<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="/rss.xsl.xml"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/">
<channel>
    <title>Changes in basiclink.b.ll</title>
    <description></description>
    <language>en</language>
    <copyright>Copyright 2015</copyright>
    <generator>Java</generator><item>
        <title>23af6484 - [opaque pointer type] Add textual IR support for explicit type parameter to the call instruction</title>
        <link>http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/Linker/Inputs/basiclink.b.ll#23af6484</link>
        <description>[opaque pointer type] Add textual IR support for explicit type parameter to the call instructionSee r230786 and r230794 for similar changes to gep and loadrespectively.Call is a bit different because it often doesn&apos;t have a single explicittype - usually the type is deduced from the arguments, and just thereturn type is explicit. In those cases there&apos;s no need to change theIR.When that&apos;s not the case, the IR usually contains the pointer type ofthe first operand - but since typed pointers are going away, thatrepresentation is insufficient so I&apos;m just stripping the &quot;pointerness&quot;of the explicit type away.This does make the IR a bit weird - it /sort of/ reads like the type ofthe first operand: &quot;call void () %x(&quot; but %x is actually of type &quot;void()*&quot; and will eventually be just of type &quot;ptr&quot;. But this seems not toobad and I don&apos;t think it would benefit from repeating the type(&quot;void (), void () * %x(&quot; and then eventually &quot;void (), ptr %x(&quot;) as hasbeen done with gep and load.This also has a side benefit: since the explicit type is no longer apointer, there&apos;s no ambiguity between an explicit type and a functionthat returns a function pointer. Previously this case needed an explicittype (eg: a function returning a void() function was written as&quot;call void () () * @x(&quot; rather than &quot;call void () * @x(&quot; because of theambiguity between a function returning a pointer to a void() functionand a function returning void).No ambiguity means even function pointer return types can just bewritten alone, without writing the whole function&apos;s type.This leaves /only/ the varargs case where the explicit type is required.Given the special type syntax in call instructions, the regex-fu usedfor migration was a bit more involved in its own unique way (as everyone of these is) so here it is. Use it in conjunction with the apply.shscript and associated find/xargs commands I&apos;ve provided in rr230786 tomigrate your out of tree tests. Do let me know if any of this doesn&apos;tcover your cases &amp; we can iterate on a more general script/regexes tohelp others with out of tree tests.About 9 test cases couldn&apos;t be automatically migrated - half of thosewere functions returning function pointers, where I just had to manuallydelete the function argument types now that we didn&apos;t need an explicitfunction type there. The other half were typedefs of function types usedin calls - just had to manually drop the * from those.import fileinputimport sysimport repat = re.compile(r&apos;((?:=|:|^|\s)call\s(?:[^@]*?))(\s*$|\s*(?:(?:\[\[[a-zA-Z0-9_]+\]\]|[@%](?:(&quot;)?[\\\?@a-zA-Z0-9_.]*?(?(3)&quot;|)|{{.*}}))(?:\(|$)|undef|inttoptr|bitcast|null|asm).*$)&apos;)addrspace_end = re.compile(r&quot;addrspace\(\d+\)\s*\*$&quot;)func_end = re.compile(&quot;(?:void.*|\)\s*)\*$&quot;)def conv(match, line):  if not match or re.search(addrspace_end, match.group(1)) or not re.search(func_end, match.group(1)):    return line  return line[:match.start()] + match.group(1)[:match.group(1).rfind(&apos;*&apos;)].rstrip() + match.group(2) + line[match.end():]for line in sys.stdin:  sys.stdout.write(conv(re.search(pat, line), line))llvm-svn: 235145

            List of files:
            /llvm-project-15.0.7/llvm/test/Linker/Inputs/basiclink.b.ll</description>
        <pubDate>Thu, 16 Apr 2015 23:24:18 +0000</pubDate>
        <dc:creator>David Blaikie &lt;dblaikie@gmail.com&gt;</dc:creator>
    </item>
<item>
        <title>8bdfe1ec - Switch a bunch of Linker tests from using elaborate echo productions to</title>
        <link>http://172.16.0.5:8080/history/llvm-project-15.0.7/llvm/test/Linker/Inputs/basiclink.b.ll#8bdfe1ec</link>
        <description>Switch a bunch of Linker tests from using elaborate echo productions tojust provide and reference separate input files from an Inputssubdirectory. This pattern works very well in the Clang tree and iseasier to understand in my opinion. It also has fewer limitations andwill remove one particularly annoying use of TCL-style {} quoting fromthe testsuite.Also teach the LLVM lit configuration to avoid recursing into &apos;Inputs&apos;subdirectories. This wasn&apos;t required for the previous &apos;Inputs&apos;subdirectories used due to fortuitous suffix patterns.This is the first step to completely removing support for TCL-style tests.llvm-svn: 159520

            List of files:
            /llvm-project-15.0.7/llvm/test/Linker/Inputs/basiclink.b.ll</description>
        <pubDate>Mon, 02 Jul 2012 10:18:06 +0000</pubDate>
        <dc:creator>Chandler Carruth &lt;chandlerc@gmail.com&gt;</dc:creator>
    </item>
</channel>
</rss>
