From: Daniel J. <dan...@gm...> - 2014-08-18 22:53:11
|
On Aug 18, 2014, at 7:01 AM, Hanspeter Niederstrasser <han...@sn...> wrote: > On 8/18/2014 1:47 AM, Daniel Macks wrote: >> A huge public "thank you" for doing this! >> >> While working on one of the warnings seen in: >> >> http://www.snaggledworks.com/fink/buildworld/2014-07-24/logs/io-aio-pm5124.log >> >> I found a symptom of a general problem with portability of fdatasync(). >> Compiling "succeeds", but the function doesn't actually work as >> expected. Could you grep the build logs for mentions of that string to >> see if there are other packages that need attention? > > $ grep -l "implicit declaration of function 'fdatasync' is invalid in C99" * > erlang-otp.log > file-sync-pm5123.log > file-sync-pm5124.log > file-sync-pm5162.log > io-aio-pm5123.log > io-aio-pm5124.log > io-aio-pm5162.log > libapr.0-shlibs.log > mysql-unified.log > > I tried to set up the search form to also search through the logs, but > 1) the logs are almost 10x bigger than the filelists and the search > times out, and 2) it can't handle phrases. If anyone has a suggestion > for a web accessible simple indexer that can run on a shared Debian box, > please let me know. Alternatively, can use this URL once Google decides > to crawl the new directory: > > https://www.google.com/?gws_rd=ssl#q=site:www.snaggledworks.com%2Ffink%2Fbuildworld%2F2014-07-24%2F > > Hanspeter > Yes, thanks indeed! I've fixed libapr.0-shlibs by changing fdatasync(x) to fcntl(x, F_FULLSYNC) which does the same thing. Daniel |