entries, so allocate at startup.
Use an 'int jobPipe[2]' within the job structure, and create pipes directly
into it. Common up the code that creates all the pipes - making them all
non-block on the read side in the process.
Call Job_CatchChildren() directly from Job_CatchOutput() so that it only
gets called when a child actually exits.
NB: Something causes a 'pregnant pause' if (for example) you call 'nbmake obj'
in src/tools. Introduced between netbsd 3 and 4.
messages whan make itself is suspended (ie by ^Z) before make actually
suspends, supress the messages during this sequence.
This means we don't care that they would be output after the suspend
and we can stop attempting to reap child status from withing the signal
handler (which doesn't work for recursive parallel makes).
The code simplification means that we can remove much of the code that
blocked signals - since the signal handlers (expect that for ^C and friends)
now do almost no work.
- Send each type of signal to its own handler.
- Only call JobFinish when a process exits, in particular don't 'fake up'
'exitstatus' for jobs being continued, nor call it for suspends.
- When a job is stopped, use an entire variable to remember the fact, so
we know we need to send a SIGCONT. Don't change any other state.
- In order to report '*** [job3] Suspended' before we suspend ourselves we
have to call waitpid() from the signal handler - where we don't want to
process job termination events. Save the exit status and process later.
The code now handles:
- jobs that suspend themselves
- jobs exiting while suspended
- jobs that don't actually suspend at all
Hoewever it still does printfs() from the signal handler, and I haven't yet
stopped it thrashing the signal mask.
with a table that is malloced with 'maxJobs' entries.
Add a 'job_state' field to the Job type that exactly follows which of
the old lists the job was on (or not).
Change all the code that scanned the lists to scan the array.
No logic changes in this commit.
(Soon we'll no longer need to lock out signals for the changes to job
statuses that are done from signal handlers now that there is no linked list.)
It isn't clear that it ever worked, if it did it has almost certainly
bitrotted in the last 12 years. I'm not even sure all the required
components were present.
I suspect it was written to attempt to use a 'farm' of diskless sun3s.
In any case the apparant random assignment fo jobs to other systems doesn't
actually seem like a good idea!
Things like 'distcc' han be used to help slow systems run native builds.
Removing this code also simplifies make, and should let me speed up some of
its processing - without worrying about bitrotting it further.
of trying to de-jobify the make.
You can now put .NOTPARALLEL in a submake of a recursive make (where it is
using a job-token pipe from the outer make and have it only run a single job.
You can also specify .NOTPARRALEL in the root makefile of a large recursive
make and have the submakes run multiple commands.
Add some diagnostics printfs (enabled with -dp) to the parser.
job_pipe and collect another one for the next job.
If we are aborting, remove all the 'normal' job tokens and add an 'error' one.
If we get an 'error' token, remove any other tokens, re-insert the error
token and exit (with error 'cos that is easier).
Add the current pid to some of the DEBUG(JOB) traces.
Combined effect is that parallel makes actually stop some fairly shortly
after an error, rather than running on long enough to fill the scrollback.
Using -e in sh/ksh to stop on error doesn't work with grouped commands. At
least for any SUSE compliant sh(1). Instead, extend the Shell struct and add
errOut which provides a template to use to check error status from commands.
seconds if a SIGCHLD arrived while make was not blocked in poll(),
by making the SIGCHLD handler write to a pipe included in the poll.
Avoided the need to implement a duplicate fix for the USE_SELECT case
by emulating poll() in terms of select() when USE_SELECT is defined.
Fixes bin/18895.
1. make -dx turns on DEBUG_SHELL which causes sh -x to be used where
possible.
2. PrintOnError() is now called when make is stopping due to an error.
This routine reports the curdir and the value of any variables listed
in MAKE_PRINT_VAR_ON_ERROR.
3. Variables set via command line, are propagated to child-makes via
MAKEFLAGS. This behaviour appears to be necessary for POSIX (according
to the GNU folk anyway).
4. Do not reset MAKEFILE when reading ".depend" as this rather eliminates the
usefulness of ${MAKEFILE}.
5. Added ${.newline} as a simple means of being able to include \n in the
result of a :@ loop expansion.
6. Set ${MAKE_VERSION} if defined. Need to come up with a useful value.
Reviewed: christos
use -j; all make's in a recursive build cooperate to limit the total
number of jobs, using a token-passing scheme.
The current token passing algorithm is similar to the one implemented
by gmake; there is a single pipe which is inherited through the entire
process hierarchy; tokens are obtained by reading a byte from the
"read end" of the pipe, and are returned by writing them to the "write
end". This exact algorithm is likely to change in the future.
Implementation details:
- Use the new trace facility to allow measurement of the
effectiveness of different token-passing schemes
- Get a token in MakeStartJobs(), return it in Make_Update()
- Eliminate Job_Full() and the jobFull global since they are
redundant with token system.
- Add an "internal" -J option (to pass the token pipe fd's down to
submakes) and a -T option for tracing.
- Change how compatMake is forced so that -j means something when
inherited by submakes.
- When waiting for a token, poll the token-passing pipe as well as
the output pipes of existing jobs.
poll() for parallel make:
- Make the poll() code behave more like the select() code: sleep for
a bit waiting for output rather than busy-wait (eww).
- Install a no-op SIGCHLD handler so that poll/select wake up early
(with -1/EINTR) when a child exits.
- Change the default sleep time from 500ms to 5 seconds since we now
wake up promptly when a child exits.
Unfortunately this revealed a deeper problem with the brk_string code.
To fix it:
- remove sharing of the buffer between brk_string invocations
- change the semantics of brk_string so that the argument array
starts with 0, and return the buffer where the strings are
stored
dsl