mdbtools @0.7 should not be built in parallel

[cooljeanius (Eric Gallager)]

So I was recently trying to install mdbtools, and got the following error:

lexer.l:22:10: fatal error: 'parser.h' file not found
updating parser.h
#include "parser.h"
         ^
1 error generated.
  CC       parser.lo
make[2]: *** [lexer.lo] Error 1
make[2]: *** Waiting for unfinished jobs....
make[2]: Leaving directory `/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_databases_mdbtools/mdbtools/work/mdbtools-0.7/src/sql'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_databases_mdbtools/mdbtools/work/mdbtools-0.7/src'
make: *** [all-recursive] Error 1
make: Leaving directory `/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_databases_mdbtools/mdbtools/work/mdbtools-0.7'
Command failed:  cd "/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_databases_mdbtools/mdbtools/work/mdbtools-0.7" && /usr/bin/make -j4 -w all 
Exit code: 2
Error: org.macports.build for port mdbtools returned: command execution failed

I tried building again immediately after that and it built perfectly fine. What I'm assuming happened was that one make job that required "parser.h" got reached before the make job that actually created "parser.h" got reached. Then, the second time around, the file was already there from the first time around. To work around this issue, I suggest that mdbtools use use_parallel_build no in its portfile.

[cooljeanius (Eric Gallager)]

darn it I just realized I got my cc and port fields mixed up...

[anddam (Andrea D'Amore)]

I cannot reproduce, I can build fine with any number of build jobs. Please run:

port clean mdbtools
port build mdbtools build.jobs=1

and confirm it builds correctly, use sudo if needed.

Then run

port clean mdbtools
port build mdbtools build.jobs=4

and confirm the build breaks.

[jmroot (Joshua Root)]

A race condition by definition won't cause a 100% reproducible problem.

[anddam (Andrea D'Amore)]

Replying to jmr@…:

A race condition by definition won't cause a 100% reproducible problem.

I understand that but I think that a clean and rebuild is the minimum test that has to be performed on the user side.

[cooljeanius (Eric Gallager)]

Replying to and.damore@…:

Replying to jmr@…:

Replying to and.damore@...:

I cannot reproduce, I can build fine with any number of build jobs. Please run:

port clean mdbtools
port build mdbtools build.jobs=1

and confirm it builds correctly, use sudo if needed.

Then run

port clean mdbtools
port build mdbtools build.jobs=4

and confirm the build breaks.

A race condition by definition won't cause a 100% reproducible problem.

I understand that but I think that a clean and rebuild is the minimum test that has to be performed on the user side.

Hm, that's strange, I can't reproduce it now anymore, either... I wonder what went wrong the first time?

[anddam (Andrea D'Amore)]

Idk but as rule of thumb a clean&build is appropriate on an error. I'm closing the ticket as worksforme but, if some other user reports issues using multiple jobs I'll disable parallel build.