finesse issueshttps://git.ligo.org/finesse/finesse/-/issues2022-08-30T01:17:29Zhttps://git.ligo.org/finesse/finesse/-/issues/29Finesse 2.3.1 fails to build for Debian Bullseye (GCC 10.x); plus non-X86 bui...2022-08-30T01:17:29ZSteffen GrunewaldFinesse 2.3.1 fails to build for Debian Bullseye (GCC 10.x); plus non-X86 build issuesWhen building finesse 2.3.1 for Debian Bullseye on amd64, I'm running into heaps of warnings about "signedness" and "implicit fall-through".
In the end, the build fails with
```
/usr/bin/ld: kat_server.o:(.bss+0x2008): multiple definitio...When building finesse 2.3.1 for Debian Bullseye on amd64, I'm running into heaps of warnings about "signedness" and "implicit fall-through".
In the end, the build fails with
```
/usr/bin/ld: kat_server.o:(.bss+0x2008): multiple definition of `cq'; kat.o:(.bss+0x7330): first defined here
/usr/bin/ld: kat_server.o:(.bss+0x2000): multiple definition of `lastid'; kat.o:(.bss+0x7328): first defined here
```
Builds for Stretch (GCC 6.3.0) and Buster (8.3.0) are fine AFAICT.
The full log can be downloaded from https://hypatia.aei.mpg.de/lsc-amd64-bullseye/finesse_2.3.1-0.01+deb11u0_amd64.--pbuilderlog .https://git.ligo.org/finesse/finesse/-/issues/20finesse.sh: Mix of architectures and flags assigned to $CPUARCH2022-01-13T14:07:16ZSteffen Grunewaldfinesse.sh: Mix of architectures and flags assigned to $CPUARCHThat `-m32`, `-m64` can be assigned to a variable named `CPUARCH` (other values like `x86_64` look more reasonable) is mildly confusing.
I'm asking whether it'd be possible to rewrite this part so that `CPUARCH` always reflects some har...That `-m32`, `-m64` can be assigned to a variable named `CPUARCH` (other values like `x86_64` look more reasonable) is mildly confusing.
I'm asking whether it'd be possible to rewrite this part so that `CPUARCH` always reflects some hardware architecture, and an extra variable (`CPUFLAGS`?) gets flags like `-m64 -mfpmath=sse -ffast-math` (these three are scattered all over the code right now).
This should, in a not so far future, allow for addition of more --build* options, e.g. for non-X86 architectures (aarch64 wants `-a64` and doesn't like Intel-specific flags)https://git.ligo.org/finesse/finesse/-/issues/18Is finesse supposed to build on i386 (32-bit x86)?2019-10-30T13:40:20ZSteffen GrunewaldIs finesse supposed to build on i386 (32-bit x86)?```
./finesse.sh --build-linux ...```
./finesse.sh --build-linux
************************************
Starting new build!
Mon Oct 28 09:06:17 UTC 2019
************************************
Running on "linux"
Building Linux version :
Setting optimisation flags for gcc compiler
Required source and library folders exist?
- Source folder exists
- Library folder exists
Checking prerequisites...
- Found "gcc"...
- Found "ar"...
- Found "ranlib"...
- GSL found! (/usr/lib/i386-linux-gnu)
Found libgsl.a, use static linking
-m64
- Required toolchain found!
Configuring CUBA
rm: cannot remove 'libcuba.a': No such file or directory
Calling make file, see make.log for more details...
In file included from /usr/include/stdio.h:27:0,
from ../source/nics_config.h:6,
from ../source/nicsluc.h:8,
from ../source/nicsluc_aegraph.c:5:
/usr/include/features.h:364:25: fatal error: sys/cdefs.h: No such file or directory
# include <sys/cdefs.h>
```
`-m64` is certainly wrong - this is a 32-bit chroot running on a 64-bit machine though (which is a rather common setup for cross-arch builds). Do I need 32-bit metal to build the 32-bit version of finesse?https://git.ligo.org/finesse/finesse/-/issues/17Backup files in release tarball2019-11-17T22:42:14ZSteffen GrunewaldBackup files in release tarballDuring the Debian packaging process it was found that some backup files hadn't been removed from the tarball:
```
dpkg-source: warning: ignoring deletion of file lib/SuperLU_MT_2.1/EXAMPLE/pdlinsolx.c.bak, use --include-removal to overri...During the Debian packaging process it was found that some backup files hadn't been removed from the tarball:
```
dpkg-source: warning: ignoring deletion of file lib/SuperLU_MT_2.1/EXAMPLE/pdlinsolx.c.bak, use --include-removal to override
dpkg-source: warning: ignoring deletion of file lib/SuperLU_MT_2.1/EXAMPLE/pdspmd.c.bak, use --include-removal to override
dpkg-source: warning: ignoring deletion of file lib/SuperLU_MT_2.1/SRC/sp_coletree.c.orig, use --include-removal to override
```
Please try to avoid such extra files (which would be removed by `make clean`) in a release tarball.https://git.ligo.org/finesse/finesse/-/issues/16FINESSE fails to build on arm64, ppc64el2019-09-23T15:46:45ZSteffen GrunewaldFINESSE fails to build on arm64, ppc64elFINESSE fails to build on arm64 (and perhaps every non-X86 architecture) because `-m64` and `-mfpmath=sse` appear to be hard-coded. Can (and will) this be changed in a future release?FINESSE fails to build on arm64 (and perhaps every non-X86 architecture) because `-m64` and `-mfpmath=sse` appear to be hard-coded. Can (and will) this be changed in a future release?https://git.ligo.org/finesse/finesse/-/issues/15slink bug warnning and/or segfault2020-10-15T08:56:36ZAndreas Freiseslink bug warnning and/or segfaultFile:
```
l l1 1 0 n1
s s1 1 n1 n2
mod mod 10k 0.3 1 pm n2 n3
s s2 1 n3 nbs_in
bs bs 0.5 0.5 0 0 nbs_in dump nbs_out nout
s s3 1 nbs_out ncav_in
m2 m1 0.98 0 0 ncav_in ncav1
s scav 1 ncav1 ncav2
m2 m2 0.999 0 0.4 ncav2 dump
pd1 pdo $fs ...File:
```
l l1 1 0 n1
s s1 1 n1 n2
mod mod 10k 0.3 1 pm n2 n3
s s2 1 n3 nbs_in
bs bs 0.5 0.5 0 0 nbs_in dump nbs_out nout
s s3 1 nbs_out ncav_in
m2 m1 0.98 0 0 ncav_in ncav1
s scav 1 ncav1 ncav2
m2 m2 0.999 0 0.4 ncav2 dump
pd1 pdo $fs ncav_in
qnoisedN qn 1 $fs nout
tf m_F_z 1 0 p 10 1000
attr m2 mass 1 zmech m_F_z
tf d_z -10 0
slink sl1 pdo m2 z d_z
fsig sig m1 10k 0
xaxis sig f log 100 10M 100
yaxis log abs:deg
```
Produces:
```
------------------------------------------------------------------------
FINESSE 2.2 (build 2.2-49-g20a0c2f)
o_.-=. Frequency domain INterferomEter Simulation SoftwarE
(\'".\| 30.10.2017 http://www.gwoptics.org/finesse/
.>' (_--.
_=/d ,^\ Input file bug.kat,
~~ \)-' ' Output file bug.out,
/ | Gnuplot file bug.gnu
' ' Mon Nov 13 20:16:11 2017
------------------------------------------------------------------------
** Found dump nodes attached to components that have non-zero transmission.
This will produce incorrect results for quantum noise calculations as noise
is lost and not injected back in. Change dump nodes to proper nodes at the
at the following components:
** - bs
** - m2
** Modulator mod is coupling a lower signal sideband to a non-conjugated sideband, will give wrong results.
Calculating - 1%
Bug detected!
In function fill_qnoise_input_matrix() of file kat_quant.c at line 1206.
Error message: unhandled value 1079574528
Please report the version of Finesse you are using and the input
file that caused the bug at www.gwoptics.org/finesse/bug/. Thank you.
```https://git.ligo.org/finesse/finesse/-/issues/13Minus sign on mirror RP2020-10-15T08:56:36ZDaniel David Browndaniel.brown@ligo.orgMinus sign on mirror RPAnnoyingly, a minus sign is missing for the RP force on a suspended mirror somewhere it seems. This was noticed when looking trying to match Finesse to analytics.Annoyingly, a minus sign is missing for the RP force on a suspended mirror somewhere it seems. This was noticed when looking trying to match Finesse to analytics.Daniel David Browndaniel.brown@ligo.orgDaniel David Browndaniel.brown@ligo.orghttps://git.ligo.org/finesse/finesse/-/issues/12Phase noise modulation lower sideband conjugate2019-12-15T22:50:21ZDaniel David Browndaniel.brown@ligo.orgPhase noise modulation lower sideband conjugateModulator component wasn't conjugating lower sideband on phase noise creationModulator component wasn't conjugating lower sideband on phase noise creationDaniel David Browndaniel.brown@ligo.orgDaniel David Browndaniel.brown@ligo.orghttps://git.ligo.org/finesse/finesse/-/issues/10dbs warn on refractive index2017-08-24T13:30:06ZDaniel David Browndaniel.brown@ligo.orgdbs warn on refractive indexDBS component doesn't throw an error when refractive indices are different at each node.DBS component doesn't throw an error when refractive indices are different at each node.https://git.ligo.org/finesse/finesse/-/issues/8lock command offset2017-07-03T10:13:51ZDaniel David Browndaniel.brown@ligo.orglock command offsetWould be useful to be able to set offsets to locks more easily. Seems a bit cumbersome to always have to define another function for this. Proposed: `lock name $var gain accuracy [offset]`Would be useful to be able to set offsets to locks more easily. Seems a bit cumbersome to always have to define another function for this. Proposed: `lock name $var gain accuracy [offset]`Finesse 2.2Daniel David Browndaniel.brown@ligo.orgDaniel David Browndaniel.brown@ligo.orghttps://git.ligo.org/finesse/finesse/-/issues/7qnoised/qshot doesn't recognize k, M, etc.2017-06-21T22:30:08ZHang Yuqnoised/qshot doesn't recognize k, M, etc.[filt_cav.kat](/uploads/32bf9ce05ccc2568f3b4fe4950799c17/filt_cav.kat)
The kat file is attached. There are also additional commands
cmd=
"""
pd2 SUM1_I_ac $f_sb 0 10 n_wfs1
qnoised SUM1_I_n 2 2248443.435 0 10 n_wfs1
...[filt_cav.kat](/uploads/32bf9ce05ccc2568f3b4fe4950799c17/filt_cav.kat)
The kat file is attached. There are also additional commands
cmd=
"""
pd2 SUM1_I_ac $f_sb 0 10 n_wfs1
qnoised SUM1_I_n 2 2248443.435 0 10 n_wfs1
fsig sig1 M2HR ybeta 10 0 1u
fsig sig2 M2AR ybeta 10 0 1u
xaxis sig1 f log 0.1 1k 30
put SUM1_I_ac f2 $x1
put SUM1_I_n f2 $x1
yaxis re:im
"""
For qnoised the '2248443.435' one can be handled properly, but if I use '2.248443435M' instead and run the file with pykat, I get "ValueError: could not convert string to float: '2.248443435M'"Daniel David Browndaniel.brown@ligo.orgDaniel David Browndaniel.brown@ligo.org