Control: tag -1 + confirmed Control: severity -1 minor On Mon, 30 May 2022 18:35:57 -0300, Jesse Smith wrote:
> > when I install the cpulimit package in testing, version 2.7 downloads. > > But when it is run, cpulimit itself shows, that it is actually version > > 2.4. Aren't these two supposed to match? > > > Yes, they should match. Upstream, the cpulimit executable displays the > correct (2.7) version number when it is run: > > $ cpulimit > CPUlimit version 2.7 > Usage: ./cpulimit TARGET [OPTIONS...] [-- PROGRAM] Thanks Radim for the bug report, and thank Jesse for the quick reply. Interestingly I can confirm this behaviour here on Debian/unstable with the same cpulimit version as in testing (2.7-2): % /usr/bin/cpulimit Error: You must specify a target process CPUlimit version 2.4 Usage: /usr/bin/cpulimit TARGET [OPTIONS...] [-- PROGRAM] Now where does this 2.4 come from? % grep -r 2\\.4 cpulimit.c:#define VERSION 2.4 … And how/where is VERSION used? % grep -r VERSION cpulimit.c:#ifndef VERSION cpulimit.c:#define VERSION 2.4 cpulimit.c: fprintf(stream, "CPUlimit version %1.1f\n", VERSION); … Makefile:VERSION?=2.7 Makefile:CFLAGS?=-Wall -O2 -DVERSION=$(VERSION) Makefile: cd .. && tar czf cpulimit-$(VERSION).tar.gz --exclude=.svn cpulimit-$(VERSION) I think what is happening here is that CFLAGS is not taken from the Makefile but overridden from the Debian build machinery, so -DVERSION (which would be 2.7) is not set/defined and VERSION is taken from cpulimit.c and that still says 2.4. I guess maintaining a variable in two places can easily lead to this mismatch, and that the build process in Debian is slightly different doesn't help. I assume fixing this in a future upstream release is enough … Cheers, gregor -- .''`. https://info.comodo.priv.at -- Debian Developer https://www.debian.org : :' : OpenPGP fingerprint D1E1 316E 93A7 60A8 104D 85FA BB3A 6801 8649 AA06 `. `' Member VIBE!AT & SPI Inc. -- Supporter Free Software Foundation Europe `-
signature.asc
Description: Digital Signature
