Truncate MIDI SysEx messages after termination byte

[Christian Costa]

Le 09/01/2013 19:32, Johannes Kroll a écrit :
> On Wed, 9 Jan 2013 19:04:03 +0100
> Christian Costa <titan.costa at gmail.com> wrote:
>
>> I would say consistency would be 2 spaces or none.
> You are, of course, entitled to your opinion.
If my opinion does not matter as well as code style practices we can 
found in wine
that would be better to stick to file code style. Off course you're free 
to submit
what you want.
>
>
>> Using only 1 space for assignment does not seem very common in wine AFAIK.
>> In general it's better to follow the file style or the function to avoid
>> mixing styles
>> except if you work heavyly on these files.
>>
>>
>>> If anybody sees non-style-related issues with the patch, please let me
>>> know.
>>>
>> dwBytesRecorded is supposed to tell how many bytes are valid. What's wrong
>> with it ?
> Is it? It was always zero in midiOutLongMsg when I checked. The program
> works on Windows though. This means that either dwBytesRecorded is not
> supposed to be valid in that place, possibly only when data is read from
> some MIDI device. This would seem to be consistent with the "-Recorded"
> in the variable name.
>
> Or it is supposed to be valid, which would mean that there is a bug
> somewhere else in Wine.
>
> SysEx messages where the real message length equals dwBufferSize do get
> through properly even without the patch, which means that the code in
> MMDRV_Message and onwards does not care about dwBytesRecorded. It uses
> dwBufferSize to check how many bytes it should send.
This applies only for recording which is not the case here.
After a better look the changes seem correct.
The condition i < lpMidioutHdr->dwBufferLength-1 is not necessary tough.
And a single call to MMDRV_Message would be better.

>
>
>>> Please keep any replies on the list(s).
>>>
>>>
>> Good mail client which sucks.
> This sentence makes no sense.
Yours too.
>
>
>> wine-patches is only for patch submission.
> Use -devel then. Just don't send me copies.
>
...