Hey all, I looked into it more carefully, and I have been running a "patched" version of the gcc on the IRIX machine, in which a call for: >gcc -ar file1.o file2.o -o output.a generates an implicit call to > ar -rc output.a file1.o file2.o The fact that the gcc on red-hat did not throw an error on the "-ar" parameter threw me off-course. It seems that the problem lies in the tool that links the final outcome. Thank you all for your help! Cedric Roux-4 wrote: > > On 02/22/2011 05:32 PM, horseloverFat wrote: >> >> Cedric, it acts as if the types and variables that were defined in the 2 >> input files (I named them file1.o and file2.o) were not declared >> anywhere. >> Namely, running: >> >>> ld output.o file3.o >> >> yeilds the same results as simply running: >> >>> ld file3.o > > what results? errors? what messages does ld print? > Maybe a full example would be easier to diagnose > for people on this mailing list. > >> Cedric Roux-4 wrote: >>> >>> On 02/22/2011 04:47 PM, horseloverFat wrote: >>>> >>>> Thanks a lot for your help, >>>> yet I already tried that, but simply running "ar" doesn't seem to >>>> generate >>>> the appropriate output for the >>>> final link stage. >>> >>> How does the link stage fails? what errors do you get? >>> >>> >> > > > :handshake::handshake: -- View this message in context: http://old.nabble.com/Help-with-gcc-partial-link-%28option--ar%29-tp30983560p31001286.html Sent from the gcc - Help mailing list archive at Nabble.com.
