remove trailing whitespace (#4984)

R/IDateTime.R

@@ -240,20 +240,20 @@ rep.ITime = function (x, ...)
   class(y) = "ITime"   # unlass and rep could feasibly not copy, hence use class<- not setattr()
   y
 }
-                           
-round.ITime <- function(x, digits = c("hours", "minutes"), ...) 
+
+round.ITime <- function(x, digits = c("hours", "minutes"), ...)
 {
   (setattr(switch(match.arg(digits),
                   hours = as.integer(round(unclass(x)/3600)*3600),
-                  minutes = as.integer(round(unclass(x)/60)*60)), 
+                  minutes = as.integer(round(unclass(x)/60)*60)),
            "class", "ITime"))
-} 
+}
 
-trunc.ITime <- function(x, units = c("hours", "minutes"), ...) 
+trunc.ITime <- function(x, units = c("hours", "minutes"), ...)
 {
   (setattr(switch(match.arg(units),
                   hours = as.integer(unclass(x)%/%3600*3600),
-                  minutes = as.integer(unclass(x)%/%60*60)), 
+                  minutes = as.integer(unclass(x)%/%60*60)),
            "class", "ITime"))
 }
 

R/as.data.table.R

@@ -131,7 +131,7 @@ as.data.table.list = function(x,
   eachncol = integer(n)
   missing.check.names = missing(check.names)
   origListNames = if (missing(.named)) names(x) else NULL  # as.data.table called directly, not from inside data.table() which provides .named, #3854
-  empty_atomic = FALSE  
+  empty_atomic = FALSE
   for (i in seq_len(n)) {
     xi = x[[i]]
     if (is.null(xi)) next    # eachncol already initialized to 0 by integer() above

R/data.table.R

@@ -160,7 +160,7 @@ replace_dot_alias = function(e) {
       else if (!isTRUEorFALSE(keyby))
         stop("When by and keyby are both provided, keyby must be TRUE or FALSE")
     }
-  }      
+  }
   bynull = !missingby && is.null(by) #3530
   byjoin = !is.null(by) && is.symbol(bysub) && bysub==".EACHI"
   naturaljoin = FALSE
@@ -1347,7 +1347,7 @@ replace_dot_alias = function(e) {
 
     if (is.data.table(jval)) {
       # should set the parent class only when jval is a plain data.table #4324
-      if (identical(class(jval), c('data.table', 'data.frame'))) 
+      if (identical(class(jval), c('data.table', 'data.frame')))
         setattr(jval, 'class', class(x)) # fix for #64
       if (haskey(x) && all(key(x) %chin% names(jval)) && is.sorted(jval, by=key(x)))
         setattr(jval, 'sorted', key(x))

R/setops.R

@@ -63,7 +63,7 @@ fintersect = function(x, y, all=FALSE) {
     x = shallow(x)[, ".seqn" := rowidv(x)]
     y = shallow(y)[, ".seqn" := rowidv(y)]
     jn.on = c(".seqn",setdiff(names(y),".seqn"))
-    # fixes #4716 by preserving order of 1st (uses y[x] join) argument instead of 2nd (uses x[y] join) 
+    # fixes #4716 by preserving order of 1st (uses y[x] join) argument instead of 2nd (uses x[y] join)
     y[x, .SD, .SDcols=setdiff(names(y),".seqn"), nomatch=NULL, on=jn.on]
   } else {
     z = funique(x)  # fixes #3034. When .. prefix in i= is implemented (TODO), this can be x[funique(..y), on=, multi=]

src/assign.c

@@ -156,21 +156,21 @@ static SEXP shallow(SEXP dt, SEXP cols, R_len_t n)
   SET_OBJECT(newdt, OBJECT(dt));
   IS_S4_OBJECT(dt) ? SET_S4_OBJECT(newdt) : UNSET_S4_OBJECT(newdt);  // To support S4 objects that incude data.table
   //SHALLOW_DUPLICATE_ATTRIB(newdt, dt);  // SHALLOW_DUPLICATE_ATTRIB would be a bit neater but is only available from R 3.3.0
-  
+
   // TO DO: keepattr() would be faster, but can't because shallow isn't merely a shallow copy. It
   //        also increases truelength. Perhaps make that distinction, then, and split out, but marked
   //        so that the next change knows to duplicate.
   //        keepattr() also merely points to the entire attrbutes list and thus doesn't allow replacing
   //        some of its elements.
-  
+
   // We copy all attributes that refer to column names so that calling setnames on either
   // the original or the shallow copy doesn't break anything.
   SEXP index = PROTECT(getAttrib(dt, sym_index)); protecti++;
   setAttrib(newdt, sym_index, shallow_duplicate(index));
-  
+
   SEXP sorted = PROTECT(getAttrib(dt, sym_sorted)); protecti++;
   setAttrib(newdt, sym_sorted, duplicate(sorted));
-  
+
   SEXP names = PROTECT(getAttrib(dt, R_NamesSymbol)); protecti++;
   SEXP newnames = PROTECT(allocVector(STRSXP, n)); protecti++;
   if (isNull(cols)) {

src/dogroups.c

@@ -49,7 +49,7 @@ static bool anySpecialStatic(SEXP x) {
   if (isNewList(x)) {
     if (TRUELENGTH(x)<0)
       return true;  // test 2158
-    for (int i=0; i<n; ++i) {  
+    for (int i=0; i<n; ++i) {
       if (anySpecialStatic(VECTOR_ELT(x,i)))
         return true;
     }
@@ -127,7 +127,7 @@ SEXP dogroups(SEXP dt, SEXP dtcols, SEXP groups, SEXP grpcols, SEXP jiscols, SEX
   SEXP names = PROTECT(getAttrib(SDall, R_NamesSymbol)); nprotect++;
   if (length(names) != length(SDall)) error(_("length(names)!=length(SD)"));
   SEXP *nameSyms = (SEXP *)R_alloc(length(names), sizeof(SEXP));
-  
+
   for(int i=0; i<length(SDall); ++i) {
     SEXP this = VECTOR_ELT(SDall, i);
     if (SIZEOF(this)==0)
@@ -158,12 +158,12 @@ SEXP dogroups(SEXP dt, SEXP dtcols, SEXP groups, SEXP grpcols, SEXP jiscols, SEX
   ansloc = 0;
   const int *istarts = INTEGER(starts);
   const int *iorder = INTEGER(order);
-  
-  // We just want to set anyNA for later. We do it only once for the whole operation 
+
+  // We just want to set anyNA for later. We do it only once for the whole operation
   // because it is a rare edge case for it to be true. See #4892.
   bool anyNA=false, orderedSubset=false;
   check_idx(order, length(VECTOR_ELT(dt, 0)), &anyNA, &orderedSubset);
-  
+
   for(int i=0; i<ngrp; ++i) {   // even for an empty i table, ngroup is length 1 (starts is value 0), for consistency of empty cases
 
     if (istarts[i]==0 && (i<ngrp-1 || estn>-1)) continue;

src/fifelse.c

@@ -154,7 +154,7 @@ SEXP fcaseR(SEXP na, SEXP rho, SEXP args) {
             "Note that the default argument must be named explicitly, e.g., default=0"), narg);
   }
   if (narg==0) return R_NilValue;
-  
+
   SEXP cons0 = PROTECT(eval(SEXPPTR_RO(args)[0], rho));
   SEXP value0 = PROTECT(eval(SEXPPTR_RO(args)[1], rho)); // value0 will be compared to from loop so leave it protected throughout
   SEXPTYPE type0 = TYPEOF(value0);
@@ -167,7 +167,7 @@ SEXP fcaseR(SEXP na, SEXP rho, SEXP args) {
   SEXP tracker = PROTECT(allocVector(INTSXP, len0));
   int *restrict p = INTEGER(tracker);
   copyMostAttrib(value0, ans);
-  
+
   bool nonna=!isNull(na);
   if (nonna) {
     if (xlength(na) != 1) {
@@ -194,7 +194,7 @@ SEXP fcaseR(SEXP na, SEXP rho, SEXP args) {
       }
     }
   }
-  
+
   const int n = narg/2;
   for (int i=0; i<n; ++i) {
     SEXP cons = PROTECT(i==0 ? cons0 : eval(SEXPPTR_RO(args)[2*i], rho)); // protect cons0 again for easy unprotect at the end of this loop

src/forder.c

@@ -216,7 +216,7 @@ int StrCmp(SEXP x, SEXP y)
   if (x == y) return 0;             // same cached pointer (including NA_STRING==NA_STRING)
   if (x == NA_STRING) return -1;    // x<y
   if (y == NA_STRING) return 1;     // x>y
-  return strcmp(CHAR(x), CHAR(y));  // bmerge calls ENC2UTF8 on x and y before passing here 
+  return strcmp(CHAR(x), CHAR(y));  // bmerge calls ENC2UTF8 on x and y before passing here
 }
 
 static void cradix_r(SEXP *xsub, int n, int radix)
@@ -1257,7 +1257,7 @@ SEXP issorted(SEXP x, SEXP by)
   // returning NA when NA present, and is multi-column.
   // TODO: test in big steps first to return faster if unsortedness is at the end (a common case of rbind'ing data to end)
   // These are all sequential access to x, so quick and cache efficient. Could be parallel by checking continuity at batch boundaries.
-  
+
   if (!isNull(by) && !isInteger(by)) STOP(_("Internal error: issorted 'by' must be NULL or integer vector"));
   if (isVectorAtomic(x) || length(by)==1) {
     // one-column special case is very common so specialize it by avoiding column-type switches inside the row-loop later

src/fread.c

@@ -1056,7 +1056,7 @@ static void parse_iso8601_timestamp(FieldParseContext *ctx)
       if (!args.noTZasUTC)
         goto fail;
       // if neither Z nor UTC offset is present, then it's local time and that's not directly supported yet; see news for v1.13.0
-      // but user can specify that the unmarked datetimes are UTC by passing tz="UTC" 
+      // but user can specify that the unmarked datetimes are UTC by passing tz="UTC"
       // if local time is UTC (env variable TZ is "" or "UTC", not unset) then local time is UTC, and that's caught by fread at R level too
     }
   }
@@ -2271,7 +2271,7 @@ int freadMain(freadMainArgs _args) {
             fun[abs(thisType)](&fctx);
             if (*tch!=sep) break;
             int8_t thisSize = size[j];
-            if (thisSize) ((char **) targets)[thisSize] += thisSize;  // 'if' for when rereading to avoid undefined NULL+0 
+            if (thisSize) ((char **) targets)[thisSize] += thisSize;  // 'if' for when rereading to avoid undefined NULL+0
             tch++;
             j++;
           }

src/fread.h

@@ -145,7 +145,7 @@ typedef struct freadMainArgs
   bool logical01;
 
   bool keepLeadingZeros;
-  
+
   // should datetime with no Z or UTZ-offset be read as UTC?
   bool noTZasUTC;
 

src/freadR.c

@@ -335,7 +335,7 @@ bool userOverride(int8_t *type, lenOff *colNames, const char *anchor, const int
               type[i]=CT_STRING; // e.g. CT_ISO8601_DATE changed to character here so that as.POSIXct treats the date-only as local time in tests 1743.122 and 2150.11
               SET_STRING_ELT(colClassesAs, i, tt);
             }
-          } else { 
+          } else {
             type[i] = typeEnum[w-1];                           // freadMain checks bump up only not down
             if (w==NUT) SET_STRING_ELT(colClassesAs, i, tt);
           }

src/fsort.c

@@ -154,7 +154,7 @@ SEXP fsort(SEXP x, SEXP verboseArg) {
   // TODO: -0ULL should allow negatives
   //       avoid twiddle function call as expensive in recent tests (0.34 vs 2.7)
   //       possibly twiddle once to *ans, then untwiddle at the end in a fast parallel sweep
- 
+
   union {double d; uint64_t u64;} u;
   u.d = max;
   uint64_t maxULL = u.u64;
@@ -262,7 +262,7 @@ SEXP fsort(SEXP x, SEXP verboseArg) {
       double *restrict myworking = NULL;
       // the working memory for the largest group per thread is allocated when the thread receives its first iteration
       int myfirstmsb = -1;  // for the monotonicity check
-      
+
       #pragma omp for schedule(monotonic_dynamic,1)
       // We require here that a thread can never be assigned to an earlier iteration; e.g. threads 0:(nth-1)
       // get iterations 0:(nth-1), possibly out of order, then first-come-first-served in order after that.
@@ -291,7 +291,7 @@ SEXP fsort(SEXP x, SEXP verboseArg) {
         if (myfirstmsb==-1 || msb<myfirstmsb) {
           failed=true; non_monotonic=true; continue;  // # nocov
         }
-        
+
         // Depends on msbCounts being sorted largest first before this parallel loop
         // Could be significant RAM saving if the largest msb is
         // a lot larger than the 2nd largest msb, especially as nth grows to perhaps 128 on X1.
@@ -317,7 +317,7 @@ SEXP fsort(SEXP x, SEXP verboseArg) {
       error(_("Unable to allocate working memory")); // # nocov
   }
   t[7] = wallclock();
-  
+
   // TODO: parallel sweep to check sorted using <= on original input. Feasible that twiddling messed up.
   //       After a few years of heavy use remove this check for speed, and move into unit tests.
   //       It's a perfectly contiguous and cache efficient parallel scan so should be relatively negligible.

src/fwriteR.c

@@ -174,7 +174,7 @@ SEXP fwriteR(
   )
 {
   if (!isNewList(DF)) error(_("fwrite must be passed an object of type list; e.g. data.frame, data.table"));
-  
+
   fwriteMainArgs args = {0};  // {0} to quieten valgrind's uninitialized, #4639
   args.is_gzip = LOGICAL(is_gzip_Arg)[0];
   args.bom = LOGICAL(bom_Arg)[0];

src/nqrecreateindices.c

@@ -9,7 +9,7 @@ SEXP nqRecreateIndices(SEXP xo, SEXP len, SEXP indices, SEXP nArg, SEXP nomatch)
   ans = PROTECT(allocVector(VECSXP, 2));
   SET_VECTOR_ELT(ans, 0, (newstarts = allocVector(INTSXP, n)));
   SET_VECTOR_ELT(ans, 1, (newlen = allocVector(INTSXP, n)));
-  
+
   int *inewlen = INTEGER(newlen);
   const int *iindices = INTEGER(indices);
   const int *ilen = INTEGER(len);

src/openmp-utils.c

@@ -61,7 +61,7 @@ int getDTthreads(const int64_t n, const bool throttle) {
   // this is the main getter used by all parallel regions; they specify num_threads(n, true|false).
   // Keep this light, simple and robust. initDTthreads() ensures 1 <= DTthreads <= omp_get_num_proc()
   // throttle introduced in 1.12.10 (see NEWS item); #4484
-  // throttle==true  : a number of iterations per thread (DTthrottle) is applied before a second thread is utilized 
+  // throttle==true  : a number of iterations per thread (DTthrottle) is applied before a second thread is utilized
   // throttle==false : parallel region is already pre-chunked such as in fread; e.g. two batches intended for two threads
   if (n<1) return 1; // 0 or negative could be deliberate in calling code for edge cases where loop is not intended to run at all
   int64_t ans = throttle ? 1+(n-1)/DTthrottle :  // 1 thread for n<=1024, 2 thread for n<=2048, etc

src/snprintf.c

@@ -109,7 +109,7 @@ int dt_win_snprintf(char *dest, const size_t n, const char *fmt, ...)
     strncpy(ch2, strp[i], strl[i]); // write the reordered specifers without the n$ part
     ch2 += strl[i];
     strcpy(ch2, delim); // includes '\0'
-    ch2 += NDELIM;      // now resting on the '\0'    
+    ch2 += NDELIM;      // now resting on the '\0'
   }
   char *buff = malloc(n); // for the result of the specifiers
   if (!buff) {
@@ -202,46 +202,46 @@ SEXP test_dt_win_snprintf()
 
   dt_win_snprintf(buff, 50, "%1$d %s", 9, "foo");
   if (strcmp(buff, "3 some %n$ but not all"))                                       error(_("dt_win_snprintf test %d failed: %s"), 5, buff);
-  
+
   dt_win_snprintf(buff, 50, "%%1$foo%d", 9);  // The %1$f is not a specifier because % is doubled
   if (strcmp(buff, "%1$foo9"))                                                      error(_("dt_win_snprintf test %d failed: %s"), 6, buff);
-  
+
   dt_win_snprintf(buff, 40, "long format string more than n==%d chopped", 40); // regular library (no %n$) chops to 39 chars + '/0'
   if (strlen(buff)!=39 || strcmp(buff, "long format string more than n==40 chop"))  error(_("dt_win_snprintf test %d failed: %s"), 7, buff);
-  
+
   dt_win_snprintf(buff, 40, "long %3$s %2$s more than n==%1$d chopped", 40, "string", "format"); // same with dt_win_snprintf
   if (strlen(buff)!=39 || strcmp(buff, "long format string more than n==40 chop"))  error(_("dt_win_snprintf test %d failed: %s"), 8, buff);
-  
+
   int res = dt_win_snprintf(buff, 10, "%4$d%2$d%3$d%5$d%1$d", 111, 222, 33, 44, 555); // fmt longer than n
   if (strlen(buff)!=9 || strcmp(buff, "442223355"))                                 error(_("dt_win_snprintf test %d failed: %s"), 9, buff);
   if (res!=13) /* should return what would have been written if not chopped */      error(_("dt_win_snprintf test %d failed: %s"), 10, res);
-  
+
   dt_win_snprintf(buff, 39, "%l", 3);
   if (strlen(buff)!=38 || strcmp(buff, "0 %l    does not end with recognized t"))   error(_("dt_win_snprintf test %d failed: %s"), 11, buff);
-  
+
   dt_win_snprintf(buff, 19, "%l", 3);
   if (strlen(buff)!=18 || strcmp(buff, "0 %l    does not e"))                       error(_("dt_win_snprintf test %d failed: %s"), 12, buff);
-  
+
   dt_win_snprintf(buff, 50, "%1$d == %0$d", 1, 2);
   if (strcmp(buff, "1 %0$ outside range [1,99]"))                                   error(_("dt_win_snprintf test %d failed: %s"), 13, buff);
-  
+
   dt_win_snprintf(buff, 50, "%1$d == %$d", 1, 2);
   if (strcmp(buff, "1 %$ outside range [1,99]"))                                    error(_("dt_win_snprintf test %d failed: %s"), 14, buff);
-  
+
   dt_win_snprintf(buff, 50, "%1$d == %100$d", 1, 2);
   if (strcmp(buff, "1 %100$ outside range [1,99]"))                                 error(_("dt_win_snprintf test %d failed: %s"), 15, buff);
-  
+
   dt_win_snprintf(buff, 50, "%1$d == %-1$d", 1, 2);
   if (strcmp(buff, "1 %-1$ outside range [1,99]"))                                  error(_("dt_win_snprintf test %d failed: %s"), 16, buff);
-  
+
   dt_win_snprintf(buff, 50, "%1$d == %3$d", 1, 2, 3);
   if (strcmp(buff, "5 %2$ missing"))                                                error(_("dt_win_snprintf test %d failed: %s"), 17, buff);
-  
+
   dt_win_snprintf(buff, 50, "%1$d == %1$d", 42);
   if (strcmp(buff, "2 %1$ appears twice"))                                          error(_("dt_win_snprintf test %d failed: %s"), 18, buff);
-  
+
   dt_win_snprintf(buff, 50, "%1$d + %3$d - %2$d == %3$d", 1, 1, 2);
   if (strcmp(buff, "2 %3$ appears twice"))                                          error(_("dt_win_snprintf test %d failed: %s"), 19, buff);
-  
+
   return R_NilValue;
 }

src/utils.c

@@ -64,7 +64,7 @@ bool allNA(SEXP x, bool errorForBadType) {
     return true;
   case CPLXSXP: {
     const Rcomplex *xd = COMPLEX(x);
-    for (int i=0; i<n; ++i) if (!ISNAN_COMPLEX(xd[i])) { 
+    for (int i=0; i<n; ++i) if (!ISNAN_COMPLEX(xd[i])) {
       return false;
     }
     return true;
@@ -187,7 +187,7 @@ SEXP copyAsPlain(SEXP x) {
   // For non-ALTREP this should do the same as R's duplicate().
   // Intended for use on columns; to either un-ALTREP them or duplicate shared memory columns; see copySharedColumns() below
   // Not intended to be called on a DT VECSXP where a concept of 'deep' might refer to whether the columns are copied
-  
+
   if (isNull(x)) {
     // deal with up front because isNewList(R_NilValue) is true
     return R_NilValue;
@@ -245,7 +245,7 @@ void copySharedColumns(SEXP x) {
     const SEXP thiscol = xp[i];
     savetl[i] = ALTREP(thiscol) ? 0 : TRUELENGTH(thiscol);
     SET_TRUELENGTH(thiscol, 0);
-  } 
+  }
   int nShared=0;
   for (int i=0; i<ncol; ++i) {
     SEXP thiscol = xp[i];
@@ -258,7 +258,7 @@ void copySharedColumns(SEXP x) {
                                       // 'shared' means a later column shares an earlier column
       SET_TRUELENGTH(thiscol, -i-1);  // -i-1 so that if, for example, column 3 shares column 1, in iteration 3 we'll know not
                                       // only that the 3rd column is shared with an earlier column, but which one too. Although
-                                      // we don't use that information currently, we could do in future.  
+                                      // we don't use that information currently, we could do in future.
     }
   }
   // now we know nShared and which ones they are (if any), restore original tl back to the unique set of columns