[llvm-rc] Serialize HTML resources to .res files (serialization, pt 1).

[mnbvmar]

Summary

This allows to process HTML resources defined in .rc scripts and output them to resulting .res files. Additionally, some infrastructure allowing to output these files is created.

This is the first resource type we can operate on.

[mnbvmar]

One file was missing from the patch.

[ecbeckmann]

llvm/tools/llvm-rc/ResourceScriptStmt.cpp

148 ↗

(On Diff #113184)

It would be good if we could reuse the .res file structures defined in WindowsResource.h, to ensure the sizes of each field are kept in sync and also have a single unified place in the code for defining the .res format. So instead of writing each field individually, you could define a local struct WinResHeaderPrefix and then call Writer.write<struct WinResHeaderPrefix>(the struct)

[mnbvmar]

Changed the code according to Eric's suggestion; some bugfixes.

[ecbeckmann]

llvm/tools/llvm-rc/ResourceScriptStmt.cpp

117 ↗

(On Diff #113295)

Instead of an overload you should rename this to writeIntOrString for clarity.

[mnbvmar]

Refactored the code to simplify the patches following this one.

Previously, resource classes had ability to output themselves, using and possibly modifying the current context. Now, Visitor class is introduced which holds the current context and processes resources one by one and serializes them.

This still holds the ability to output HTML resources.

[mnbvmar]

This fixes @ecbeckmann's comment https://reviews.llvm.org/D37824#inline-329768 (set type IDs and memory flags as constants). It's probably better to fix this in the first patch this appears in.

[rnk]

llvm/test/tools/llvm-rc/tag-html.test

2–4

nit: personally I like RUN: rm -rf %t && mkdir %t && cd %t as the standard one-line idiom for "get this test a clean working directory"

7

Do we have any tools that can dump resource files? These kind of "golden file" tests are pretty fragile and aren't easy to debug when they fail.

llvm/tools/llvm-rc/ResourceSerializator.h

1 ↗

(On Diff #115144)

Even if "Serializator" is a word, I'd recommend something more common, like ResourceFileWriter or something.

56 ↗

(On Diff #115144)

Maybe call this writeBytes, and take ArrayRef<uint8_t>?

58 ↗

(On Diff #115144)

I'm concerned that an object could easily contain endian-swapped fields, but I can't think of a clean way to defend against it, so let's go with this.

[mnbvmar]

llvm/test/tools/llvm-rc/tag-html.test

7

We could pipe the output of llvm-cvtres to llvm-readobj. However, cvtres needs to reorder the resources, and readobj can only extract header information.

llvm/tools/llvm-rc/ResourceSerializator.h

58 ↗

(On Diff #115144)

The only thing I can think of is some std::disable_if detecting basic integral types larger than a byte. This is nowhere close to perfect, though.

[zturner]

llvm/tools/llvm-rc/ResourceSerializator.h

56 ↗

(On Diff #115144)

It seems like all of these functions are reproducing logic that already exists in BinaryStreamWriter. Is there some reason we can't use that here?

The nice thing about them is they prevent endianness errors up front, because you don't have to remember to call writeObject() with some endian-aware type, because as soon as you set the endianness on the stream, all integers get written appropriately.

It also contains functions for writing c strings, arrays, arbitrary objects, and seeking, so it seems like a good fit here and reduces some boilerplate?

llvm/tools/llvm-rc/llvm-rc.cpp

156

Can this be a unique_ptr?

[mnbvmar]

llvm/tools/llvm-rc/ResourceSerializator.h

56 ↗

(On Diff #115144)

BinaryStreamWriter expects to know the size of the underlying buffer beforehand, so I can't really use it here.

However, is it a good idea to teach this to write little-endian integers only (and convert native-endian ints to ulittle{}_t when necessary)?

[zturner]

llvm/tools/llvm-rc/ResourceSerializator.h

56 ↗

(On Diff #115144)

Yes, just call it writeInteger and do the endian conversion inside the function. Then call the function writeInteger<uint64_t>(12);

[mnbvmar]

Style fixes suggested in the comments. Integrated llvm-readobj into test invocation.

(I changed file and class names to ResourceFileWriter. However, I leave the revision as-is not to make confusion.)

[rnk]

lgtm, thanks!

[mnbvmar]

Seems this revision didn't close automatically. Fixing that.

llvm/test/tools/llvm-rc/Inputs/tag-html-wrong.rc

+1 HTML "some-really-nonexistent-file.html"

llvm/test/tools/llvm-rc/Inputs/tag-html.rc

+100 HTML "webpage1.html"
+Kitten HTML "webpage2.html"

llvm/test/tools/llvm-rc/Inputs/webpage1.html

+<html>
+  <body>
+    Hello!
+  </body>
+</html>

llvm/test/tools/llvm-rc/Inputs/webpage2.html

+<!-- Should not embed the image. -->
+<img src="kittens.bmp">

llvm/test/tools/llvm-rc/helpmsg.test

 ; RUN: llvm-rc /h > %t1
 ; RUN: llvm-rc '/?' > %t2
 ; RUN: diff %t1 %t2
 ; RUN: FileCheck -input-file=%t1 %s
 
 ; CHECK:  OVERVIEW: Resource Converter
 ; CHECK-DAG:  USAGE: rc [options] <inputs>
 ; CHECK-DAG:  OPTIONS:
 ; CHECK-NEXT:    /?          Display this help and exit.
+; CHECK-NEXT:    /dry-run    Don't compile the input; only try to parse it.
 ; CHECK-NEXT:    /D <value>  Define a symbol for the C preprocessor.
 ; CHECK-NEXT:    /FO <value> Change the output file location.
 ; CHECK-NEXT:    /H          Display this help and exit.
 ; CHECK-NEXT:    /I <value>  Add an include path.
 ; CHECK-NEXT:    /LN <value> Set the default language name.
 ; CHECK-NEXT:    /L <value>  Set the default language identifier.
 ; CHECK-NEXT:    /N          Null-terminate all strings in the string table.
 ; CHECK-NEXT:    /U <value>  Undefine a symbol for the C preprocessor.
 ; CHECK-NEXT:    /V          Be verbose.
 ; CHECK-NEXT:    /X          Ignore 'include' variable.
 ; CHECK-NEXT:    /Y          Suppress warnings on duplicate resource IDs.

llvm/test/tools/llvm-rc/parser-expr.test

-; RUN: llvm-rc /V %p/Inputs/parser-expr.rc | FileCheck %s
+; RUN: llvm-rc /dry-run /V %p/Inputs/parser-expr.rc | FileCheck %s
 
 ; CHECK:  Language: 5, Sublanguage: 1
 ; CHECK-NEXT:  Language: 3, Sublanguage: 2
 ; CHECK-NEXT:  Language: 4294967293, Sublanguage: 4294967292
 ; CHECK-NEXT:  Language: 0, Sublanguage: 1
 ; CHECK-NEXT:  Language: 2, Sublanguage: 4
 ; CHECK-NEXT:  Language: 3, Sublanguage: 5
 ; CHECK-NEXT:  Language: 2, Sublanguage: 0
 ; CHECK-NEXT:  Language: 4294967295, Sublanguage: 1
 ; CHECK-NEXT:  Language: 1, Sublanguage: 4294967295
 ; CHECK-NEXT:  Language: 4294967294, Sublanguage: 1
 ; CHECK-NEXT:  Language: 1, Sublanguage: 4294967294
 ; CHECK-NEXT:  Language: 2, Sublanguage: 1
 ; CHECK-NEXT:  Language: 3, Sublanguage: 5
 ; CHECK-NEXT:  Language: 0, Sublanguage: 13
 ; CHECK-NEXT:  Language: 5, Sublanguage: 7
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-expr-bad-binary-1.rc 2>&1 | FileCheck %s --check-prefix BINARY1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-expr-bad-binary-1.rc 2>&1 | FileCheck %s --check-prefix BINARY1
 
 ; BINARY1: llvm-rc: Error parsing file: expected '-', '~', integer or '(', got &
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-expr-bad-binary-2.rc 2>&1 | FileCheck %s --check-prefix BINARY2
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-expr-bad-binary-2.rc 2>&1 | FileCheck %s --check-prefix BINARY2
 
 ; BINARY2: llvm-rc: Error parsing file: expected '-', '~', integer or '(', got |
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-expr-bad-binary-3.rc 2>&1 | FileCheck %s --check-prefix BINARY3
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-expr-bad-binary-3.rc 2>&1 | FileCheck %s --check-prefix BINARY3
 
 ; BINARY3: llvm-rc: Error parsing file: expected '-', '~', integer or '(', got +
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-expr-bad-unary.rc 2>&1 | FileCheck %s --check-prefix UNARY
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-expr-bad-unary.rc 2>&1 | FileCheck %s --check-prefix UNARY
 
 ; UNARY: llvm-rc: Error parsing file: expected ',', got ~
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-expr-unbalanced-1.rc 2>&1 | FileCheck %s --check-prefix UNBALANCED1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-expr-unbalanced-1.rc 2>&1 | FileCheck %s --check-prefix UNBALANCED1
 
 ; UNBALANCED1: llvm-rc: Error parsing file: expected ')', got ,
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-expr-unbalanced-2.rc 2>&1 | FileCheck %s --check-prefix UNBALANCED2
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-expr-unbalanced-2.rc 2>&1 | FileCheck %s --check-prefix UNBALANCED2
 
 ; UNBALANCED2: llvm-rc: Error parsing file: expected ',', got )
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-expr-unbalanced-3.rc 2>&1 | FileCheck %s --check-prefix UNBALANCED3
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-expr-unbalanced-3.rc 2>&1 | FileCheck %s --check-prefix UNBALANCED3
 
 ; UNBALANCED3: llvm-rc: Error parsing file: expected ',', got )

llvm/test/tools/llvm-rc/parser.test

-; RUN: llvm-rc /V %p/Inputs/parser-correct-everything.rc | FileCheck %s --check-prefix PGOOD
+; RUN: llvm-rc /dry-run /V %p/Inputs/parser-correct-everything.rc | FileCheck %s --check-prefix PGOOD
 
 ; PGOOD:  Icon (meh): "hello.bmp"
 ; PGOOD-NEXT:  Icon (Icon): "Icon"
 ; PGOOD-NEXT:  Language: 5, Sublanguage: 12
 ; PGOOD-NEXT:  StringTable:
 ; PGOOD-NEXT:    Option: Language: 1, Sublanguage: 1
 ; PGOOD-NEXT:    Option: Characteristics: 500
 ; PGOOD-NEXT:    Option: Language: 3, Sublanguage: 4

 ; PGOOD-NEXT:  User-defined (type: 600, name: 500): "other filename"
 ; PGOOD-NEXT:  User-defined (type: INTEGERS, name: HELLO): data = 1 2 3 4
 ; PGOOD-NEXT:  User-defined (type: STRINGS, name: HELLO): data = "1" "2" "3" "4"
 ; PGOOD-NEXT:  User-defined (type: MIXED, name: 4): data = 1 "2" 3 "4"
 
 
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-stringtable-no-string.rc 2>&1 | FileCheck %s --check-prefix PSTRINGTABLE1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-stringtable-no-string.rc 2>&1 | FileCheck %s --check-prefix PSTRINGTABLE1
 
 ; PSTRINGTABLE1:  llvm-rc: Error parsing file: expected string, got }
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-stringtable-weird-option.rc 2>&1 | FileCheck %s --check-prefix PSTRINGTABLE2
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-stringtable-weird-option.rc 2>&1 | FileCheck %s --check-prefix PSTRINGTABLE2
 
 ; PSTRINGTABLE2:  llvm-rc: Error parsing file: expected optional statement type, BEGIN or '{', got NONSENSETYPE
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-eof.rc 2>&1 | FileCheck %s --check-prefix PEOF
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-eof.rc 2>&1 | FileCheck %s --check-prefix PEOF
 
 ; PEOF:  llvm-rc: Error parsing file: expected '-', '~', integer or '(', got <EOF>
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-no-characteristics-arg.rc 2>&1 | FileCheck %s --check-prefix PCHARACTERISTICS1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-no-characteristics-arg.rc 2>&1 | FileCheck %s --check-prefix PCHARACTERISTICS1
 
 ; PCHARACTERISTICS1:  llvm-rc: Error parsing file: expected '-', '~', integer or '(', got BEGIN
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-nonsense-token.rc 2>&1 | FileCheck %s --check-prefix PNONSENSE1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-nonsense-token.rc 2>&1 | FileCheck %s --check-prefix PNONSENSE1
 
 ; PNONSENSE1:  llvm-rc: Error parsing file: expected int or identifier, got &
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-nonsense-type.rc 2>&1 | FileCheck %s --check-prefix PNONSENSE2
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-nonsense-type.rc 2>&1 | FileCheck %s --check-prefix PNONSENSE2
 
 ; PNONSENSE2:  llvm-rc: Error parsing file: expected filename, '{' or BEGIN, got <EOF>
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-nonsense-type-eof.rc 2>&1 | FileCheck %s --check-prefix PNONSENSE3
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-nonsense-type-eof.rc 2>&1 | FileCheck %s --check-prefix PNONSENSE3
 
 ; PNONSENSE3:  llvm-rc: Error parsing file: expected int or identifier, got <EOF>
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-language-no-comma.rc 2>&1 | FileCheck %s --check-prefix PLANGUAGE1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-language-no-comma.rc 2>&1 | FileCheck %s --check-prefix PLANGUAGE1
 
 ; PLANGUAGE1:  llvm-rc: Error parsing file: expected ',', got 7
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-language-too-many-commas.rc 2>&1 | FileCheck %s --check-prefix PLANGUAGE2
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-language-too-many-commas.rc 2>&1 | FileCheck %s --check-prefix PLANGUAGE2
 
 ; PLANGUAGE2:  llvm-rc: Error parsing file: expected '-', '~', integer or '(', got ,
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-html-bad-string.rc 2>&1 | FileCheck %s --check-prefix PHTML1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-html-bad-string.rc 2>&1 | FileCheck %s --check-prefix PHTML1
 
 ; PHTML1:  llvm-rc: Error parsing file: expected string, got ThisPassesInTheOriginalToolButDocSaysItShouldBeQuoted
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-html-extra-comma.rc 2>&1 | FileCheck %s --check-prefix PHTML2
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-html-extra-comma.rc 2>&1 | FileCheck %s --check-prefix PHTML2
 
 ; PHTML2:  llvm-rc: Error parsing file: expected string, got ,
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-accelerators-bad-flag.rc 2>&1 | FileCheck %s --check-prefix PACCELERATORS1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-accelerators-bad-flag.rc 2>&1 | FileCheck %s --check-prefix PACCELERATORS1
 
 ; PACCELERATORS1:  llvm-rc: Error parsing file: expected ASCII/VIRTKEY/NOINVERT/ALT/SHIFT/CONTROL, got HELLO
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-accelerators-bad-int-or-string.rc 2>&1 | FileCheck %s --check-prefix PACCELERATORS2
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-accelerators-bad-int-or-string.rc 2>&1 | FileCheck %s --check-prefix PACCELERATORS2
 
 ; PACCELERATORS2:  llvm-rc: Error parsing file: expected int or string, got NotIntOrString
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-accelerators-no-comma.rc 2>&1 | FileCheck %s --check-prefix PACCELERATORS3
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-accelerators-no-comma.rc 2>&1 | FileCheck %s --check-prefix PACCELERATORS3
 
 ; PACCELERATORS3:  llvm-rc: Error parsing file: expected int or string, got CONTROL
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-accelerators-no-comma-2.rc 2>&1 | FileCheck %s --check-prefix PACCELERATORS4
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-accelerators-no-comma-2.rc 2>&1 | FileCheck %s --check-prefix PACCELERATORS4
 
 ; PACCELERATORS4:  llvm-rc: Error parsing file: expected ',', got 10
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-menu-bad-id.rc 2>&1 | FileCheck %s --check-prefix PMENU1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-menu-bad-id.rc 2>&1 | FileCheck %s --check-prefix PMENU1
 
 ; PMENU1:  llvm-rc: Error parsing file: expected '-', '~', integer or '(', got A
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-menu-bad-flag.rc 2>&1 | FileCheck %s --check-prefix PMENU2
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-menu-bad-flag.rc 2>&1 | FileCheck %s --check-prefix PMENU2
 
 ; PMENU2:  llvm-rc: Error parsing file: expected CHECKED/GRAYED/HELP/INACTIVE/MENUBARBREAK/MENUBREAK, got ERRONEOUS
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-menu-missing-block.rc 2>&1 | FileCheck %s --check-prefix PMENU3
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-menu-missing-block.rc 2>&1 | FileCheck %s --check-prefix PMENU3
 
 ; PMENU3:  llvm-rc: Error parsing file: expected '{', got POPUP
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-menu-misspelled-separator.rc 2>&1 | FileCheck %s --check-prefix PMENU4
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-menu-misspelled-separator.rc 2>&1 | FileCheck %s --check-prefix PMENU4
 
 ; PMENU4:  llvm-rc: Error parsing file: expected SEPARATOR or string, got NOTSEPARATOR
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-dialog-cant-give-helpid.rc 2>&1 | FileCheck %s --check-prefix PDIALOG1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-dialog-cant-give-helpid.rc 2>&1 | FileCheck %s --check-prefix PDIALOG1
 
 ; PDIALOG1:  llvm-rc: Error parsing file: expected identifier, got ,
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-dialog-too-few-args.rc 2>&1 | FileCheck %s --check-prefix PDIALOG2
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-dialog-too-few-args.rc 2>&1 | FileCheck %s --check-prefix PDIALOG2
 
 ; PDIALOG2:  llvm-rc: Error parsing file: expected ',', got }
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-dialog-too-many-args.rc 2>&1 | FileCheck %s --check-prefix PDIALOG3
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-dialog-too-many-args.rc 2>&1 | FileCheck %s --check-prefix PDIALOG3
 
 ; PDIALOG3:  llvm-rc: Error parsing file: expected identifier, got ,
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-dialog-unknown-type.rc 2>&1 | FileCheck %s --check-prefix PDIALOG4
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-dialog-unknown-type.rc 2>&1 | FileCheck %s --check-prefix PDIALOG4
 
 ; PDIALOG4:  llvm-rc: Error parsing file: expected control type, END or '}', got UNKNOWN
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-dialog-unnecessary-string.rc 2>&1 | FileCheck %s --check-prefix PDIALOG5
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-dialog-unnecessary-string.rc 2>&1 | FileCheck %s --check-prefix PDIALOG5
 
 ; PDIALOG5:  llvm-rc: Error parsing file: expected '-', '~', integer or '(', got "This shouldn't be here"
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-versioninfo-wrong-fixed.rc 2>&1 | FileCheck %s --check-prefix PVERSIONINFO1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-versioninfo-wrong-fixed.rc 2>&1 | FileCheck %s --check-prefix PVERSIONINFO1
 
 ; PVERSIONINFO1:  llvm-rc: Error parsing file: expected fixed VERSIONINFO statement type, got WEIRDFIXED
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-versioninfo-named-main-block.rc 2>&1 | FileCheck %s --check-prefix PVERSIONINFO2
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-versioninfo-named-main-block.rc 2>&1 | FileCheck %s --check-prefix PVERSIONINFO2
 
 ; PVERSIONINFO2:  llvm-rc: Error parsing file: expected fixed VERSIONINFO statement type, got BLOCK
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-versioninfo-unnamed-inner-block.rc 2>&1 | FileCheck %s --check-prefix PVERSIONINFO3
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-versioninfo-unnamed-inner-block.rc 2>&1 | FileCheck %s --check-prefix PVERSIONINFO3
 
 ; PVERSIONINFO3:  llvm-rc: Error parsing file: expected string, got {
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-versioninfo-unnamed-value.rc 2>&1 | FileCheck %s --check-prefix PVERSIONINFO4
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-versioninfo-unnamed-value.rc 2>&1 | FileCheck %s --check-prefix PVERSIONINFO4
 
 ; PVERSIONINFO4:  llvm-rc: Error parsing file: expected string, got END
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-versioninfo-bad-type.rc 2>&1 | FileCheck %s --check-prefix PVERSIONINFO5
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-versioninfo-bad-type.rc 2>&1 | FileCheck %s --check-prefix PVERSIONINFO5
 
 ; PVERSIONINFO5:  llvm-rc: Error parsing file: expected BLOCK or VALUE, got INCORRECT
 
 
-; RUN: not llvm-rc /V %p/Inputs/parser-user-invalid-contents.rc 2>&1 | FileCheck %s --check-prefix PUSER1
+; RUN: not llvm-rc /dry-run /V %p/Inputs/parser-user-invalid-contents.rc 2>&1 | FileCheck %s --check-prefix PUSER1
 
 ; PUSER1:  llvm-rc: Error parsing file: expected int or string, got InvalidToken

llvm/test/tools/llvm-rc/tag-html.test

+; RUN: rm -rf %t && mkdir %t && cd %t
+; RUN: cp %p/Inputs/webpage*.html .
+; RUN: llvm-rc /FO %t/tag-html.res %p/Inputs/tag-html.rc
+; RUN: llvm-readobj %t/tag-html.res | FileCheck %s --check-prefix HTML
+
+; HTML: Resource type (int): 23
+; HTML-NEXT: Resource name (int): 100
+; HTML-NEXT: Data version: 0
+; HTML-NEXT: Memory flags: 0x30
+; HTML-NEXT: Language ID: 1033
+; HTML-NEXT: Version (major): 0
+; HTML-NEXT: Version (minor): 0
+; HTML-NEXT: Characteristics: 0
+; HTML-NEXT: Data size: 45
+; HTML-NEXT: Data: (
+; HTML-NEXT:   0000: 3C68746D 6C3E0A20 203C626F 64793E0A  |<html>.  <body>.|
+; HTML-NEXT:   0010: 20202020 48656C6C 6F210A20 203C2F62  |    Hello!.  </b|
+; HTML-NEXT:   0020: 6F64793E 0A3C2F68 746D6C3E 0A        |ody>.</html>.|
+; HTML-NEXT: )
+
+; HTML-DAG: Resource type (int): 23
+; HTML-NEXT: Resource name (string): KITTEN
+; HTML-NEXT: Data version: 0
+; HTML-NEXT: Memory flags: 0x30
+; HTML-NEXT: Language ID: 1033
+; HTML-NEXT: Version (major): 0
+; HTML-NEXT: Version (minor): 0
+; HTML-NEXT: Characteristics: 0
+; HTML-NEXT: Data size: 61
+; HTML-NEXT: Data: (
+; HTML-NEXT:   0000: 3C212D2D 2053686F 756C6420 6E6F7420  |<!-- Should not |
+; HTML-NEXT:   0010: 656D6265 64207468 6520696D 6167652E  |embed the image.|
+; HTML-NEXT:   0020: 202D2D3E 0A3C696D 67207372 633D226B  | -->.<img src="k|
+; HTML-NEXT:   0030: 69747465 6E732E62 6D70223E 0A        |ittens.bmp">.|
+; HTML-NEXT: )
+
+
+; RUN: not llvm-rc /FO %t/tag-html-wrong.res %p/Inputs/tag-html-wrong.rc 2>&1 | FileCheck %s --check-prefix NOFILE
+
+; NOFILE: llvm-rc: Error in HTML statement (ID 1):
+; NOFILE-NEXT: Error opening file 'some-really-nonexistent-file.html':

llvm/tools/llvm-rc/CMakeLists.txt

 set(LLVM_LINK_COMPONENTS
   Option
   Support
   )
 
 set(LLVM_TARGET_DEFINITIONS Opts.td)
 
 tablegen(LLVM Opts.inc -gen-opt-parser-defs)
 add_public_tablegen_target(RcTableGen)
 
 add_llvm_tool(llvm-rc
   llvm-rc.cpp
+  ResourceFileWriter.cpp
   ResourceScriptParser.cpp
   ResourceScriptStmt.cpp
   ResourceScriptToken.cpp
   )

llvm/tools/llvm-rc/Opts.td

def DUPID_NOWARN : Flag<[ "/", "-" ], "Y">,
                    HelpText<"Suppress warnings on duplicate resource IDs.">;
 
 def VERBOSE : Flag<[ "/", "-" ], "V">, HelpText<"Be verbose.">;
 def HELP : Flag<[ "/", "-" ], "?">, HelpText<"Display this help and exit.">;
 def H : Flag<[ "/", "-" ], "H">,
         Alias<HELP>,
         HelpText<"Display this help and exit.">;
 
+def DRY_RUN : Flag<[ "/", "-" ], "dry-run">,
+              HelpText<"Don't compile the input; only try to parse it.">;
+
 // Unused switches (at least for now). These will stay unimplemented
 // in an early stage of development and can be ignored. However, we need to
 // parse them in order to preserve the compatibility with the original tool.
 
 def NOLOGO : Flag<[ "/", "-" ], "NOLOGO">;
 def R : Flag<[ "/", "-" ], "R">;
 def SL : Flag<[ "/", "-" ], "SL">;
 

llvm/tools/llvm-rc/ResourceFileWriter.h

+//===-- ResourceSerializator.h ----------------------------------*- C++-*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// This defines a visitor serializing resources to a .res stream.
+//
+//===---------------------------------------------------------------------===//
+
+#ifndef LLVM_TOOLS_LLVMRC_RESOURCESERIALIZATOR_H
+#define LLVM_TOOLS_LLVMRC_RESOURCESERIALIZATOR_H
+
+#include "ResourceScriptStmt.h"
+#include "ResourceVisitor.h"
+
+#include "llvm/Support/Endian.h"
+
+namespace llvm {
+namespace rc {
+
+class ResourceFileWriter : public Visitor {
+public:
+  ResourceFileWriter(std::unique_ptr<raw_fd_ostream> Stream)
+      : FS(std::move(Stream)) {
+    assert(FS && "Output stream needs to be provided to the serializator");
+  }
+
+  Error visitNullResource(const RCResource *) override;
+  Error visitHTMLResource(const RCResource *) override;
+
+  Error visitLanguageStmt(const LanguageResource *) override;
+
+  struct ObjectInfo {
+    uint16_t LanguageInfo;
+
+    ObjectInfo() : LanguageInfo(0) {}
+  } ObjectData;
+
+private:
+  Error handleError(Error &&Err, const RCResource *Res);
+
+  Error
+  writeResource(const RCResource *Res,
+                Error (ResourceFileWriter::*BodyWriter)(const RCResource *));
+
+  Error writeNullBody(const RCResource *);
+  Error writeHTMLBody(const RCResource *);
+
+  // Output stream handling.
+  std::unique_ptr<raw_fd_ostream> FS;
+
+  uint64_t tell() const { return FS->tell(); }
+
+  uint64_t writeObject(const ArrayRef<uint8_t> Data);
+
+  template <typename T> uint64_t writeInt(const T &Value) {
+    support::detail::packed_endian_specific_integral<T, support::little,
+                                                     support::unaligned>
+        Object(Value);
+    return writeObject(Object);
+  }
+
+  template <typename T> uint64_t writeObject(const T &Value) {
+    return writeObject(ArrayRef<uint8_t>(
+        reinterpret_cast<const uint8_t *>(&Value), sizeof(T)));
+  }
+
+  template <typename T> void writeObjectAt(const T &Value, uint64_t Position) {
+    FS->pwrite((const char *)&Value, sizeof(T), Position);
+  }
+
+  Error writeCString(StringRef Str, bool WriteTerminator = true);
+
+  Error writeIdentifier(const IntOrString &Ident);
+  Error writeIntOrString(const IntOrString &Data);
+
+  Error appendFile(StringRef Filename);
+
+  void padStream(uint64_t Length);
+};
+
+} // namespace rc
+} // namespace llvm
+
+#endif

llvm/tools/llvm-rc/ResourceFileWriter.cpp

+//===-- ResourceFileWriter.cpp --------------------------------*- C++-*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// This implements the visitor serializing resources to a .res stream.
+//
+//===---------------------------------------------------------------------===//
+
+#include "ResourceFileWriter.h"
+
+#include "llvm/Object/WindowsResource.h"
+#include "llvm/Support/ConvertUTF.h"
+#include "llvm/Support/Endian.h"
+#include "llvm/Support/EndianStream.h"
+
+using namespace llvm::support;
+
+// Take an expression returning llvm::Error and forward the error if it exists.
+#define RETURN_IF_ERROR(Expr)                                                  \
+  if (auto Err = (Expr))                                                       \
+    return Err;
+
+namespace llvm {
+namespace rc {
+
+static Error createError(Twine Message,
+                         std::errc Type = std::errc::invalid_argument) {
+  return make_error<StringError>(Message, std::make_error_code(Type));
+}
+
+static Error checkNumberFits(uint32_t Number, size_t MaxBits, Twine FieldName) {
+  assert(1 <= MaxBits && MaxBits <= 32);
+  if (!(Number >> MaxBits))
+    return Error::success();
+  return createError(FieldName + " (" + Twine(Number) + ") does not fit in " +
+                         Twine(MaxBits) + " bits.",
+                     std::errc::value_too_large);
+}
+
+template <typename FitType>
+static Error checkNumberFits(uint32_t Number, Twine FieldName) {
+  return checkNumberFits(Number, sizeof(FitType) * 8, FieldName);
+}
+
+static Error checkIntOrString(IntOrString Value, Twine FieldName) {
+  if (!Value.isInt())
+    return Error::success();
+  return checkNumberFits<uint16_t>(Value.getInt(), FieldName);
+}
+
+static bool stripQuotes(StringRef &Str, bool &IsLongString) {
+  if (!Str.contains('"'))
+    return false;
+
+  // Just take the contents of the string, checking if it's been marked long.
+  IsLongString = Str.startswith_lower("L");
+  if (IsLongString)
+    Str = Str.drop_front();
+
+  bool StripSuccess = Str.consume_front("\"") && Str.consume_back("\"");
+  (void)StripSuccess;
+  assert(StripSuccess && "Strings should be enclosed in quotes.");
+  return true;
+}
+
+// Describes a way to handle '\0' characters when processing the string.
+// rc.exe tool sometimes behaves in a weird way in postprocessing.
+// If the string to be output is equivalent to a C-string (e.g. in MENU
+// titles), string is (predictably) truncated after first 0-byte.
+// When outputting a string table, the behavior is equivalent to appending
+// '\0\0' at the end of the string, and then stripping the string
+// before the first '\0\0' occurrence.
+// Finally, when handling strings in user-defined resources, 0-bytes
+// aren't stripped, nor do they terminate the string.
+
+enum class NullHandlingMethod {
+  UserResource,   // Don't terminate string on '\0'.
+  CutAtNull,      // Terminate string on '\0'.
+  CutAtDoubleNull // Terminate string on '\0\0'; strip final '\0'.
+};
+
+// Parses an identifier or string and returns a processed version of it.
+// For now, it only strips the string boundaries, but TODO:
+//   * Squash "" to a single ".
+//   * Replace the escape sequences with their processed version.
+// For identifiers, this is no-op.
+static Error processString(StringRef Str, NullHandlingMethod NullHandler,
+                           bool &IsLongString, SmallVectorImpl<UTF16> &Result) {
+  assert(NullHandler == NullHandlingMethod::CutAtNull);
+
+  bool IsString = stripQuotes(Str, IsLongString);
+  convertUTF8ToUTF16String(Str, Result);
+
+  if (!IsString) {
+    // It's an identifier if it's not a string. Make all characters uppercase.
+    for (UTF16 &Ch : Result) {
+      assert(Ch <= 0x7F && "We didn't allow identifiers to be non-ASCII");
+      Ch = toupper(Ch);
+    }
+    return Error::success();
+  }
+
+  // We don't process the string contents. Only cut at '\0'.
+
+  for (size_t Pos = 0; Pos < Result.size(); ++Pos)
+    if (Result[Pos] == '\0')
+      Result.resize(Pos);
+
+  return Error::success();
+}
+
+uint64_t ResourceFileWriter::writeObject(const ArrayRef<uint8_t> Data) {
+  uint64_t Result = tell();
+  FS->write((const char *)Data.begin(), Data.size());
+  return Result;
+}
+
+Error ResourceFileWriter::writeCString(StringRef Str, bool WriteTerminator) {
+  SmallVector<UTF16, 128> ProcessedString;
+  bool IsLongString;
+  RETURN_IF_ERROR(processString(Str, NullHandlingMethod::CutAtNull,
+                                IsLongString, ProcessedString));
+  for (auto Ch : ProcessedString)
+    writeInt<uint16_t>(Ch);
+  if (WriteTerminator)
+    writeInt<uint16_t>(0);
+  return Error::success();
+}
+
+Error ResourceFileWriter::writeIdentifier(const IntOrString &Ident) {
+  return writeIntOrString(Ident);
+}
+
+Error ResourceFileWriter::writeIntOrString(const IntOrString &Value) {
+  if (!Value.isInt())
+    return writeCString(Value.getString());
+
+  writeInt<uint16_t>(0xFFFF);
+  writeInt<uint16_t>(Value.getInt());
+  return Error::success();
+}
+
+Error ResourceFileWriter::appendFile(StringRef Filename) {
+  bool IsLong;
+  stripQuotes(Filename, IsLong);
+
+  // Filename path should be relative to the current working directory.
+  // FIXME: docs say so, but reality is more complicated, script
+  // location and include paths must be taken into account.
+  ErrorOr<std::unique_ptr<MemoryBuffer>> File =
+      MemoryBuffer::getFile(Filename, -1, false);
+  if (!File)
+    return make_error<StringError>("Error opening file '" + Filename +
+                                       "': " + File.getError().message(),
+                                   File.getError());
+  *FS << (*File)->getBuffer();
+  return Error::success();
+}
+
+void ResourceFileWriter::padStream(uint64_t Length) {
+  assert(Length > 0);
+  uint64_t Location = tell();
+  Location %= Length;
+  uint64_t Pad = (Length - Location) % Length;
+  for (uint64_t i = 0; i < Pad; ++i)
+    writeInt<uint8_t>(0);
+}
+
+Error ResourceFileWriter::handleError(Error &&Err, const RCResource *Res) {
+  if (Err)
+    return joinErrors(createError("Error in " + Res->getResourceTypeName() +
+                                  " statement (ID " + Twine(Res->ResName) +
+                                  "): "),
+                      std::move(Err));
+  return Error::success();
+}
+
+Error ResourceFileWriter::visitNullResource(const RCResource *Res) {
+  return writeResource(Res, &ResourceFileWriter::writeNullBody);
+}
+
+Error ResourceFileWriter::visitHTMLResource(const RCResource *Res) {
+  return writeResource(Res, &ResourceFileWriter::writeHTMLBody);
+}
+
+Error ResourceFileWriter::visitLanguageStmt(const LanguageResource *Stmt) {
+  RETURN_IF_ERROR(checkNumberFits(Stmt->Lang, 10, "Primary language ID"));
+  RETURN_IF_ERROR(checkNumberFits(Stmt->SubLang, 6, "Sublanguage ID"));
+  ObjectData.LanguageInfo = Stmt->Lang | (Stmt->SubLang << 10);
+  return Error::success();
+}
+
+Error ResourceFileWriter::writeResource(
+    const RCResource *Res,
+    Error (ResourceFileWriter::*BodyWriter)(const RCResource *)) {
+  // We don't know the sizes yet.
+  object::WinResHeaderPrefix HeaderPrefix{ulittle32_t(0U), ulittle32_t(0U)};
+  uint64_t HeaderLoc = writeObject(HeaderPrefix);
+
+  auto ResType = Res->getResourceType();
+  RETURN_IF_ERROR(checkIntOrString(ResType, "Resource type"));
+  RETURN_IF_ERROR(checkIntOrString(Res->ResName, "Resource ID"));
+  RETURN_IF_ERROR(handleError(writeIdentifier(ResType), Res));
+  RETURN_IF_ERROR(handleError(writeIdentifier(Res->ResName), Res));
+
+  padStream(sizeof(uint32_t));
+  object::WinResHeaderSuffix HeaderSuffix{
+      ulittle32_t(0), // DataVersion; seems to always be 0
+      ulittle16_t(Res->getMemoryFlags()), ulittle16_t(ObjectData.LanguageInfo),
+      ulittle32_t(0),  // VersionInfo
+      ulittle32_t(0)}; // Characteristics
+  writeObject(HeaderSuffix);
+
+  uint64_t DataLoc = tell();
+  RETURN_IF_ERROR(handleError((this->*BodyWriter)(Res), Res));
+  // RETURN_IF_ERROR(handleError(dumpResource(Ctx)));
+
+  // Update the sizes.
+  HeaderPrefix.DataSize = tell() - DataLoc;
+  HeaderPrefix.HeaderSize = DataLoc - HeaderLoc;
+  writeObjectAt(HeaderPrefix, HeaderLoc);
+  padStream(sizeof(uint32_t));
+
+  return Error::success();
+}
+
+Error ResourceFileWriter::writeNullBody(const RCResource *) {
+  return Error::success();
+}
+
+Error ResourceFileWriter::writeHTMLBody(const RCResource *Base) {
+  return appendFile(cast<HTMLResource>(Base)->HTMLLoc);
+}
+
+} // namespace rc
+} // namespace llvm

llvm/tools/llvm-rc/ResourceScriptStmt.h

 // This lists all the resource and statement types occurring in RC scripts.
 //
 //===---------------------------------------------------------------------===//
 
 #ifndef LLVM_TOOLS_LLVMRC_RESOURCESCRIPTSTMT_H
 #define LLVM_TOOLS_LLVMRC_RESOURCESCRIPTSTMT_H
 
 #include "ResourceScriptToken.h"
+#include "ResourceVisitor.h"
 
 #include "llvm/ADT/StringSet.h"
 
 namespace llvm {
 namespace rc {
 
 // A class holding a name - either an integer or a reference to the string.
 class IntOrString {
public:
   IntOrString(StringRef Value) : Data(Value), IsInt(0) {}
   IntOrString(const RCToken &Token)
       : Data(Token), IsInt(Token.kind() == RCToken::Kind::Int) {}
 
   bool equalsLower(const char *Str) {
     return !IsInt && Data.String.equals_lower(Str);
   }
 
+  bool isInt() const { return IsInt; }
+
+  uint32_t getInt() const {
+    assert(IsInt);
+    return Data.Int;
+  }
+
+  const StringRef &getString() const {
+    assert(!IsInt);
+    return Data.String;
+  }
+
+  operator Twine() const {
+    return isInt() ? Twine(getInt()) : Twine(getString());
+  }
+
   friend raw_ostream &operator<<(raw_ostream &, const IntOrString &);
 };
 
+enum ResourceKind {
+  // These resource kinds have corresponding .res resource type IDs
+  // (TYPE in RESOURCEHEADER structure). The numeric value assigned to each
+  // kind is equal to this type ID.
+  RkNull = 0,
+  RkMenu = 4,
+  RkDialog = 5,
+  RkAccelerators = 9,
+  RkVersionInfo = 16,
+  RkHTML = 23,
+
+  // These kinds don't have assigned type IDs (they might be the resources
+  // of invalid kind, expand to many resource structures in .res files,
+  // or have variable type ID). In order to avoid ID clashes with IDs above,
+  // we assign the kinds the values 256 and larger.
+  RkInvalid = 256,
+  RkBase,
+  RkCursor,
+  RkIcon,
+  RkUser
+};
+
+// Non-zero memory flags.
+// Ref: msdn.microsoft.com/en-us/library/windows/desktop/ms648027(v=vs.85).aspx
+enum MemoryFlags {
+  MfMoveable = 0x10,
+  MfPure = 0x20,
+  MfPreload = 0x40,
+  MfDiscardable = 0x1000
+};
+
 // Base resource. All the resources should derive from this base.
 class RCResource {
-protected:
+public:
   IntOrString ResName;
 
-public:
   RCResource() = default;
   RCResource(RCResource &&) = default;
   void setName(const IntOrString &Name) { ResName = Name; }
   virtual raw_ostream &log(raw_ostream &OS) const {
     return OS << "Base statement\n";
   };
   virtual ~RCResource() {}
+
+  virtual Error visit(Visitor *) const {
+    llvm_unreachable("This is unable to call methods from Visitor base");
+  }
+
+  // By default, memory flags are DISCARDABLE | PURE | MOVEABLE.
+  virtual uint16_t getMemoryFlags() const {
+    return MfDiscardable | MfPure | MfMoveable;
+  }
+  virtual ResourceKind getKind() const { return RkBase; }
+  static bool classof(const RCResource *Res) { return true; }
+
+  virtual IntOrString getResourceType() const {
+    llvm_unreachable("This cannot be called on objects without types.");
+  }
+  virtual Twine getResourceTypeName() const {
+    llvm_unreachable("This cannot be called on objects without types.");
+  };
+};
+
+// An empty resource. It has no content, type 0, ID 0 and all of its
+// characteristics are equal to 0.
+class NullResource : public RCResource {
+public:
+  raw_ostream &log(raw_ostream &OS) const override {
+    return OS << "Null resource\n";
+  }
+  Error visit(Visitor *V) const override { return V->visitNullResource(this); }
+  IntOrString getResourceType() const override { return 0; }
+  Twine getResourceTypeName() const override { return "(NULL)"; }
+  uint16_t getMemoryFlags() const override { return 0; }
 };
 
 // Optional statement base. All such statements should derive from this base.
 class OptionalStmt : public RCResource {};
 
 class OptionalStmtList : public OptionalStmt {
   std::vector<std::unique_ptr<OptionalStmt>> Statements;
 
 public:
   OptionalStmtList() {}
   virtual raw_ostream &log(raw_ostream &OS) const;
 
   void addStmt(std::unique_ptr<OptionalStmt> Stmt) {
     Statements.push_back(std::move(Stmt));
   }
 };
 
 // LANGUAGE statement. It can occur both as a top-level statement (in such
 // a situation, it changes the default language until the end of the file)
 // and as an optional resource statement (then it changes the language
 // of a single resource).
 //
 // Ref: msdn.microsoft.com/en-us/library/windows/desktop/aa381019(v=vs.85).aspx
 class LanguageResource : public OptionalStmt {
+public:
   uint32_t Lang, SubLang;
 
-public:
   LanguageResource(uint32_t LangId, uint32_t SubLangId)
       : Lang(LangId), SubLang(SubLangId) {}
   raw_ostream &log(raw_ostream &) const override;
+
+  // This is not a regular top-level statement; when it occurs, it just
+  // modifies the language context.
+  Error visit(Visitor *V) const override { return V->visitLanguageStmt(this); }
+  Twine getResourceTypeName() const override { return "LANGUAGE"; }
 };
 
 // ACCELERATORS resource. Defines a named table of accelerators for the app.
 //
 // Ref: msdn.microsoft.com/en-us/library/windows/desktop/aa380610(v=vs.85).aspx
 class AcceleratorsResource : public RCResource {
 public:
   class Accelerator {

 };
 
 // HTML resource. Represents a local webpage that is to be embedded into the
 // resulting resource file. It embeds a file only - no additional resources
 // (images etc.) are included with this resource.
 //
 // Ref: msdn.microsoft.com/en-us/library/windows/desktop/aa966018(v=vs.85).aspx
 class HTMLResource : public RCResource {
+public:
   StringRef HTMLLoc;
 
-public:
   HTMLResource(StringRef Location) : HTMLLoc(Location) {}
   raw_ostream &log(raw_ostream &) const override;
+
+  Error visit(Visitor *V) const override { return V->visitHTMLResource(this); }
+
+  // Curiously, file resources don't have DISCARDABLE flag set.
+  uint16_t getMemoryFlags() const override { return MfPure | MfMoveable; }
+  IntOrString getResourceType() const override { return RkHTML; }
+  Twine getResourceTypeName() const override { return "HTML"; }
+  ResourceKind getKind() const override { return RkHTML; }
+  static bool classof(const RCResource *Res) {
+    return Res->getKind() == RkHTML;
+  }
 };
 
 // -- MENU resource and its helper classes --
 // This resource describes the contents of an application menu
 // (usually located in the upper part of the dialog.)
 //
 // Ref: msdn.microsoft.com/en-us/library/windows/desktop/aa381025(v=vs.85).aspx
 

llvm/tools/llvm-rc/ResourceVisitor.h

+//===-- ResourceVisitor.h ---------------------------------------*- C++-*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// This defines a base class visiting resource script resources.
+//
+//===---------------------------------------------------------------------===//
+
+#ifndef LLVM_TOOLS_LLVMRC_RESOURCEVISITOR_H
+#define LLVM_TOOLS_LLVMRC_RESOURCEVISITOR_H
+
+#include "llvm/Support/Error.h"
+
+namespace llvm {
+namespace rc {
+
+class RCResource;
+
+class LanguageResource;
+
+class Visitor {
+public:
+  virtual Error visitNullResource(const RCResource *) = 0;
+  virtual Error visitHTMLResource(const RCResource *) = 0;
+
+  virtual Error visitLanguageStmt(const LanguageResource *) = 0;
+
+  virtual ~Visitor() {}
+};
+
+} // namespace rc
+} // namespace llvm
+
+#endif

llvm/tools/llvm-rc/llvm-rc.cpp

 //===-- llvm-rc.cpp - Compile .rc scripts into .res -------------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 //
 // Compile .rc scripts into .res files. This is intended to be a
 // platform-independent port of Microsoft's rc.exe tool.
 //
 //===----------------------------------------------------------------------===//
 
-#include "ResourceScriptToken.h"
+#include "ResourceFileWriter.h"
 #include "ResourceScriptParser.h"
+#include "ResourceScriptStmt.h"
+#include "ResourceScriptToken.h"
 
 #include "llvm/Option/Arg.h"
 #include "llvm/Option/ArgList.h"
 #include "llvm/Support/Error.h"
+#include "llvm/Support/FileSystem.h"
 #include "llvm/Support/ManagedStatic.h"
 #include "llvm/Support/PrettyStackTrace.h"
 #include "llvm/Support/Process.h"
 #include "llvm/Support/Signals.h"
 #include "llvm/Support/raw_ostream.h"
 
 #include <system_error>
 
 using namespace llvm;
+using namespace llvm::rc;
 
 namespace {
 
 // Input options tables.
 
 enum ID {
   OPT_INVALID = 0, // This is not a correct option ID.
 #define OPTION(PREFIX, NAME, ID, KIND, GROUP, ALIAS, ALIASARGS, FLAGS, PARAM,  \
for (const RCToken &Token : Tokens) {
              << Token.value();
       if (Token.kind() == RCToken::Kind::Int)
         outs() << "; int value = " << Token.intValue();
 
       outs() << "\n";
     }
   }
 
+  std::unique_ptr<ResourceFileWriter> Visitor;
+  bool IsDryRun = InputArgs.hasArg(OPT_DRY_RUN);
+
+  if (!IsDryRun) {
+    auto OutArgsInfo = InputArgs.getAllArgValues(OPT_FILEOUT);
+    if (OutArgsInfo.size() != 1)
+      fatalError(
+          "Exactly one output file should be provided (using /FO flag).");
+
+    std::error_code EC;
+    auto FOut =
+        llvm::make_unique<raw_fd_ostream>(OutArgsInfo[0], EC, sys::fs::F_RW);
+    if (EC)
+      fatalError("Error opening output file '" + OutArgsInfo[0] +
+                 "': " + EC.message());
+    Visitor = make_unique<ResourceFileWriter>(std::move(FOut));
+
+    ExitOnErr(NullResource().visit(Visitor.get()));
+
+    // Set the default language; choose en-US arbitrarily.
+    ExitOnErr(LanguageResource(0x09, 0x01).visit(Visitor.get()));
+  }
+
   rc::RCParser Parser{std::move(Tokens)};
   while (!Parser.isEof()) {
     auto Resource = ExitOnErr(Parser.parseSingleResource());
     if (BeVerbose)
       Resource->log(outs());
+    if (!IsDryRun)
+      ExitOnErr(Resource->visit(Visitor.get()));
   }
 
   return 0;
 }