Page 153 of 211
Re: 43S News
Posted: Thu Jan 13, 2022 6:32 pm
by dlachieze
H2X wrote: ↑Thu Jan 13, 2022 1:59 pm
And if I am not mistaken, this is also exactly how it would appear in any other code or text editor with line numbers enabled. Correct, dlachieze?
Yes.
Here is an issue with ROUNDI : this instruction applies currently only to reals, not to integer. This is an issue because it means that when the user has to enter data in a program including the ROUNDI instruction he has also to specifically enter numbers with a decimal separator. Otherwise there could be an error message "Invalid input type for this operation" if ROUNDI is applied to the result of a calculation that happens to be integer.
For example, let's look at the simple program below just for illustration:
Code: Select all
001 LBL 'TST'
002 RCL+X
003 SQRT
004 ROUNDI
005 END
If you enter 1 XEQ 'TST' it will return correctly 1.
If you enter 3 XEQ 'TST' it will return correctly 2.
But if you enter 2 XEQ 'TST' you'll get an "Invalid input type for this operation" error message because 2+2=4 and the square root of 4 is 2 which is considered as an integer to which ROUNDI cannot be applied.
The only way to get a result for 2 is to enter 2. XEQ 'TST' so SQRT(2.+2.) will return 2. which is a valid type for ROUNDI.
For consistency and ease of use ROUNDI should not return an error when applied to integers (same as IP which works correctly with integers).
Re: 43S News
Posted: Thu Jan 13, 2022 6:48 pm
by Walter
dlachieze wrote: ↑Thu Jan 13, 2022 6:32 pm
... just for illustration:
Code: Select all
001 LBL 'TST'
002 RCL+X
003 SQRT
004 ->REAL @This step was inserted by me
005 ROUNDI
006 END
My step 004 solves your problem as well.
Re: 43S News
Posted: Thu Jan 13, 2022 7:33 pm
by dlachieze
Well, this is a workaround but if you have to systematically add ->REAL before ROUNDI to avoid unneeded error messages, my take is that it should be part of the ROUNDI instruction itself.
Why would 2 ROUNDI return an error message and not 2 IP ?
Re: 43S News
Posted: Thu Jan 13, 2022 7:41 pm
by Over_score
dlachieze wrote: ↑Thu Jan 13, 2022 7:33 pm
Why would 2 ROUNDI return an error message and not 2 IP ?
1+
Re: 43S News
Posted: Thu Jan 13, 2022 7:42 pm
by akaTB
dlachieze wrote: ↑Thu Jan 13, 2022 7:33 pm
Why would 2 ROUNDI return an error message and not 2 IP ?
The answer is to modify IP so it issues an error too.
SCNR
Re: 43S News
Posted: Fri Jan 14, 2022 10:45 am
by Walter
Release 0.21.8 is here
https://gitlab.com/Over_score/wp43s/-/releases. Santa Claus has come again (very early):
- Program editing is refined following user request.
- ROUNDI digests also integers now without throwing an error.
- Improvements of the Solver were implemented.
- Handling of local registers was improved.
- SHUFFLE (a.k.a. <>) can be inserted in a program now.
- Further refinements in program entry.
For this release, no updated manuals were uploaded.
Included in the release package is everything you may want to try, test, and check the 43S on your computer or your DM42. Please look into the
Reference Manual,
App. E and
F for how to do this.
Enjoy!
If you have any questions, don't hesitate to ask early. And if you discover anything suboptimal unknown yet, please report. There are certainly some issues - who finds one will get 12 test points.
Re: 43S News
Posted: Fri Jan 14, 2022 1:10 pm
by dlachieze
Cool, this is a quick bug fix turn around time
I will check it in details.
In the mean time here is an issue I found with 0.21.7 and confirmed with 0.21.8: RCL operations with L are using the future value of L instead of the current one.
For example
5 x^2 RCL+ L returns
50 instead of
30 .
Re: 43S News
Posted: Fri Jan 14, 2022 3:19 pm
by Walter
This is a cute one!
12 more points for you (you may need an account, don't you?).
As written in the foreword, we strive for short maintenance times.
(EDIT 15:56: Repaired. Will be available with next release.)
Re: 43S News
Posted: Fri Jan 14, 2022 4:36 pm
by toml_12953
dlachieze wrote: ↑Fri Jan 14, 2022 1:10 pm
Cool, this is a quick bug fix turn around time
I will check it in details.
In the mean time here is an issue I found with 0.21.7 and confirmed with 0.21.8: RCL operations with L are using the future value of L instead of the current one.
For example
5 x^2 RCL+ L returns
50 instead of
30 .
I was able to run 0.21.7 with no problem but when I try to run 0.21.8 on Win 11, the black window opens with a message about the code page, then a white calculator-sized window opens then both close with no error messages.
Re: 43S News
Posted: Fri Jan 14, 2022 5:32 pm
by Walter
toml_12953 wrote: ↑Fri Jan 14, 2022 4:36 pm
I was able to run 0.21.7 with no problem but when I try to run 0.21.8 on Win 11, the black window opens with a message about the code page, then a white calculator-sized window opens then both close with no error messages.
Sorry, I work with Windows 10. We feel sympathetic to you but:
Never change a solid OS unless inevitable (fundamental law #1 of computing).
