Use portable flag with nm in extract_symbols.py

[daltenty]

Summary

nm is one of the tools that extract_symbols.py can use to extract
symbols from llvm libraries as part of the build process. This patch
updates the invocation of nm to use the -P POSIX option for "portable
output" so we get a consistently parsable output format on all
platforms.

A link to the relevant nm format: https://pubs.opengroup.org/onlinepubs/9699919799/utilities/nm.html

[stevewan]

llvm/utils/extract_symbols.py

51

So I see we are changing the target of group in the regex match, but I don't see clearly the rationale behind it. Is it based on some testing outputs?

[daltenty]

llvm/utils/extract_symbols.py

51

The nm portable output format is:

"%s%s %s %x %x\n", 
<library/object name>, <name>, <type>, <value>, <size>

(where <library/object name> will be empty in our usage)

So we change the target group to deal with the new position of the symbol name.

[stevewan]

LGTM

[emaste]

Is this using the system nm? ELF Tool Chain nm does not (currently) support the -P flag.

[daltenty]

In D69004#1719247, @emaste wrote:

Is this using the system nm? ELF Tool Chain nm does not (currently) support the -P flag.

extract_symbol.py does normally invoke the system nm (which is why we wanted to apply the portability flag).

I was under the impression that -P was supported on FreeBSD. If not, perhaps there is an equivalent posix format option?

[emaste]

I was under the impression that -P was supported on FreeBSD.

Hrm, I was mistaken and nm -P is supported on FreeBSD (with the ELF Tool Chain nm in the base system); I'm not sure what I ended up trying earlier. Sorry for the noise.

llvm/utils/extract_symbols.py

def dumpbin_get_symbols(lib):
     for line in process.stdout:
         # Look for external symbols that are defined in some section
         match = re.match("^.+SECT.+External\s+\|\s+(\S+).*$", line)
         if match:
             yield match.group(1)
     process.wait()
 
 def nm_get_symbols(lib):
-    process = subprocess.Popen(['nm',lib], bufsize=1,
+    process = subprocess.Popen(['nm','-P',lib], bufsize=1,
                                stdout=subprocess.PIPE, stdin=subprocess.PIPE,
                                universal_newlines=True)
     process.stdin.close()
     for line in process.stdout:
         # Look for external symbols that are defined in some section
-        match = re.match("^\S+\s+[BDGRSTVW]\s+(\S+)$", line)
+        match = re.match("^(\S+)\s+[BDGRSTVW]\s+\S+\s+\S+$", line)
         if match:
             yield match.group(1)
     process.wait()
 
 def readobj_get_symbols(lib):
     process = subprocess.Popen(['llvm-readobj','-symbols',lib], bufsize=1,
                                stdout=subprocess.PIPE, stdin=subprocess.PIPE,
                                universal_newlines=True)