| Type | Bug | Status | built | Date | 9-Nov-2010 20:19 |
|---|---|---|---|---|---|
| Version | alpha 110 | Category | Math | Submitted by | Sunanda |
| Platform | All | Severity | minor | Priority | normal |
| Summary | TANGENT returns a decimal MOLD can't handle |
|---|---|
| Description |
Working close to TANGENT's outer limit: mold tangent 89.99999999999986 == "399405529304859.^T" ;; That is not a really a decimal *** Also, changing that last 6 to a 7: R2 can return a result (though wolframalpha.com says it is not very accurate): tangent 89.99999999999987 == 438274250673198.0 While R3 can't: tangent 89.99999999999987 ** Math error: math or number overflow ** Where: tangent ** Near: tangent 89.9999999999999 |
| Example code |
decimal? load mold tangent 89.99999999999986 ;; given the unusual fractional part, ;; you might expect this to fail. But it ;; works, suggesting there is an edge case ;; for LOAD or TRANSCODE in there somewhere. |
| Assigned to | n/a | Fixed in | alpha 111 | Last Update | 22-Nov-2010 12:17 |
|---|
| Comments | |
|---|---|
|
(0002814)
BrianH 9-Nov-2010 21:01 |
Is this only a problem on Windows? What about the other platforms? |
|
(0002816)
Sunanda 9-Nov-2010 21:14 |
If it is only on some platforms, then it may somehow be related to #1122 (which is still a bug, at least on Windows) |
|
(0002820)
abolka 10-Nov-2010 16:12 |
Same problem on Linux. Likely to be a problem in MOLD (as `tangent 89.99999999999986` returns a decimal!). |
|
(0002821)
BrianH 10-Nov-2010 23:00 |
Note: The nonsense character is treated as whitespace by TRANSCODE - this is why LOAD works with corrupt output. |
|
(0002909)
Ladislav 21-Nov-2010 16:09 |
in the core-tests suite |
|
(0002910)
Carl 21-Nov-2010 21:14 |
The main bug is fixed (MOLD output), BUT, there may be other issues. Test in A111, and then post new bugs if needed. |
|
(0002916)
Ladislav 22-Nov-2010 12:17 |
I am sure, that the other issue is related to the EQUAL? function and the test used:
>> equal? pi / 2 89.99999999999986 * pi / 180 == false >> equal? pi / 2 89.99999999999987 * pi / 180 == true So, this looks like the reason, why the overflow error is "fired". I do not see any problem in it, despite the fact, that the test in R2 worked differently. |
| Date | User | Field | Action | Change |
|---|---|---|---|---|
| 22-Nov-2010 12:17 | Ladislav | Comment : 0002916 | Added | - |
| 21-Nov-2010 21:14 | carl | Comment : 0002910 | Added | - |
| 21-Nov-2010 21:13 | carl | Fixedin | Modified | => alpha 111 |
| 21-Nov-2010 21:13 | carl | Status | Modified | reviewed => built |
| 21-Nov-2010 16:09 | Ladislav | Comment : 0002909 | Added | - |
| 10-Nov-2010 23:00 | BrianH | Comment : 0002821 | Added | - |
| 10-Nov-2010 22:58 | BrianH | Summary | Modified | TANGENT returns a non-decimal => TANGENT returns a decimal MOLD can't handle |
| 10-Nov-2010 22:58 | BrianH | Description | Modified | - |
| 10-Nov-2010 22:58 | BrianH | Code | Modified | - |
| 10-Nov-2010 22:58 | BrianH | Status | Modified | submitted => reviewed |
| 10-Nov-2010 22:58 | BrianH | Platform | Modified | Windows => All |
| 10-Nov-2010 16:12 | abolka | Comment : 0002820 | Added | - |
| 9-Nov-2010 21:14 | sunanda | Comment : 0002816 | Added | - |
| 9-Nov-2010 21:01 | BrianH | Comment : 0002814 | Added | - |
| 9-Nov-2010 20:19 | sunanda | Ticket | Added | - |