http://www.purebasic.com
https://www.purebasic.fr/english/
I wonder how that should work if there is a backslash right in the code.Bisonte wrote:Also from TS-Soft (this should work on all platforms) slightly modified :
Code: Select all
tmp + "\"
Code: Select all
;-! OS_PATH_SEPARATOR
; Not #OS_PATH_SEP$ due to ;BUG; PB v573 does not support @#OS_PATH_SEP$
CompilerIf #PB_Compiler_OS = #PB_OS_Windows
Global.s OS_PATH_SEP$ = #PS$
CompilerElse
Global.s OS_PATH_SEP$ = #NPS$
CompilerEndIf
Macro PB_CP
;ProgramFilename() + #COLON2$ + #PB_Compiler_File + #COLON2$ + #PB_Compiler_Procedure
GetFilePart(#PB_Compiler_File, #PB_FileSystem_NoExtension) + "::" + #PB_Compiler_Procedure
EndMacro
Macro MRE(txt) ; MessageRequester Error = #MB_ICONHAND | #MB_ICONSTOP | #MB_ICONERROR
MessageRequester(PB_CP, txt, #MB_ICONERROR)
EndMacro
ImportC ""
CompilerIf #PB_Compiler_Processor = #PB_Processor_x64
_wcslen_(*String) As "wcslen"
_wcsstr_(*String1, *String2) As "wcsstr"
CompilerElse
_wcslen_(*String) As "_wcslen"
_wcsstr_(*String1, *String2) As "_wcsstr"
CompilerEndIf
EndImport
Procedure.i SplitC(*String.Character, Array A$(1), *Delimiter)
;nco2k, http://www.purebasic.fr/english/viewtopic.php?p=516082#p516082
;skywalk; modified to catch trailing null's.
; Ex. "1,2,3,," Splits with ',' -> [1][2][3][][].
Protected.i *StringEnd, DelimiterLength, ArrayIndex
Protected.i Resize = 511
Protected.i nStrings = ArraySize(A$())
If *String ; Avoid 'Pointer is null' if pointing to an array element = Empty$
If *String\c
DelimiterLength = _wcslen_(*Delimiter) * SizeOf(Character)
If DelimiterLength
Repeat
*StringEnd = _wcsstr_(*String, *Delimiter)
If *StringEnd
A$(ArrayIndex) = PeekS(*String, (*StringEnd - *String) / SizeOf(Character))
ArrayIndex + 1
If ArrayIndex > nStrings
nStrings + Resize
ReDim A$(nStrings)
EndIf
*String = *StringEnd + DelimiterLength
Else
Break
EndIf
ForEver
EndIf
ReDim A$(ArrayIndex)
A$(ArrayIndex) = PeekS(*String) ; Catch final trailing null's if any.
ArrayIndex + 1
EndIf
EndIf
ProcedureReturn ArrayIndex
EndProcedure
Procedure.i SF_HasBadChar(s$)
; Verify that string does not contain illegal characters, since it will be used for a filename.
; such as: /, \, :, *, ?, ", <, >, | ; Windows 10
Protected Dim A$(10)
Protected.i i, ri
A$(0) = #PS$ ;"\"
A$(1) = #NPS$ ;"/"
A$(2) = ":"
A$(3) = "*"
A$(4) = "?"
A$(5) = #DQUOTE$
A$(6) = "<"
A$(7) = ">"
A$(8) = "|"
For i = 0 To 8
If FindString(s$, A$(i))
ri = 1
Break
EndIf
Next i
ProcedureReturn ri
EndProcedure
Procedure.i FL_CreateFullDirPath(DirPath$)
; REV: 100405, skywalk
; REV: 180325, skywalk
; Bug in FileSize() for existing path queried with a bad char?
; Allows wildcards in its search: ?,*, including '<','>'.
; No recursion.
; Windows rejects certain characters within pathname:
; / \ : * ? " < > |
Protected.i i, k, ri
CompilerIf 0
; Check if path exists?
If FL_FileExists(DirPath$) = -1
ProcedureReturn 1
Else
ri = CreateDirectory(DirPath$)
If FL_FileExists(DirPath$) = -1
ProcedureReturn 1
EndIf
EndIf
CompilerEndIf
Protected.s TestPath$
Protected Dim tmppath$(0)
k = SplitC(@DirPath$, tmppath$(), @OS_PATH_SEP$) ;BUG; PB v573 does not support @#OS_PATH_SEP$
If k > 1
TestPath$ = tmppath$(0)
i = 1 ; Do not allow creation of 1st dir entry, "c:\"
Repeat
If SF_HasBadChar(tmppath$(i)) = 0
TestPath$ = TestPath$ + OS_PATH_SEP$ + tmppath$(i)
ri = CreateDirectory(TestPath$)
Delay(5)
If (ri = 0) And (i = k)
i = 0
MRE("Error creating dir structure:" + #CRLF$ + DirPath$ + #CRLF$ + "Verify Write access to drive.")
Break
EndIf
i + 1
Else
i = 0
MRE("Error creating dir structure:" + #CRLF$ + DirPath$ + #CRLF$ + "May contain invalid character.")
Break
EndIf
Until i = k
ProcedureReturn i
Else
ProcedureReturn i
EndIf
EndProcedure
Debug FL_CreateFullDirPath("c:\try3\z1\z2\z3\z4\z5")
Debug FL_CreateFullDirPath("c:/try3/z1/z2/z3/z4/")
Debug FL_CreateFullDirPath("c:\try3\z\zz\|txt")
;FL_CreateFullDirPath("c:\try3\z\z*z\")
;FL_CreateFullDirPath("c:\try2\try 3\try 4\")Code: Select all
CompilerIf #PB_Compiler_OS = #PB_OS_Windows
Global.s OS_PATH_SEP$ = #PS$
CompilerElse
Global.s OS_PATH_SEP$ = #NPS$
CompilerEndIf
Constants like #PB_Compiler_#PB_Compiler_Debugger change so why not one that changes with the OS used?skywalk wrote:Hmmm, I needed a non constant string variable for @path$. @#PS$ is not allowed.
I don't think the contents of #PS$ change per OS, else there would only be 1 constant. No?
I agree the name is a little vague. I'm guessing it was to keep it short and as terse as possible, similar to the #LF$, #CR$, #CRLF$ constants.skywalk wrote:Just a hunch, cause there are not 2 #debugger constants like #PS and #NPS.
Why is there no #PB_OS_PATH_SEP$?
No. That's why #PS$ was introduced, to work on all OSes. It's in the Help:skywalk wrote:I don't think the contents of #PS$ change per OS, else there would only be 1 constant. No?
Code: Select all
Added: #PS, #NPS, #PS$ and #NPS$ constants (Path seperator character depending of the OS)There is - just what I said above.skywalk wrote:Why is there no #PB_OS_PATH_SEP$?
Code: Select all
Debug #PS$
Debug #NPS$
NicTheQuick wrote:Consider the following code:On Linux the result is:Code: Select all
Debug #PS Debug #NPS Debug #PS$ Debug #NPS$On Windows the result is:Code: Select all
47 92 / \Code: Select all
92 47 \ /
I didn't write the manual, so I don't know. All I know is that #PS$ stands for path separator, but I don't know what the N in #NPS$ would mean. The manual doesn't say (as I quoted it).skywalk wrote:Dude? You are not explaining why there is #PS$ and #NPS$????
If the compiler changes the value per OS, why the 2 constants?
The procedure "SF_HasBadChar" is wrong on Linux. On Linux the only character which is not allowed for file names is the forward slash or path separator.skywalk wrote:This is what I use, and it is not recursive.Code: Select all
...