fwrite(): when col.names=FALSE, produce a gzip header, avoid a leak (#6853)

* fwrite(): emit gzip header even if !col.names

Previously, fwrite() emitted a raw DEFLATE stream if col.names was FALSE
(and thus headerLen was zero). Decompressing such files is possible, but
not trivial.

* Avoid a memory leak in fwrite(compress="gzip")

Previously, the zlib object zstrm was initialized whenever compression
was enabled, at least in order to calculate the buffer size for the
compressed data stream, but was only used for compression and
subsequently deallocated only when there was a header with column names.

Make sure to deflateEnd(&strm) even if there was no header.

* NEWS entry

* comment the source of this #endif for clarity

* Clean up the test file

Co-Authored-By: Michael Chirico <[email protected]>

* Report write() errno instead of just plain -1

Co-Authored-By: Michael Chirico <[email protected]>

---------

Co-authored-by: Michael Chirico <[email protected]>

Author: aitap

NEWS.md

@@ -10,6 +10,8 @@
 
 1. Custom binary operators from the `lubridate` package now work with objects of class `IDate` as with a `Date` subclass, [#6839](https://github.com/Rdatatable/data.table/issues/6839). Thanks @emallickhossain for the report and @aitap for the fix.
 
+2. `fwrite(compress="gzip")` once again produces a gzip header when the column names are missing or disabled, [@6852](https://github.com/Rdatatable/data.table/issues/6852). Thanks @maxscheiber for the report and @aitap for the fix.
+
 ## NOTES
 
 1. Continued work to remove non-API C functions, [#6180](https://github.com/Rdatatable/data.table/issues/6180). Thanks Ivan Krylov for the PRs and for writing a clear and concise guide about the R API: https://aitap.codeberg.page/R-api/.

inst/tests/tests.Rraw

@@ -10024,10 +10024,13 @@ if (!haszlib()) {
   fwrite(DT, file=f3<-tempfile(), compress="gzip")   # compress to filename not ending .gz
   fwrite(DT, file=f4<-tempfile(), compress="gzip", compressLevel=1)   # test compressLevel
   fwrite(DT, file=f5<-tempfile(), compress="gzip", compressLevel=9)
+  # col.names=FALSE must not disable gzip header, #6852
+  fwrite(DT, file=f6<-tempfile(), compress="gzip", col.names=FALSE)
   test(1658.441, file.info(f3)$size, file.info(f1)$size)
   test(1658.442, file.info(f4)$size >= file.info(f1)$size)
   test(1658.443, file.info(f1)$size >= file.info(f5)$size)
-  unlink(c(f1,f2,f3,f4,f5))
+  test(1658.444, fread(f6, col.names = c("a", "b")), DT)
+  unlink(c(f1,f2,f3,f4,f5,f6))
 }
 DT = data.table(a=1:3, b=list(1:4, c(3.14, 100e10), c("foo", "bar", "baz")))
 test(1658.45, fwrite(DT), output=c("a,b","1,1|2|3|4","2,3.14|1e+12","3,foo|bar|baz"))

src/fwrite.c

@@ -829,9 +829,28 @@ void fwriteMain(fwriteMainArgs args)
            zbuffSize / MEGA, nth, errno, strerror(errno));
       // # nocov end
     }
-  }
+    len = 0;
+    crc = crc32(0L, Z_NULL, 0);
 
-#endif
+    if (f != -1) {
+      // write minimal gzip header, but not on the console
+      static const char header[] = "\037\213\10\0\0\0\0\0\0\3";
+      int ret0 = WRITE(f, header, (sizeof header) - 1);
+      compress_len += (sizeof header) - 1;
+
+      if (ret0 == -1) {
+        // # nocov start
+        int errwrite = errno; // capture write errno now in case close fails with a different errno
+        CLOSE(f);
+        free(buffPool);
+        free(zbuffPool);
+        deflateEnd(&strm);
+        STOP(_("Failed to write gzip header. Write returned %d"), errwrite);
+        // # nocov end
+      }
+    }
+  }
+#endif // #NOZLIB
 
   // write header
 
@@ -873,15 +892,10 @@ void fwriteMain(fwriteMainArgs args)
       *ch = '\0';
       DTPRINT("%s", buff); // # notranslate
     } else {
-      int ret0=0, ret1=0, ret2=0;
+      int ret1=0, ret2=0;
 #ifndef NOZLIB
       if (args.is_gzip) {
         char* zbuff = zbuffPool;
-        // write minimal gzip header
-        char* header = "\037\213\10\0\0\0\0\0\0\3";
-        ret0 = WRITE(f, header, 10);
-        compress_len += 10;
-        crc = crc32(0L, Z_NULL, 0);
 
         size_t zbuffUsed = zbuffSize;
         len = (size_t)(ch - buff);
@@ -898,21 +912,26 @@ void fwriteMain(fwriteMainArgs args)
 #ifndef NOZLIB
       }
 #endif
-      if (ret0 == -1 || ret1 || ret2 == -1) {
+      if (ret1 || ret2 == -1) {
         // # nocov start
         int errwrite = errno; // capture write errno now in case close fails with a different errno
         CLOSE(f);
         free(buffPool);
 #ifndef NOZLIB
         free(zbuffPool);
 #endif
-        if (ret0 == -1) STOP(_("Failed to write gzip header. Write returned %d"), ret0);
-        else if (ret1) STOP(_("Failed to compress gzip. compressbuff() returned %d"), ret1);
+        if (ret1) STOP(_("Failed to compress gzip. compressbuff() returned %d"), ret1);
         else STOP(_("%s: '%s'"), strerror(errwrite), args.filename);
         // # nocov end
       }
     }
   }
+#ifndef NOZLIB
+  else {
+    // was unconditionally initialized for zbuffSize, not used for header
+    deflateEnd(&strm);
+  }
+#endif
   if (verbose)
     DTPRINT(_("Initialization done in %.3fs\n"), 1.0*(wallclock()-t0));