Re: Symlink badness. . .

2005-02-23 Thread Edward Ciramella
Yes - sigh - I've added all this functionality and a series of dependencies to test for a symlink and then create them as needed. I honestly think that branching every folder each time and asking people to ignore via client spec is a better solution.

Re: Symlink badness. . .

2005-02-23 Thread Matt Benson
--- Mark Lundquist <[EMAIL PROTECTED]> wrote: [SNIP] > > I'm no Ant guru, but here are some suggestions: > > — you could just punt on , and just > an ln -shf...? > > — or, use and the "unless" > attribute on something > so that you only make the link if the directory > doesn't exist... I thi

Re: Symlink badness. . .

2005-02-23 Thread Mark Lundquist
On Feb 23, 2005, at 7:54 AM, Edward Ciramella wrote: What I'm trying to do is create a symlink to a directory which may/may not exist in the current branch. So if the child branch doesn't have a particular folder, create a symlink to it prior to starting the build. It's possible for the child br

Re: Symlink badness. . .

2005-02-23 Thread Edward Ciramella
What I'm trying to do is create a symlink to a directory which may/may not exist in the current branch. So if the child branch doesn't have a particular folder, create a symlink to it prior to starting the build. It's possible for the child branch to contain this folder, but it may not always b

Re: Symlink badness. . .

2005-02-23 Thread Matt Benson
--- Edward Ciramella <[EMAIL PROTECTED]> wrote: > With overwrite set to true and failonerror set to > false, I still get > the link in the sub directory and the second time > it's run, I get the > build failed message. Okay, are all these tests based on the situation that ${basedir}/testing-si

Re: Symlink badness. . .

2005-02-23 Thread Edward Ciramella
With overwrite set to true and failonerror set to false, I still get the link in the sub directory and the second time it's run, I get the build failed message. Here is the exact line from the build file: On Feb 22, 2005, at 4:55 PM, Matt Benson wrote: --- Edward Ciramella <[EMAIL PROTECTED]> w

Re: Symlink badness. . .

2005-02-22 Thread Matt Benson
--- Edward Ciramella <[EMAIL PROTECTED]> wrote: > Hmmm, doesn't seem to matter. In my example below, > if the "test-link" > directory exists, and regardless of if it is a > symlink, ant creates > another symlink dir, one level deeper. The second > time this is run > with overwrite="true" (and

Re: Symlink badness. . .

2005-02-22 Thread Edward Ciramella
Hmmm, doesn't seem to matter. In my example below, if the "test-link" directory exists, and regardless of if it is a symlink, ant creates another symlink dir, one level deeper. The second time this is run with overwrite="true" (and with failonerror="false") I get "BUILD FAILED" and a message

Re: Symlink badness. . .

2005-02-22 Thread Matt Benson
EJ, what you are seeing, AFAICT, is the by-design behavior of the ln executable. We can debate whether allowing this is a bug in Ant, since the documentation of the symlink task makes no mention of this alternate interpretation of the link entity specified as a parameter to ln, but in the meantime

Symlink badness. . .

2005-02-22 Thread Edward Ciramella
Hello ant world - I'm attempting to create a symlink regardless if a directory exists where the symlink is or if the symlink already exists. What I'm seeing is kind of interesting. Say I'm creating a symlink from /Users/username/test-link to /Users/username/someotherdir/test-link If the "to"