On Mon, Sep 15, 2014 at 2:25 PM, David Woodhouse <dwmw2 at infradead.org> wrote:
> On Tue, 2014-08-26 at 16:00 +0200, Nikos Mavrogiannopoulos wrote:
>> + ret = -GetLastError();
> Hm, that makes us return a value that isn't in the normal errno space,
> doesn't it?
> Admittedly, I don't think anyone ever actually *does* anything with the
> return value from script_config_tun(), but still I think we should use
> FormatMessage() to print the error correctly, and then return something
> like -EIO rather than a Windows error code.
I thought it didn't make much sense to bother as there was no check of
this function's error code, but strictly speaking you're right.
>> - if (ret == 0x2331) {
>> - /* This is what cmd.exe returns for unrecognised commands */
> That made some sense...
>
>> if (ret) {
>> + /* This is what cmd.exe returns for unrecognised commands */
> ... but this doesn't. For a start we're not using cmd.exe any more...
> and you seem to be suggesting that *all* values other than -1 mean that
> it was an unrecognised command :)
Indeed. Do you want an updated patch?
regards,
Nikos
