#
f11debc0 |
|
30-Oct-2019 |
Michael Norrish <Michael.Norrish@nicta.com.au> |
Holmake: key monitor function on combination of PID and tag-string The latter was probably unique when the tags were the filenames that were either from the current directory, or had paths attached. But if the printing form is the tag, which means the path is always stripped, multiple files of the same name may be being built in parallel. Then they can be distinguished by the PID of the process that is building them.
|
#
737a28c9 |
|
25-Oct-2019 |
Michael Norrish <Michael.Norrish@nicta.com.au> |
Make sure hollog files are stored under right directory Also normalise theory log files to always name without the .dat/sig/sml suffix. Closes #607 (as a side effect). With test.
|
#
3a08e8cb |
|
16-Oct-2019 |
Michael Norrish <Michael.Norrish@nicta.com.au> |
Progress: now some recursive building happens occasionally
|
#
7485218f |
|
18-Apr-2017 |
Michael Norrish <Michael.Norrish@nicta.com.au> |
Make multibuild Holmake pass on reloc environment variable
|
#
20eb6cc0 |
|
21-Apr-2016 |
Michael Norrish <michael.norrish@nicta.com.au> |
Make parallel Holmake's output prettier When a process fails, a chunk of its output up to the failure point is output to the main screen. Also, a process that is killed because of a general desire to give up (a failure elsewhere and no -k option, say), now gets a red M-KILLED ("monitor killed") report.
|
#
18d1c2f7 |
|
20-Apr-2016 |
Michael Norrish <Michael.Norrish@nicta.com.au> |
Enable multiplexed builds for Poly/ML Output is horrendous at the moment
|
#
9ee1eb0b |
|
19-Apr-2016 |
Michael Norrish <Michael.Norrish@nicta.com.au> |
Implement monitor cmds (e.g., kill) for Holmake These are the commands that the client sends to the running worklist via the callbacks. It even makes sense to allow for a GiveUpAndDie when the client is generating a fresh job via the genjob callback.
|
#
63e469ab |
|
18-Apr-2016 |
Michael Norrish <Michael.Norrish@nicta.com.au> |
Move ProcessMultiplexor to tools/Holmake/poly
|