bit you will find what impossible to use sis packages with ARM compiled exe's on Windows x86 emulator.

[rnair510]

If you google a bit you will find what impossible to use sis packages with ARM compiled exe's on Windows x86 emulator.

Thank you. I've primarily done Python coding on S60 so I was unfamiliar with this issue - thanks again.