After trying a succession of tests on Jacob's machine, I think I've

finally isolated the _important_ difference between Romano Trampus's
working printing.c and my failing one: he ignores the error return
from the first exploratory how-big-does-my-buffer-need-to-be call to
EnumPrinters(), because not having enough buffer space counts as an
error condition. Hence I am officially a klutz, but this should now
work. (Also reverted ENUM_LEVEL to 1, again, because that seems to
be the choice of people whose code works.)

git-svn-id: svn://svn.tartarus.org/sgt/putty@1915 cda61777-01e9-0310-a592-d414129be87e

diff --git a/printing.c b/printing.c
index ebec7ac..17bf6ff 100644 (file)
--- a/printing.c
+++ b/printing.c
@@ -11,7 +11,7 @@
  * under PRINTER_ENUM_CONNECTIONS on NT 4. I don't pretend to
  * understand this...
  */
-#if 1
+#if 0
 #define ENUM_LEVEL 5
 #define ENUM_PTR LPPRINTER_INFO_5
 #define ENUM_TYPE PRINTER_INFO_5
@@ -39,9 +39,13 @@ static char *printer_add_enum(int param, char *buffer,
 
     buffer = srealloc(buffer, offset+512);
 
-    if (EnumPrinters(param, NULL, ENUM_LEVEL, buffer+offset,
-                     512, &needed, &nprinters) == 0)
-        return NULL;
+    /*
+     * Exploratory call to EnumPrinters to determine how much space
+     * we'll need for the output. Discard the return value since it
+     * will almost certainly be a failure due to lack of space.
+     */
+    EnumPrinters(param, NULL, ENUM_LEVEL, buffer+offset, 512,
+                &needed, &nprinters);
 
     if (needed < 512)
         needed = 512;