[4.6,patch,libgfortran/io] Fix build issue with _commit [PR5016]

[Tobias Burnus]

Pre-remark: Jakub (4.6.2 RM) had no objects - and thus I have committed 
it to the branch.

Steve Kargl wrote:
> I don't have an objection, just a question. Does this issue appear on 
> trunk?

Yes and no. The MinGW build issue this patch fixes was unique to the branch.

However, the actual issue - the slow down - exists both on the 4.6 
branch and on the trunk. The problem is that Janne and I couldn't agree 
when to call _commit(). We solved that by deferring it to later for the 
trunk (still time before the 4.7.0 release) while we wanted to have some 
fix for 4.6.2 - thus, we agreed on some fix for 4.6. The fix on the 
branch now calls significatly less often _commit() than before - but it 
probably still calls it too often.

Regarding rushing with the fix: I agree that is was less optimal; the 
main problem is that we do not have any MinGW gfortran developer who can 
properly test such patches and who can detect such regressions more 
timely. (The performance regression dates back over a year.)

Tobias