It works, finds just x and goes happily on its way past it as it should.
cd ..\dbm
for %d in ( x) do if not %d == x NMAKE -f apr_dbm_%d.mak
CFG="apr_dbm_%d - Win32 Release" RECURSE=0
C:\build\httpd-2.2.x-dev\srclib\apr-util\dbm>if not x == x NMAKE -f
apr_dbm_x.mak CFG="apr_dbm_x - Win32 Release" RECURSE=0
cd ..\..\..
cd modules\ssl
I am a bit less concerned why it worked fine before for you and not for
me since when DBD was added to the build, there was a check for an
undefined list before it would let it into the loop that was taken out
when Tom added the odbc to the build. So this obviously tripped others
before, and it has now reared it's head again but now in the DBM
section. As I look at R664253 it was you that did that ;-) most likely
just to skip going into the loop unnecessarily I guess but also it kept
this problem from showing cause I know it is not just me this is
happening to this time.
It is strange tho because at the command line it does not do this.
C:\build\httpd-2.2.x-dev> for %d in () do echo "hello world"
C:\build\httpd-2.2.x-dev>
I thought I had tried a command line build as well as IDE, I'm sure of
it. Newbies (those 'spoiled folks' who eventually do get guts and decide
to try compiling) always use the IDE so this keeps this from happening
to them if it is just an IDE problem. Oh course, on VC9 w/ just the
included Express SDK only, it'll go down before that at odbc, but so
will mod_deflate, mod_ssl and abs and I see no way around that. I'm not
a fan of downloading and installing a pretty much unneeded 1.2G of stuff
just for devenv, I just add them to the buildbin project skipping the
check cause I know support is there and it saves time. On that note, in
Windows, couldn't the odbc project be added to buildbin (or the normal
make run) instead of going out to the makefile now? not that it matters
much.
Gregg
William A. Rowe, Jr. wrote:
Gregg L. Smith wrote:
William A. Rowe, Jr. wrote:
+/-1
[+1] Release httpd-2.2.13 as GA
This is looking fine on windows, is up at /dev/dist as -src-win32,
and binaries will follow in the morning.
non-binding +9/10 on Windows. Please see PR 47659 for the last 1/10.
This is only a build process failure due to a small oversight in the
makefile.
I offered an alternative solution (one based on too many /bin/sh for
loops :-) Could you test 2.2 branch or even just apr-util/1.3.x branch
so we have this right in the next release?
I'm still puzzled, since my modern cmd.exe shells never complain about
empty for %x in () do loops.
