Simon Michael
ee276afc56
ci: fall back on checking last 5 commit messages, instead of 20
...
This will allow CI status to recover faster when a bad commit message
is accidentally merged. In some cases (eg force pushes of PRs with
more than 5 commits) not all commit messages will be checked at all
times.
2021-11-14 15:49:04 -10:00
Simon Michael
6f96d21a3e
ci: fix windows, make bash requirement explicit
2021-08-29 08:10:09 -10:00
Simon Michael
07eb3a9086
ci: tweak
2021-08-19 08:55:44 -10:00
Simon Michael
cf95fc7abf
ci: tweak
2021-08-18 21:47:40 -10:00
Simon Michael
90612c1444
ci: tweak
2021-08-18 18:05:02 -10:00
Simon Michael
818e098b04
ci: handle another PR failure case; more debug output
...
"Skip remaining CI steps" now also handles git log failure, continuing on instead of failing. Getting pretty gnarly.
2021-08-18 17:39:46 -10:00
Simon Michael
df6667232f
ci: if git log fails, don't skip steps and succeed
...
Eg git log fails after force-push to a PR right now.
2021-08-17 22:05:29 -10:00
Simon Michael
a291f5d33b
ci: push/linux/mac/windows: fix ; conditional step execution
2021-07-20 23:24:18 -10:00
Simon Michael
88c6d47e32
ci: linux/mac/windows: skip steps if message begins with ;
2021-07-20 23:01:57 -10:00
Simon Michael
41d432c30a
ci: mac: use ghc 9.0
2021-07-20 20:38:39 -10:00
Simon Michael
14e1efc843
stack: 9.0: rename yaml file, making 9.0 the default
2021-07-20 20:38:39 -10:00
Simon Michael
82b74ec41e
ci: commitlint: when range is invalid (eg force push), check last 20
...
This is imperfect but better than nothing, and should avoid spurious
failures on force pushes to PRs.
2021-07-13 12:35:48 -10:00
Simon Michael
01f9c70545
ci: run commitlint to check commits ( #1602 )
...
As first step in our main "check" workflows (push, linux, mac,
windows), check all the commit messages with commitlint.
These workflows can be triggered in various ways:
pull requests, pushes, manually, or scheduled.
For (each push to) a pull request, all commits currently in the PR
branch are checked.
For a regular push, all the pushed commits are checked, usually.
Subcases: push to master, push to other branch, force push;
I think at least the first two work, I don't care to spend more time
on it.
For a manual run, it seemed to check the same commits as a push (which
push ? Not sure how this works).
For a scheduled run - we'll see.
2021-07-10 23:26:11 -10:00
Simon Michael
d2fd039460
;ci: let all workflows be triggered manually
...
For when the intended triggers fail.
2021-07-03 14:51:51 -10:00
Simon Michael
fc7df75f43
;ci: no need for extra tags fetching ( #1543 )
...
fetch-depth: 0 also fetches tags.
2021-06-03 22:53:01 -10:00
Simon Michael
77be26514f
;ci: fix syntax ( #1543 )
2021-06-03 22:51:17 -10:00
Simon Michael
73de79d3d6
;ci: fix syntax ( #1543 )
2021-06-03 22:49:39 -10:00
Simon Michael
d16cd267f9
;ci: get all history to fix git describe/--version ( #1543 )
2021-06-03 22:47:56 -10:00
Simon Michael
d089f1b0d1
tools: ghc 8.6 is now the oldest, drop older stack configs
2021-06-03 14:05:02 -10:00
Simon Michael
31868a6892
;stack: build with lts-17/ghc-8.10.3 by default
2021-01-29 15:13:06 -08:00
Simon Michael
1023a3cc1a
;ci: run bin/ compilation tests only with GHC 8.2 CI for now ( #1453 )
2021-01-12 11:47:57 -08:00
Simon Michael
10bf0e5cf3
;ci: also run when bin/*.hs are changed ( #1453 )
2021-01-12 11:01:41 -08:00
Simon Michael
3ae6cf3200
ci: really add addon-building tests to CI ( #1453 )
2021-01-11 09:03:09 -08:00
Simon Michael
55e8ecf9e3
;ci: mac: cache bump to fix the usual failure
2021-01-10 07:01:42 -08:00
Simon Michael
06b466d847
ci: also test compilation of bin/ add-on scripts
2020-12-30 08:59:03 -08:00
Simon Michael
5eb69785e5
;ci: fix missing changes for moved tests
2020-11-22 09:48:46 -08:00
Simon Michael
6cc1312ca8
;ci: ignore example files; drop redundant hledger/tests/ entry
2020-11-17 12:55:13 -08:00
Simon Michael
19eb014ced
;ci: also ignore *.m4 (extra doc source files)
2020-11-17 10:57:42 -08:00
Simon Michael
ec2a774b18
cli, lib: move functional tests into hledger/test/
2020-11-16 08:47:41 -08:00
Simon Michael
5d42f1f198
;ci: exclude underscore-disabled tests
2020-11-13 10:23:47 -08:00
Simon Michael
fe5a97be81
;ci: mac: try 8.10 to avoid no symbol warnings
2020-10-18 23:04:03 -07:00
Simon Michael
035a590227
;ci: mac: try ghc 8.8 to fix cannot execute binary file
2020-10-18 22:57:11 -07:00
Simon Michael
8c9ca24adf
ci: tweak
2020-10-18 22:45:25 -07:00
Simon Michael
b2fcdcda16
;ci: tweak
2020-10-18 22:35:22 -07:00
Simon Michael
917cbdb2fa
;ci tweak
2020-10-18 22:23:37 -07:00
Simon Michael
655ae2a9bb
;ci: tweak
2020-10-18 22:18:48 -07:00
Simon Michael
aaaa1bd5cc
;ci: tweak
2020-10-18 22:10:57 -07:00
Simon Michael
484c69f795
;ci: work on getting latest stack on all platforms
2020-10-18 22:04:32 -07:00
Simon Michael
d2a1afdfc2
;ci: mac: invalidate mac cache to work around the usual failure
...
(cf https://github.com/haskell/cabal/issues/1076#issuecomment-685879454 )
2020-09-02 10:35:28 -07:00
Simon Michael
73332c0cbc
;ci: mac: try again to bump the cache to fix Cabal artifact
...
That last key still matched. I will defeat you, cache.
2020-08-23 08:40:53 -07:00
Simon Michael
c477e9c7dd
;ci: mac: try again to bump the right cache to fix Cabal artifact
...
CI error:
hledger-lib>
/Users/runner/.stack/setup-exe-cache/x86_64-osx/Cabal-simple_mPHDZzAJ_2.4.0.1_ghc-8.6.5:
cannot execute binary file
2020-08-23 08:33:41 -07:00
Simon Michael
3c32aa2cd4
;ci: mac: try bumping cache to fix ghc 8.6 breakage again
2020-08-23 07:53:27 -07:00
Simon Michael
3a32f28042
;ci: mac: bust cache, trying to fix strange "cannot execute" error
2020-08-10 08:47:11 -07:00
Simon Michael
22091dfd31
Revert ";ci: bump actions/cache from 2 to 2.1"
...
I guess that was not necessary or advisable.
This reverts commit 703c698b16 .
2020-08-01 12:59:28 -07:00
Simon Michael
703c698b16
;ci: bump actions/cache from 2 to 2.1
2020-08-01 12:49:53 -07:00
Simon Michael
31038e2888
;ci: cleanups
2020-07-13 11:49:54 -07:00
Simon Michael
a1c2c82c03
ci: spread ghc versions among workflows, clean up restore keys
2020-07-13 11:49:54 -07:00
Simon Michael
1f707ab0d8
;ci: linux: disable unreliable artifact upload causing CI failures
...
Also we don't need artifacts for every GHC version and every PR.
2020-06-25 18:07:02 -07:00
Simon Michael
b0ad7c138b
;ci: linux: use -split-sections to shrink binaries by ~30% (slower ?)
2020-06-21 18:18:42 -07:00
Simon Michael
269b981dec
ci: separate quick and full tests again, update
...
The linux full tests no longer test GHC 8.0,
and all full tests now always compile all modules
and generate optimised binaries.
2020-06-21 10:23:13 -07:00