diff options
| author | jlam <jlam> | 2002-08-25 22:16:44 +0000 |
|---|---|---|
| committer | jlam <jlam> | 2002-08-25 22:16:44 +0000 |
| commit | a76ce2af7e5c80b17e4e8f173f41f03b55b396da (patch) | |
| tree | 5b0f321044b5af41def519faa18f0dbeb7695b07 /mk | |
| parent | 0ea65df5943de50a3e49521b1e0f6cfa9b50bc38 (diff) | |
| download | pkgsrc-a76ce2af7e5c80b17e4e8f173f41f03b55b396da.tar.gz | |
Remove estimates on increased build times as they're probably not still
accurate.
Diffstat (limited to 'mk')
| -rw-r--r-- | mk/buildlink2/buildlink2.txt | 13 |
1 files changed, 2 insertions, 11 deletions
diff --git a/mk/buildlink2/buildlink2.txt b/mk/buildlink2/buildlink2.txt index 9d409b59c4d..88e0e693518 100644 --- a/mk/buildlink2/buildlink2.txt +++ b/mk/buildlink2/buildlink2.txt @@ -1,4 +1,4 @@ -$NetBSD: buildlink2.txt,v 1.3 2002/08/25 18:51:31 jlam Exp $ +$NetBSD: buildlink2.txt,v 1.4 2002/08/25 22:16:44 jlam Exp $ The buildlink2 framework is a departure from the original buildlink framework, which tries to do its work up-front before the configure @@ -73,16 +73,7 @@ A: It should Just Work. The wrapper scripts automatically handle this Q: So what's bad about the new buildlink2 framework? A: The full build now takes longer than it used to. Since we are using wrapper scripts in place of the compilers, we bear the cost of the extra - shell processes invoked as a result. The increased build times on the - platforms on which I was able to test are roughly: - - NetBSD-1.5ZC/i386 ~1% (non-USE_LIBTOOL) - ~3% (USE_LIBTOOL) - - NetBSD-1.5.1/mac68k +?% (USE_LIBTOOL) - - The i386 box is an Intel P3 850MHz + UDMA IDE HD + 512MB SDRAM. - The mac68k box is a Quadra 650 (68040) + SCSI2 HD + 48MB RAM. + shell processes invoked as a result. Q: What happened to x11.buildlink.mk? A: x11.buildlink.mk is now deprecated, as it's all handled directly by |