Opened 11 years ago
Last modified 14 months ago
#42731 new enhancement
RFE: Add a conf option to automatically run test phase during normal install process
Reported by: | cooljeanius (Eric Gallager) | Owned by: | macports-tickets@… |
---|---|---|---|
Priority: | Normal | Milestone: | |
Component: | base | Version: | |
Keywords: | Cc: | ci42, mkae (Marko Käning), dstrubbe (David Strubbe), ryandesign (Ryan Carsten Schmidt) | |
Port: |
Description
While MacPorts currently has some support for running test suites for ports, this is hidden in the port test
command, which users must run manually separately from the normal install process. It would be much more convenient for users who are worried about the integrity of their software if they could just set something like
force_run_testsuites yes
in their macports.conf file. It would also be useful to have a Portfile option to override this, for example:
test.force no
to revert to the current behavior of only running the test suite manually with port test
, regardless of what the user has set in their conf file (in case the testsuite is broken, but the Portfile developer still wants to be able to work on fixing it).
Also, if this change were made, adding a depends_test
type of dependency as per #38208 would probably become more important.
Change History (8)
comment:1 Changed 11 years ago by ci42
Cc: | ciserlohn@… added |
---|
comment:5 follow-up: 6 Changed 10 years ago by ryandesign (Ryan Carsten Schmidt)
Cc: | ryandesign@… added |
---|---|
Version: | 2.2.1 |
What behavior are you envisioning when a forced test suite fails?
- The port install is prevented
- The failure is reported and the port installation proceeds anyway
I would favor (2).
test.force no
would also be useful for ports whose test suites are enormous and take hours to complete.
comment:6 follow-up: 8 Changed 10 years ago by cooljeanius (Eric Gallager)
Replying to ryandesign@…:
What behavior are you envisioning when a forced test suite fails?
- The port install is prevented
- The failure is reported and the port installation proceeds anyway
I would favor (2).
Yeah, I'd favor option 2 as well, but I'd just like to note that Homebrew does option 1, and I think that Fink also does that, as well...
test.force no
would also be useful for ports whose test suites are enormous and take hours to complete.
(i.e. openldap is the first one that comes to mind there...)
comment:7 Changed 8 years ago by petrrr
Cc: | Peter.Danecek@… removed |
---|
comment:8 Changed 14 months ago by cooljeanius (Eric Gallager)
Replying to cooljeanius:
Replying to ryandesign@…:
What behavior are you envisioning when a forced test suite fails?
- The port install is prevented
- The failure is reported and the port installation proceeds anyway
I would favor (2).
Yeah, I'd favor option 2 as well, but I'd just like to note that Homebrew does option 1, and I think that Fink also does that, as well...
So on further consideration, I'd just like to note that Homebrew's approach here is really annoying, and that it would be best to just completely disregard it.
test.force no
would also be useful for ports whose test suites are enormous and take hours to complete.(i.e. openldap is the first one that comes to mind there...)
(also gcc and openssl are coming to mind (as ports with enormous testsuites) more recently now, too)
Cc Me!