Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
|
|
The scanformat specification will be converted to a new Jim_Obj of type
scanFormatStringObjType, that will contain the parsed representation within
its internal object representation. This speed up multiple scanning within
e.g. a loop, of objects were cached.
For internal scanning we use sscanf currently (I am lazy right now). That
means also, we will inherit its incapability to handle string with embedded
ZERO. It would be not too difficult to implement another scanner just for
the string and charset conversion type that could be able to handle those
embedded ZEROs, however.
Furthermore two small details were fixed:
1. Jim_DoubleToString should also recognize a number if a leading '+' or
'-' occured. By recognizing I mean, add a ".0" to such a number.
2. Jim_StrDupLen should also properly handle duplication of substrings. So
now it should be possible to do this:
const char *str1 = "This is a long string";
char *substr1 = Jim_StrDupLen(str1, 4);
Now substr1 should contain a properly ZERO ended "This".
|
|
|
|
|
|
Added a regression test to test.tcl
(I plan to move everything inside regtest.tcl into test.tcl).
|
|
|
|
Jim_GetIndex() semantic changed to make [lindex] more compatible
with Tcl's one, without effects on other commnads (me).
|
|
|
|
performaces of Tcl8.4 for the specialized forms. The implementation
is a bit complex so may contain bugs... to handle with care.
Also a [for] bug about [continue] was fixed and the regression test added.
|
|
avoid problems with -command and shimmering of the objects
passed as [switch] arguments.
|
|
in the BUG file about jim.libpath lookup performed using the local
variable API that's now fixed. In order to fix the problem the new
API call Jim_GetGlobalVariableStr() was added.
Jim_GetVariableStr() and Jim_GetGlobalVariableStr() are now exported.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Also a patch about empty string and StringAppendString
causing memory corruption, and a patch about Jim_WrongNumArgs
that I modified because caused other tests to fail.
|
|
|
|
|
|
|
|
More tests. A fix in the parsing code for line number calculation.
|
|
|
|
|