Here we drop support for the slave port, since it never quite worked
as desired, nor was it ever submitted, and removes the conditionals to support
it. Also, we add a new default option FACTER_JAVA to build support for Java
and install facter.jar to be used by sysutils/puppetserver5
Details
Details
Poudriere testport
Diff Detail
Diff Detail
- Repository
- rP FreeBSD ports repository
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
Comment Actions
I think the JAVA_BUILD here is making openjdk8 a run time dependency, but I only want it to be a build-time dependency. I'm reading.
Comment Actions
Nevermind, that must have been installed in the jail as part of the build and left there. I can remove openjdk8 from the jail without de-installation of facter.
| Makefile | ||
|---|---|---|
| 5 ↗ | (On Diff #33428) | probably no need to go from 0 to 2 directly. |
| 6 ↗ | (On Diff #33428) | Why ?= ? |
| 27 ↗ | (On Diff #33428) | Why += ? Also, it happens too early. |
| 31 ↗ | (On Diff #33428) | why += ? Also, it happens too late. |
| 38 ↗ | (On Diff #33428) | typo ? (and missing tab after =) |
| 42–45 ↗ | (On Diff #33428) | FACTER_JAVA_VARS= JAVA_BUILD=yes FACTER_JAVA_CMAKE_ON= -DJAVA_HOME=... |
| 57 ↗ | (On Diff #33428) | This should be post-install-FACTER_JAVA-on: |
| 64 ↗ | (On Diff #33428) | you can drop the PKGNAMEPREFIX test here. |
| pkg-plist | ||
| 24 ↗ | (On Diff #33428) | This probably needs to be guarded by a %%FACTER_JAVA%% as it is optional. (And also add OPTIONS_SUB=yes in the Makefile.) |
Comment Actions
| Makefile | ||
|---|---|---|
| 6 ↗ | (On Diff #33428) | This was introduced when we where working on a slave port, and the salve would have been in the java category. This is usually done by setting the category using '?=' in the master port and with '=' in the slave port. Here is the commit where it was introduced: Since we drop this slave port, this can be revert to '='. |
| 27 ↗ | (On Diff #33428) | I guess it's a mistake I introduced. Could be replaced by a simple '='. |
| 31 ↗ | (On Diff #33428) | There is a USES=cmake on line 15, and since the conditional on line 17 was removed, everything should be moved to a single USES assignation. |
| 57 ↗ | (On Diff #33428) | I am pretty sure that this target is useless. I had this code in order to only install the produced jar file (without running make install). But since now we build the jar file if the option is selected, I guess that the make install stage will copy the .jar file at the right location automatically. |
| 64 ↗ | (On Diff #33428) | Yep, it was used to determine if the user was installing sysutils/facter or sysutils/java-facter. It is now useless. |
| Makefile | ||
|---|---|---|
| 5 ↗ | (On Diff #33428) | ?= was required in the master port so that the slave port could have a different one. Should be =. |
Comment Actions
Thank you for the review. I believe I've addressed the feedback, and perhaps correctly.
| Makefile | ||
|---|---|---|
| 42–45 ↗ | (On Diff #33428) | Oh neat. I see that now in the docs. |
| 57 ↗ | (On Diff #33428) | That would be nice. I'll test without it. |
| 57 ↗ | (On Diff #33428) | Yep, not required. Thanks for the pointer! |
| pkg-plist | ||
| 24 ↗ | (On Diff #33428) | I see. Interesting. |