Summary: | x11-toolkits/scintilla & editors/scite: Update to 3.6.3 | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Product: | Ports & Packages | Reporter: | lightside <lightside> | ||||||||||
Component: | Individual Port(s) | Assignee: | freebsd-ports-bugs (Nobody) <ports-bugs> | ||||||||||
Status: | Closed DUPLICATE | ||||||||||||
Severity: | Affects Some People | CC: | cyberbotx | ||||||||||
Priority: | --- | Keywords: | patch, patch-ready | ||||||||||
Version: | Latest | Flags: | cyberbotx:
maintainer-feedback+
|
||||||||||
Hardware: | Any | ||||||||||||
OS: | Any | ||||||||||||
Attachments: |
|
Description
lightside
2016-02-04 08:44:21 UTC
Created attachment 166539 [details]
Proposed patch for editors/scite (since 402318 revision)
Created attachment 166540 [details]
The poudriere testport log (FreeBSD 10.2 amd64): x11-toolkits/scintilla
Created attachment 166541 [details]
The poudriere testport log (FreeBSD 10.2 amd64): editors/scite
I've already submitted this update as bug #206751 albeit without the edit for the PIC flags in the scintilla port. I'll ask for this bug to be closed because I already submitted it, but is the PIC flags patch absolutely necessary? There were no issues with building it for amd64 or i386. Hello, Naram Qashat. (In reply to comment #4) > I've already submitted this update as bug #206751 albeit without the edit for > the PIC flags in the scintilla port. Sorry, I didn't see this on PortsMon on time of checking the port update (may be because of outdated information): http://portsmon.freebsd.org/portoverview.py?category=x11-toolkits&portname=scintilla (In reply to comment #4) > I'll ask for this bug to be closed because I already submitted it, but is the > PIC flags patch absolutely necessary? There were no issues with building it > for amd64 or i386. Sure, I can close it or it could be marked as duplicate of bug #206751. I also didn't have issues on amd64 for build and run without changes for PIC flags, but after reading the build log I found "-fPIC" added, along with "-DPIC -fpic", which was strange from the logic of the port's Makefile. Then I found following commit, which was cause of this: http://sourceforge.net/p/scintilla/code/ci/255c6d20fa88b2aa1ebbc67c0a139f6f1595d126/ After reading parts of GCC documentation, the svn logs for x11-toolkits/scintilla port, related to sparc64 architecture, I decided to propose current solution. Unfortunately, I can't test this on other architectures to be sure. Anyway, there is a possibility to use the proposed sed patch for PICFLAGS define, if you find it useful. *** This bug has been marked as a duplicate of bug 206751 *** I created following test case, based on GCC manual documentation: -8<-- #include <stdio.h> int main() { printf("__pic__ = %d, __PIC__ = %d\n", __pic__, __PIC__); return 0; } -->8- And run it with following configurations: % cc -fpic test.c -o test && ./test __pic__ = 1, __PIC__ = 1 % cc -fPIC test.c -o test && ./test __pic__ = 2, __PIC__ = 2 % cc -fPIC -fpic test.c -o test && ./test __pic__ = 1, __PIC__ = 1 % cc -fpic -fPIC test.c -o test && ./test __pic__ = 2, __PIC__ = 2 In result, the last option (either -fpic or -fPIC) was used. If this is how it works, then it should be ok for x11-toolkits/scintilla port, as long as port's CFLAGS applied last (and this is current case). |