2 things

Mon Aug 2 04:17:00 GMT 1999

Firstly, is it possible for libgcj to affect the compiler ? I just
installed a CVS version of libgcj, and now find I can't compile
because cc1 and jc1 both reply with:
jc1: Invalid option `-fno-use-divide-subroutine'
It's possible someone has installed a new version of egcs which is causing
the problem, but I don't think so (there's nothing on the internal update
list)
Also, here's a (very poor!) patch for Runtime.exec() which takes into 
account a single level of quoting in a passed-in execution string. So
you can call exec("find /tmp -name xxx -printf '%p %AH'") for example.
It really needs a lot more to make it robust - chack for escaped quotes, 
multiple level of quotes etc. Alternatively, just pass strings straight
through and create a single string from the array-type call. I did think
of doing this, but unfortunately I have a deadline to meet :-(
[simon@donerkebab lang]# diff Runtime.java /tmp/Runtime.java
39,45c39
< boolean inQuotes = false; // Whether we're aggregating tokens
< String quotedText = ""; // Text within quotes
< 
< // Use the string tokeniser to split based on whitespace, but ask
< // for the tokens so we can process quoted text as a single entity
< String delimiters = " \t\n\r'\"";
< StringTokenizer st = new StringTokenizer(prog, delimiters, true);
---
> StringTokenizer st = new StringTokenizer(prog);
47,50d40
< 
< // Process each of the tokens from the tokeniser. Should probably
< // be done using StringBuffers.
< int count = 0;
52,97c42,43
< {
< String token = st.nextToken ();
< 
< if (token.length() == 1 && delimiters.indexOf(token) >= 0)
< {
< if (token.equals("\"") || token.equals("'"))
< {
< inQuotes = ! inQuotes;
< if (!inQuotes)
< {
< a[count++] = quotedText;
< quotedText = "";
< }
< }
< else 
< {
< if (inQuotes)
< {
< quotedText += token;
< }
< }
< }
< else
< {
< if (!inQuotes)
< {
< a[count++] = token;
< }
< else
< {
< quotedText += token;
< }
< }
< }
< 
< // Copy the a[] array to a new array (b[] :-) which only contains
< // 'count' entries rather than st.countTokens()
< 
< String b[] = new String[count];
< for (int i=0; i<count; i++)
< {
< b[i] = a[i];
< }
< 
< 
< return exec (b, envp);
---
> a[i] = st.nextToken ();
> return exec (a, envp);