Sometimes, the users may erase some items in delimited string.

Consider this case: say the user want to search semicolon delimited files in a semicolon delimited directories, i.e. search "1.h;2.h" in "f:\sourcecode\1;f:\sourcecode\2".

once the file is located, it should be deleted from semicolon delimited files,  the code looks like:

typedef string_tokeniser<string,char> tokeniser;
tokeniser tokendirs("f:\sourcecode\1;f:\sourcecode\2",';');
tokeniser tokenfiles("1.h",';');

tokeniser::const_iterator dir = tokendirs.being();
tokeniser::const_iterator dirEnd = tokendirs.end();

for (; dir!= dirEnd && tokenfiles.size(); ++dir)
{
    string files = join(tokenfiles);  // i.e. get semicolon delimited
string.
    string strDir(*dir);
    findfile_sequence file_seq(strDir.c_str(), files.c_str(), ';',
findfile_sequence::files);

    findfile_sequence::const_iterator foundFile = file_seq.begin();
    findfile_sequence::const_iterator foundFileEnd    = file_seq.End();
   for (; foundFile!= foundFileEnd; ++foundFile)
  {
           string sFile((*foundFile).get_filename());
           tokeniser:: const_iterator  iFound =
find(tokenfiles.begin(),tokenfiles.end(),sFile);
           if (iFound != tokenfiles.end())
              tokefiles.erase(iFound);              // this function is
helpful!!!
  }
}

corrently , I split delimited string into vector for applying this way.

thanks
chang

I'm not 100% sure what you're driving at, but I'm pretty sure it's not appropriate for a general-purpose class template like string_tokeniser. One of its important features is that it is very efficient - in fact I believe it's pretty much the fastest tokeniser available - and adding a feature to change the string being tokenised would complicate it, and likely reduce its performance characteristics.

But I'm very glad you're making good use of the libraries! ;-)



"chenchang" <baibaichen@sohu.com> wrote in message news:dn15hr$hgk$1@digitaldaemon.com...
> Sometimes, the users may erase some items in delimited string.
>
> Consider this case: say the user want to search semicolon delimited files
> in
> a semicolon delimited directories, i.e. search "1.h;2.h" in
> "f:\sourcecode\1;f:\sourcecode\2".
>
> once the file is located, it should be deleted from semicolon delimited files,  the code looks like:
>
> typedef string_tokeniser<string,char> tokeniser;
> tokeniser tokendirs("f:\sourcecode\1;f:\sourcecode\2",';');
> tokeniser tokenfiles("1.h",';');
>
> tokeniser::const_iterator dir = tokendirs.being();
> tokeniser::const_iterator dirEnd = tokendirs.end();
>
> for (; dir!= dirEnd && tokenfiles.size(); ++dir)
> {
>    string files = join(tokenfiles);  // i.e. get semicolon delimited
> string.
>    string strDir(*dir);
>    findfile_sequence file_seq(strDir.c_str(), files.c_str(), ';',
> findfile_sequence::files);
>
>    findfile_sequence::const_iterator foundFile = file_seq.begin();
>    findfile_sequence::const_iterator foundFileEnd    = file_seq.End();
>   for (; foundFile!= foundFileEnd; ++foundFile)
>  {
>           string sFile((*foundFile).get_filename());
>           tokeniser:: const_iterator  iFound =
> find(tokenfiles.begin(),tokenfiles.end(),sFile);
>           if (iFound != tokenfiles.end())
>              tokefiles.erase(iFound);              // this function is
> helpful!!!
>  }
> }
>
> corrently , I split delimited string into vector for applying this way.
>
> thanks
> chang
>
>
> 

btw, if you're doing sophisticated file-system searching, check out http://recls.org/. The docs are no better than those of STLSoft, but the library's well-tested and pretty useful. ;-)

"chenchang" <baibaichen@sohu.com> wrote in message news:dn15hr$hgk$1@digitaldaemon.com...
> Sometimes, the users may erase some items in delimited string.
>
> Consider this case: say the user want to search semicolon delimited files
> in
> a semicolon delimited directories, i.e. search "1.h;2.h" in
> "f:\sourcecode\1;f:\sourcecode\2".
>
> once the file is located, it should be deleted from semicolon delimited files,  the code looks like:
>
> typedef string_tokeniser<string,char> tokeniser;
> tokeniser tokendirs("f:\sourcecode\1;f:\sourcecode\2",';');
> tokeniser tokenfiles("1.h",';');
>
> tokeniser::const_iterator dir = tokendirs.being();
> tokeniser::const_iterator dirEnd = tokendirs.end();
>
> for (; dir!= dirEnd && tokenfiles.size(); ++dir)
> {
>    string files = join(tokenfiles);  // i.e. get semicolon delimited
> string.
>    string strDir(*dir);
>    findfile_sequence file_seq(strDir.c_str(), files.c_str(), ';',
> findfile_sequence::files);
>
>    findfile_sequence::const_iterator foundFile = file_seq.begin();
>    findfile_sequence::const_iterator foundFileEnd    = file_seq.End();
>   for (; foundFile!= foundFileEnd; ++foundFile)
>  {
>           string sFile((*foundFile).get_filename());
>           tokeniser:: const_iterator  iFound =
> find(tokenfiles.begin(),tokenfiles.end(),sFile);
>           if (iFound != tokenfiles.end())
>              tokefiles.erase(iFound);              // this function is
> helpful!!!
>  }
> }
>
> corrently , I split delimited string into vector for applying this way.
>
> thanks
> chang
>
>
>