Description of problem: sdcc/gputils is supposed to support pic16f1704/8 among others, but fails to link with strange messages about family mismatch and undefined symbols. Version-Release number of selected component (if applicable): SDCC : mcs51/z80/z180/r2k/r3ka/gbz80/tlcs90/ds390/pic16/pic14/TININative/ds400/hc08/s08/stm8 3.4.0 #8981 (Oct 16 2014) (Linux) gplink-1.2.0 #980 (Oct 18 2014) How reproducible: always Steps to Reproduce: 1. sdcc --use-non-free -mpic14 -ppic16f1708 --opt-code-size blink.c message: using default linker script "/usr/share/gputils/lkr/16f1708_g.lkr" error: processor family mismatch in "idata.o" error: processor family mismatch in "_gptrget1.o" error: processor family mismatch in "_gptrget2.o" error: missing definition for symbol "_INDF", required by "_gptrget1.o" error: missing definition for symbol "_FSR", required by "_gptrget1.o" You can find the blink,c here: http://vserver.13thfloor.at/Stuff/MAGEIA/blink.c Reproducible: Steps to Reproduce:
Assignee: bugsquad => zen25000
(In reply to Herbert Poetzl from comment #0) Just to be sure, are you using the version from Mageia core or non-free repos?
sdcc-3.4.0-7.mga5.nonfree gputils-1.2.0-4.mga5 Best, Herbert
CC: (none) => herbert
Hi Herbert, I'm just looking at this: http://sourceforge.net/p/sdcc/mailman/message/34326697/ Looks like the issue but I need to see if it's been patched anywhere. Barry
Yes that fixes it. [baz@localhost SOURCES]$ sdcc --use-non-free -mpic14 -ppic16f1708 --opt-code-size blink.c message: using default linker script "/usr/share/gputils/lkr/16f1708_g.lkr" [baz@localhost SOURCES]$ No errors :) I will push patched versions to 5/nonfree/updates_testing and 5/core/updates_testing later. Please test.
Ping? Herbert could you please confirm that the patched version fixes this for you. Thanks, Barry
Keywords: (none) => NEEDINFO
Yep, patched version works for pic16f1704/8. Thanks! Btw, an update to sdcc 3.5 and gputils 1.4 (from 2014 :) would be appreciated. Best, Herbert
(In reply to Herbert Poetzl from comment #6) > Yep, patched version works for pic16f1704/8. > > Thanks! > > Btw, an update to sdcc 3.5 and gputils 1.4 (from 2014 :) would be > appreciated. > > Best, > Herbert 3.5 currently won't build in either Mga5 or cauldron, but maybe down to gputils version. I will look at it again when cauldron is less of a train crash.
################################### Update Advisory ################################### sdcc-3.4.0-8 has been submitted to 5/core/updates_testing and 5/nonfree/updates_testing this update only affects the non-free version. Description: This update to sdcc-3.4.0-8 fixes an issue with pic16f1704 and pic16f1708 compilation. This only affected the versions in the nonfree repository. See the following references for further information: https://bugs.mageia.org/show_bug.cgi?id=16565 and http://sourceforge.net/p/sdcc/mailman/message/34326697/ Affected rpms: sdcc-3.4.0-8.mga5.nonfree.x86_64.rpm sdcc-3.4.0-8.mga5.nonfree.i586.rpm sdcc-debuginfo-3.4.0-8.mga5.nonfree.x86_64.rpm sdcc-debuginfo-3.4.0-8.mga5.nonfree.i586.rpm sdcc-3.4.0-8.mga5.x86_64.rpm sdcc-3.4.0-8.mga5.i586.rpm sdcc-debuginfo-3.4.0-8.mga5.x86_64.rpm sdcc-debuginfo-3.4.0-8.mga5.i586.rpm sdcc-3.4.0-8.mga5.nonfree.src.rpm sdcc-3.4.0-8.mga5.src.rpm Testing: This is a really minor fix that was down to two typos in a data file. I would suggest that no further testing is needed.
Assignee: zen25000 => qa-bugs
(In reply to Herbert Poetzl from comment #6) > > Btw, an update to sdcc 3.5 and gputils 1.4 (from 2014 :) would be > appreciated. > These are both now in Cauldron and it would be really good if you could test :) Barry > Best, > Herbert
CC: (none) => zen25000
Based on comment 6 and comment 8, I have validated this update. A QA committer needs to upload the advisory to SVN The packages can then be pushed to updates
Keywords: NEEDINFO => validated_updateCC: (none) => sysadmin-bugsWhiteboard: (none) => MGA5-64-OK
Advisory uploaded. Barry, please remember next time to define a subrel instead of bumping the release when you provide an update. You should have used: %define subrel 1 And the package would have been named: sdcc-3.4.0-7.1.mga5{,.nonfree}
Whiteboard: MGA5-64-OK => MGA5-64-OK advisory
Yes - will try to remember.
An update for this issue has been pushed to Mageia Updates repository. http://advisories.mageia.org/MGAA-2015-0097.html
Status: NEW => RESOLVEDResolution: (none) => FIXED