FloatToStr(f:n:n) problem

1
2
Page 1 of 2

allan

Benutzer

Gender: n/a
Age: 63
Posts: 11
Registered: 10 / 2016

Subject:

FloatToStr(f:n:n) problem

· Posted: 04.10.2017 - 16:50 · #1

Hi,
With new compiler (5.09.18) any attempt to convert float to string causes crash!
This does not work any more, even in simulator:

Float1 := 0.1001;
Float2 := 2.0002;
st := FloatToStr(Float1:6:4) + ' ' + FloatToStr(Float2:6:4);

Gunter

Administrator

Gender:
Location: Frankfurt Main / Germany
Posts: 1697
Registered: 02 / 2003

Subject:

Re: FloatToStr(f:n:n) problem

· Posted: 04.10.2017 - 23:20 · #2

topic.php?t=3155

allan

Benutzer

Gender: n/a
Age: 63
Posts: 11
Registered: 10 / 2016

Subject:

Re: FloatToStr(f:n:n) problem

· Posted: 05.10.2017 - 13:04 · #3

OK, understood, but this is so easy to reproduce, just copy-paste these 3 lines and try.
Something bad happened to floating point numbers while I did not update (since 2012).

Code

program avrprommer;

{$NOSHADOW}

{ $W+ Warnings}            {Warnings off}

Device = mega64, VCC=3.3;

Import SysTick, WatchDog;

From System Import  Float;

Define
        ProcClock   = 6000000;

        SysTick     = 10;            {msec}
        StackSize   = $0040, iData;
        FrameSize   = $0040, iData;

        WatchDog       = msec2000;

Implementation

{$IDATA}
{--------------------------------------------------------------}
{ Type Declarations }

type

{--------------------------------------------------------------}
{ Const Declarations }

{$EEPROM}
{--------------------------------------------------------------}

Structconst

{--------------------------------------------------------------}
{ Var Declarations }
{$IDATA}

Var

        st             : string[16];
        Float1, Float2 : float;

const

{--------------------------------------------------------------}
{ functions }

{ Main Program }
{$IDATA}

begin

  // ----- DEBUG  ----
  Float1 := 0.1001;
  Float2 := 2.0002;
  st := FloatToStr(Float1:6:4)  + '  ' + FloatToStr(Float2:6:4);


end avrprommer.

Gunter

Administrator

Gender:
Location: Frankfurt Main / Germany
Posts: 1697
Registered: 02 / 2003

Subject:

Re: FloatToStr(f:n:n) problem

· Posted: 05.10.2017 - 14:33 · #4

Hi.

I found no problem with 5.09.18

Gunter

You must be logged in or your permissions are to low to see this Attachment(s).

Administrator

Gender:
Location: Germany
Age: 24
Homepage: e-lab.de
Posts: 5558
Registered: 03 / 2002

Subject:

Re: FloatToStr(f:n:n) problem

· Posted: 05.10.2017 - 15:18 · #5

Allan, Gunter,

I also checked this example

rolf

You must be logged in or your permissions are to low to see this Attachment(s).

allan

Benutzer

Gender: n/a
Age: 63
Posts: 11
Registered: 10 / 2016

Subject:

Re: FloatToStr(f:n:n) problem

· Posted: 05.10.2017 - 17:37 · #6

I see - if you just 'Make' it works, but 'Make & Optimize' was what I did...
So it must be optimizer related problem.

Gunter

Administrator

Gender:
Location: Frankfurt Main / Germany
Posts: 1697
Registered: 02 / 2003

Subject:

Re: FloatToStr(f:n:n) problem

· Posted: 05.10.2017 - 17:53 · #7

yes you are right !
Optimize creates an endless loop.

Merlin ?

Edit:
the problem is the optimisation of the FloatToStr function !
even "st := FloatToStr (1);" creates a loop.

You must be logged in or your permissions are to low to see this Attachment(s).

Administrator

Gender:
Location: Germany
Age: 24
Homepage: e-lab.de
Posts: 5558
Registered: 03 / 2002

Subject:

Re: FloatToStr(f:n:n) problem

· Posted: 05.10.2017 - 18:56 · #8

Hello Allan,

without complete environment and other infos we can't
evaluate such problems.
As always:
doesn't work ... doesn't help

rolf

1
2
Page 1 of 2

FloatToStr(f:n:n) problem

Registered users in this topic