Apache OpenOffice (AOO) Bugzilla – Issue 24980
cannot build toolkit without -fpermissive
Last modified: 2004-03-14 18:37:07 UTC
fresh update from cws_src680_ooo20031216 these errors become warnings with -fpermissive jim@sun:~/680/o2_src/toolkit$ uname -a Linux sun 2.4.18 #1 SMP Mon Jan 27 14:07:39 EST 2003 sparc64 GNU/Linux jim@sun:~/680/o2_src/toolkit$ g++ --version g++ (GCC) 3.3.3 20040110 (prerelease) (Debian) /home/jim/680/o2_src/toolkit/source/awt/vclxspinbutton.cxx: In member function `virtual void toolkit::VCLXSpinButton::setValue(long int)': /home/jim/680/o2_src/toolkit/source/awt/vclxspinbutton.cxx:193: error: assuming pointer to member `void SpinButton::SetValue(long int)' /home/jim/680/o2_src/toolkit/source/awt/vclxspinbutton.cxx:193: error: (a pointer to member can only be formed with `&SpinButton::SetValue') and 209, 215, 221, 227, 233, 239, 245 dmake: Error code 1, while making '../../unxlngs.pro/slo/vclxspinbutton.obj' ---* TG_SLO.MK *---
mh->mt: please have a look.
MT->FS: Seems to be yours...
replacing "SpinButton::SetValue" with "&SpinButton::SetValue" should do here, IIRC (had some similar problem in another file). Same in lines 209, 215, 221, 227, 233, 239, 245. sparcmoz, do you mind trying this?
yes, it will build OK without -fpermissive after replacing those "SpinButton::SetValue" with "&SpinButton::SetValue". I also rebuilt the whole toolkit module without -fpermissive and no other errors were found.
great - should I check it in, or could you (who you're already working on this branch) do this?
I will attach the patch file now. Please you commit this since I do not like to make a change that affects all platforms, especially as I have never committed anything yet ;) Anyway I do not know what branch I am "working on" since i am updating from cws_src680_ooo20031216 which is "closed".
Created attachment 12852 [details] cd toolkt; patch -p0 < toolkit.diff
ehm, well, with branch I meant cws_src680_ooo20031216 ... Thanks for the patch file - I am going to check into which processes I can embed the checkin ... fs->mh: Since we're not building without -fpermissive, I do not see any sense in putting this issue into a CWS - it cannot be verified there. What's the process for getting such changes in the tree?
Re: "Since we're not building without -fpermissive," I think waratah made all the changes to remove -fpermissive from 680, and this is in all(?) the solenv/inc/platform.mk files --> "# -fpermissive should be removed as soon as possible" i did already build all cws_src680_ooo20031216 without -fpermissive only until after it closed :( But maybe this time it is because I have gone to gcc 3.3.3? anyway I do not mean to cause any problems here, maybe it is invalid now, but i suppose it is best to fix as it will become a problem eventually? Maybe let it wait until a new branch is provide for porting, no-one else seems to be building this just now?
I wouldn't close this one as invalid, it's certainly worth fixing. However, I am just wondering where to check in the fix, and I hope Martin can enlighten me here :)
accepting
JFYI: it seems to be also on SRC680_m24 as of today.
*** Issue 25152 has been marked as a duplicate of this issue. ***
I fixed this issue on just opened cws_src680_ooo20040225.
thanks!
*** Issue 25787 has been marked as a duplicate of this issue. ***
Verified.
*** Issue 26172 has been marked as a duplicate of this issue. ***
Fixed in cws_src680_ooo20040329. Closing.