[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] Vincent's TTY fix problem ?? Was: Re: Issue with daily snapshot Mint kernel



Yes, that's fine to do, and it's not strange. It is expected behaviour
when you call Fselect(1,...) now.

But I suggest you switch to Fpoll() instead.

I'll put back the Fselect(1,...) fix in the kernel.

Alan.

On 10/10/14 22:49, OL wrote:
> Alan,
>
> I have replace all my Fselect(1,...) by Fselect(2,...) test the Mint
> build of yesterday and MyAES ... start!
>
> Very strange
>
> Olivier
>> On 10/10/14 21:51, OL wrote:
>>> Hello,
>>>> On 10/10/14 20:41, OL wrote:
>>>>> Hello I have done the test and compil Mint for Aranym.
>>>>>
>>>>> the issue is not in TTY patch, the problem is in dosfile.c patch
>>>>>
>>>> Mmm, then it couldn't have happened on the 22nd as in your first
>>>> email.
>>>>
>>>> Also, there were two dosfile.c patches. One on the 30th September, and
>>>> the other on the 2nd October. Which one cures it ?
>>> I want to said the 22nd september build version was ok, and the 2nd
>>> October build version was not ok (report by an user I have not this
>>> versions)
>>> So I suppose the problem is between 22nd September and 2nd October.
>>>
>>> In my  test I'm return to the 30th September and I have no issue so it
>>> is on the 2nd October patch I have issue. If I understand it is in
>>> Fselect with timout value at 1, I use this value to know the kernel
>>> timeslice at start this is interesting to know to better manage event
>>> timer, but it is not my problem here.
>> I've reverted this patch for now until more investigation work is done
>> to the cause.
>>
>> But now, I've just noticed you are the maintainer of MyAES.
>>
>> Do you use Fselect(1, ...) in MyAES directly or do you use the MiNTlib
>> call select() ??
>>
>> It's probably better to switch to poll() instead now if you can.
>>
>> Alan.
>>
>>
>>
>
>
>