Mail Archives: djgpp-workers/2004/12/23/00:43:09
On Wed, 22 Dec 2004 20:27:28 +0200, Eli Zaretskii <eliz AT gnu DOT org>
wrote:
>> Date: Tue, 21 Dec 2004 22:03:05 -0700
>> From: Brian Inglis <Brian DOT Inglis AT SystematicSw DOT ab DOT ca>
>>
>> As a first approx, embed the 2.04 symlink() routine, renamed to
>> softlink(), into the 2.03 symlink() module, and call softlink() where
>> the 2.03 symlink() returns EXDEV.
>
>In addition to this, there's this snippet at the end of `symlink' from
>v2.03:
>
> if (spawnlp (P_WAIT, STUBIFY, STUBIFY, "-g", dest_abs, (char *)0)
> || spawnlp (P_WAIT, STUBEDIT, STUBEDIT, dest_abs, ropt, (char *)0))
> return -1;
>
>Should we create a v2.04 style symlink if stubify and/or stubedit
>failed in this fragment? I don't know.
I guess some research into the failure modes and return codes of
stubify and stubedit are in order.
>Also, there were discussions in the past (check the archives of this
>list) about some tricky situations with using the stub-style symlinks.
>It would be nice to make sure those issues are covered.
Always good to have something new to read over Xmas.
>Finally, please make sure the code you suggest allows to make a v2.04
>style symlink foo -> bar when there's both bar and bar.exe in that
>directory.
The intermingling of 2.03 and 2.04 symlinks is looking more
interesting (as in the Chinese curse!)
- Raw text -