Re: [Gm2] ISO IO experiment

[Gaius Mulley]

Steve Giess <address@hidden> writes:

> Dear Gaius,
>
> Thank you for your feedback. I had, from reading the library
> documentation, thought that because StrIO and FpuIO were 'Base'
> libraries they could be used with ISO libraries as well without
> problems. I presume then I should use WriteLn from STextIO to keep
> everything ISO?
>
> Regards,
>
> Steve

Hi Steve,

yes indeed - it is a bug.  But using STextIO.WriteLn should provide
a work around,

regards,
Gaius


>> From: address@hidden
>> To: address@hidden
>> Subject: Re: ISO IO experiment
>> Date: Wed, 15 Apr 2015 16:54:04 +0100
>> 
>> Steve Giess <address@hidden> writes:
>> 
>> > Dear Gaius,
>> >
>> > ... well I've tried some simple screen IO operations. However I'm
>> > having lazy I/O issues. I've attached a simple test programme
> where I
>> > input 4 real numbers and see what I get on the screen. I've found
> the
>> > ordering of the output has been less than useful. I've tried to
> use
>> > WriteLn as a 'buffer flush' but without success for that role.
> 4.1.2
>> > seemed to behave better than 4.7.4 !
>> >
>> > If you try swopping around the 4 Read/write sections in the prog
> it is
>> > possible for me to get a 'frozen' style (as if the system is
> waiting
>> > for something) screen which I have to ctrl C in order to regain
> the
>> > prompt.
>> >
>> > The build command line is:
>> >
>> > gm2 -g -fiso -I. io10.mod -lm
>> >
>> > Any thoughts please?
>> >
>> > Regards,
>> >
>> > Steve
>> 
>> Hi Steve,
>> 
>> thanks for the bug report and code. Yes I suspect the critical
> problem
>> is that both PIM and ISO library modules are being used (I/O
> modules).
>> It is a bug, but if you were to use ISO IO completely I suspect the
>> buffering bugs would disappear,
>> 
>> regards,
>> Gaius