RISC OS Open: Forum: Fat32FS wildcard fail in GNU makefile

Jan 31, 2013 12:23pm

I tried to compile a C++ project last night that I had put on a 16GB USB drive attached to my Raspberry Pi. It failed because the line that get’s a list of source in the makefile returned no values. The line was:

CCSRC = $(wildcard *.cc)

I copied the whole project to the SDCard and it built correctly.

I was using GNU make from the autobuilder website (version 3.81 I think) and tried it with the Fat32FS that came with the Raspberry Pi image (1.40?) and also downloaded and tried it with 1.42.

Jan 31, 2013 1:37pm

Jeffrey Lee (213) 6048 posts

If this is a Fat32FS bug, then surely it’s something to report to Jeff Doggett? We can’t expect every developer to check the ROOL bugs forum for bug reports about their software!

Feb 1, 2013 8:07am

Jeff Doggett (257) 234 posts

Wildcards are handled by filecore(?), not by the individual filing systems. All Fat32fs does is provide the OS_GBPB calls to fetch the filenames. I rather suspect that it’s the R4 return value rearing it’s ugly head again.

Jeff

Feb 1, 2013 12:34pm

Alan Buckley (167) 232 posts

If this is a Fat32FS bug, then surely it’s something to report to Jeff Doggett? We can’t expect every developer to check the ROOL bugs forum for bug reports about their software!

My Apologies – I should have known better. I did know Jeff Doggett looked at these forums and thought it might be Raspberry Pi specific and was hoping someone who knew exactly how the line that was failing might have something to add. Still these are very weak excuses – sorry.

Wildcards are handled by filecore(?), not by the individual filing systems. All Fat32fs does is provide the OS_GBPB calls to fetch the filenames. I rather suspect that it’s the R4 return value rearing it’s ugly head again.

Does this mean it’s a fix that is needed in filecore or Fat32fs?

Feb 1, 2013 1:01pm

Chris Johnson (125) 825 posts

Does this mean it’s a fix that is needed in filecore or Fat32fs?

I would think the problem more likely lies in the application that is making the OS calls not dealing correctly with the values of offset being returned.

Feb 1, 2013 3:14pm

Jeff Doggett (257) 234 posts

I all fairness to Alan, it can be very difficult sometimes to work out exactly which piece of software is the culprit when things don’t work as expected. I think that he did the right thing in this case as it others here can help to pinpoint the problem.
As far as Fat32fs is concerned, I am denying that it’s me!

(Since as Chris points out, wildcards are handled by the application which is making the repeated os_gbpb calls).

Fat32FS wildcard fail in GNU makefile

Reply

Search forums

Social

ROOL Store

Donate! Why?

RISC OS IPR

Description

Voices

Options

Jan 31, 2013 12:23pm Alan Buckley (167) 232 posts	I tried to compile a C++ project last night that I had put on a 16GB USB drive attached to my Raspberry Pi. It failed because the line that get’s a list of source in the makefile returned no values. The line was: CCSRC = $(wildcard *.cc) I copied the whole project to the SDCard and it built correctly. I was using GNU make from the autobuilder website (version 3.81 I think) and tried it with the Fat32FS that came with the Raspberry Pi image (1.40?) and also downloaded and tried it with 1.42.

Jan 31, 2013 1:37pm Jeffrey Lee (213) 6048 posts	If this is a Fat32FS bug, then surely it’s something to report to Jeff Doggett? We can’t expect every developer to check the ROOL bugs forum for bug reports about their software!

Feb 1, 2013 8:07am Jeff Doggett (257) 234 posts	Wildcards are handled by filecore(?), not by the individual filing systems. All Fat32fs does is provide the OS_GBPB calls to fetch the filenames. I rather suspect that it’s the R4 return value rearing it’s ugly head again. Jeff

Feb 1, 2013 12:34pm Alan Buckley (167) 232 posts	If this is a Fat32FS bug, then surely it’s something to report to Jeff Doggett? We can’t expect every developer to check the ROOL bugs forum for bug reports about their software! My Apologies – I should have known better. I did know Jeff Doggett looked at these forums and thought it might be Raspberry Pi specific and was hoping someone who knew exactly how the line that was failing might have something to add. Still these are very weak excuses – sorry. Wildcards are handled by filecore(?), not by the individual filing systems. All Fat32fs does is provide the OS_GBPB calls to fetch the filenames. I rather suspect that it’s the R4 return value rearing it’s ugly head again. Does this mean it’s a fix that is needed in filecore or Fat32fs?

Feb 1, 2013 1:01pm Chris Johnson (125) 825 posts	Does this mean it’s a fix that is needed in filecore or Fat32fs? I would think the problem more likely lies in the application that is making the OS calls not dealing correctly with the values of offset being returned.

Feb 1, 2013 3:14pm Jeff Doggett (257) 234 posts	I all fairness to Alan, it can be very difficult sometimes to work out exactly which piece of software is the culprit when things don’t work as expected. I think that he did the right thing in this case as it others here can help to pinpoint the problem. As far as Fat32fs is concerned, I am denying that it’s me! (Since as Chris points out, wildcards are handled by the application which is making the repeated os_gbpb calls).