bug in gdal-1.8.0/ogr/ogrsf_frmts/mitab/mitab_utils.cpp

[iskander]

Dear

we are using GDAL 1.8.0 and we found a bug in this method when reading MIF file that has basename with '.' (e.g. shape.001.MIF)

file location gdal-1.8.0/ogr/ogrsf_frmts/mitab/mitab_utils.cpp

///code of the method

/**********************************************************************
 *                       TABGetBasename()
 *
 * Extract the basename part of a complete file path.
 *
 * Returns a newly allocated string without the leading path (dirs) and
 * the extenstion.  The returned string should be freed using CPLFree().
 **********************************************************************/
char *TABGetBasename(const char *pszFname) {
    const char *pszTmp = NULL;

    /*-----------------------------------------------------------------
     * Skip leading path or use whole name if no path dividers are
     * encountered.
     *----------------------------------------------------------------*/
    pszTmp = pszFname + strlen(pszFname) - 1;
    while ( pszTmp != pszFname 
            && *pszTmp != '/' && *pszTmp != '\\' ) 
        pszTmp--;

    if( pszTmp != pszFname )
        pszTmp++;

    /*-----------------------------------------------------------------
     * Now allocate our own copy and remove extension
     *----------------------------------------------------------------*/
    char *pszBasename = CPLStrdup(pszTmp);
    int i;
 
 //>>>> your old code
 //for(i=0; pszBasename[i] != '\0'; i++)  //>>>>
 
 //>>>> our modification 
    for(i=strlen(pszBasename)-1; i >= 0; i--)  //>>>>
    {
        if (pszBasename[i] == '.')
        {
            pszBasename[i] = '\0';
            break;
        }
    }

    return pszBasename;
}

we replaced " for(i=0; pszBasename[i] != '\0'; i++)" by this " for(i=strlen(pszBasename)-1; i >= 0; i--)" and it worked TABGetBasename() doesn't support basename with . It should look for the '.' in the basename starting from the basename end

Please rectify this minor bug

Best Regards Iskander Benhadj

[warmerdam]

I'll address this in GDAL and upstream in mitab. Daniel let me know if there is something that looks wrong about the fellows proposed solution.

[dmorissette]

Looks fine to me. Thanks for handling it.

[warmerdam]

Change applied upstream in MITAB and pulled down into trunk (r22548) and 1.8 (r22549).

Thanks!