grass-7.0.1: compiling on Gentoo shows warnings when built with -Wagressive-loop-optimizations

[wraeth]

• QA Notice: Package triggers severe warnings which indicate that it
• may exhibit random runtime failures.
• main.c:312:2: warning: iteration 14u invokes undefined behavior [-Waggressive-loop-optimizations]
• main.c:308:6: warning: iteration 14u invokes undefined behavior [-Waggressive-loop-optimizations]
• main.c:278:6: warning: iteration 14u invokes undefined behavior [-Waggressive-loop-optimizations]
• aerosolmodel.cpp:619:59: warning: iteration 82u invokes undefined behavior [-Waggressive-loop-optimizations]

This appears to be coming from the files:

imagery/i.aster.toar/main.c imagery/i.atcorr/aerosolmodel.cpp

Compile log to follow.

[wraeth]

Compressed build output

[wraeth]

Formatting messed the lines. The errors are produced when compiling the files imagery/i.aster.toar/main.c and imagery/i.atcorr/aerosolmodel.cpp

[glynn]

Replying to wraeth:

• main.c:312:2: warning: iteration 14u invokes undefined behavior [-Waggressive-loop-optimizations]
• main.c:308:6: warning: iteration 14u invokes undefined behavior [-Waggressive-loop-optimizations]
• main.c:278:6: warning: iteration 14u invokes undefined behavior [-Waggressive-loop-optimizations]

imagery/i.aster.toar/main.c

77  /*File Descriptors */
78  int infd[MAXFILES];
79  int outfd[MAXFILES];

277	for (i = 1; i <= MAXFILES; i++)
278	    Rast_get_row(infd[i], inrast[i-1], row, in_data_type[i-1]);

307	for (i = 1; i <= MAXFILES; i++)
308	    Rast_put_row(outfd[i], outrast[i-1], out_data_type);

310 for (i = 1; i <= MAXFILES; i++) {
311	G_free(inrast[i-1]);
312	Rast_close(infd[i]);
313	G_free(outrast[i-1]);
314	Rast_close(outfd[i]);
315 }

i can be equal to MAXFILES, which is the size of infd and outfd, so accessing infd[i] and outfd[i] is undefined behaviour.

• aerosolmodel.cpp:619:59: warning: iteration 82u invokes undefined behavior [-Waggressive-loop-optimizations]

imagery/i.atcorr/common.h

51	float phasel[10][83];

imagery/i.atcorr/aerosolmodel.cpp

619	for (int k = 1; k <= 83; ++k) sixs_sos.phasel[i][k] = 0.f;

Similar issue.

Was this converted from Fortran?

[neteler]

Replying to glynn:

Was this converted from Fortran?

Yes, long time back.

[neteler]

Ticket retargeted after milestone closed

[neteler]

Ticket retargeted after 7.0.3 milestone closed

[martinl]

Milestone renamed

[neteler]

Ticket retargeted after milestone closed

[martinl]

What is status of this issue?

[neteler]

I have now compiled relbranch74 (a.k.a. 7.4.0) with -Waggressive-loop-optimizations and the above reported i.atcorr compilation issues are fixed after Markus Metz's latest improvements.

Closing, feel free to reopen if needed.