Summary: | net-im/uTox update to 0.5.1.20160121 fail | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Product: | Ports & Packages | Reporter: | Ivan Rozhuk <rozhuk.im> | ||||||||||
Component: | Individual Port(s) | Assignee: | Dmitry Marakasov <amdmi3> | ||||||||||
Status: | Closed FIXED | ||||||||||||
Severity: | Affects Many People | CC: | fidaj, thierry | ||||||||||
Priority: | --- | Flags: | bugzilla:
maintainer-feedback?
(fidaj) |
||||||||||
Version: | Latest | ||||||||||||
Hardware: | Any | ||||||||||||
OS: | Any | ||||||||||||
Attachments: |
|
Description
Ivan Rozhuk
2016-01-28 10:03:22 UTC
Please, use net-im/tox with default options. Build with nacl-20110221_1 fail, with sodium - OK. May be NaCl option should be removed? (In reply to rozhuk.im from comment #2) This issue must be addressed to the port MAINTAINER= thierry@FreeBSD.org net-im/tox Regardless of whether net-im/tox has problems, uTox port is missing sodium dependency. And it should be fixed to produce full log instead of meaningless " CC src/svg.o" lines. (In reply to Dmitry Marakasov from comment #4) In any case, if net-im/tox built without SODIUM then net-im/uTox not build due to lack of libtoxav. Yes. You can also add explicit check for that. Created attachment 166238 [details]
uTox.diff
Created attachment 166243 [details]
uTox.diff
(In reply to Dmitry Marakasov from comment #6) Done. (In reply to Ivan Klymenko from comment #9) Should'nt you replace the BUILD_DEPENDS on libsodium.so by a LIB_DEPENDS? (In reply to rozhuk.im from comment #2) NaCl is fine for other Tox clients, only uTox encounter problem with it. The proposed patch (attachment 166243 [details]) should fix it, and packaged are built with libsodium by default. Created attachment 166245 [details]
uTox.diff
(In reply to Thierry Thomas from comment #10) Done. Hey. These changes are destined to fall into the ports tree? Thanks. The change is broken, have you tested it? - bsd.port.mk vs. bsd.port.post.mk - you cannot test for other ports' files before these ports are installed. You may test in pre-everything:: instead Created attachment 166797 [details] uTox.diff (In reply to Dmitry Marakasov from comment #15) Of course I test locally before sending the patch. Locally, the port build was held successfully and correctly. (In reply to Ivan Klymenko from comment #16) > Of course I test locally before sending the patch. > Locally, the port build was held successfully and correctly. The port which includes both bsd.port.pre.mk and bsd.port.mk (as opposed to bsd.port.post.mk) may not build, so it's clear that submission was not properly tested. The patch is still incorrect. The point is that IGNORE operates at time of Makefile processing, while dependencies are installed after this phase, so it'll always be ignored on cluster and will fail for end users unless they manually install net-im/tox port first. Actually I was wrong, pre-everything is not the right place either, but pre-configure is. Committing fixed version. A commit references this bug: Author: amdmi3 Date: Tue Feb 9 20:19:07 UTC 2016 New revision: 408591 URL: https://svnweb.freebsd.org/changeset/ports/408591 Log: - Add missing sodium dependency - Add a check for that net-im/tox dependency was built with correct options PR: 206706 Submitted by: fidaj@ukr.net (maintainer) Reported by: rozhuk.im@gmail.com Changes: head/net-im/uTox/Makefile |