Record type of content being downloaded.

chrome/browser/download/download_util.cc

Index: chrome/browser/download/download_util.cc
diff --git a/chrome/browser/download/download_util.cc b/chrome/browser/download/download_util.cc
index e1db3ffd38fe7e253f00c1b64ea722df4033df25..dec64a7c9e7e3dd1f3514187dc6760073a81a7c9 100644
--- a/chrome/browser/download/download_util.cc
+++ b/chrome/browser/download/download_util.cc
@@ -392,6 +392,75 @@ void RecordDownloadInterrupted(int error, int64 received, int64 total) {
   UMA_HISTOGRAM_BOOLEAN("Download.InterruptedUnknownSize", unknown_size);
 }
 
+namespace {
+
+enum DownloadContent {

[benjhayden, 2011/08/01 18:53:03]

Do all compilers number enums from 0 counting upwards? I'm thinking of
histograms.xml.

[cbentzel, 2011/08/01 19:00:33]

Not guaranteed - thanks for the catch.

+  DOWNLOAD_CONTENT_TEXT,
+  DOWNLOAD_CONTENT_IMAGE,
+  DOWNLOAD_CONTENT_AUDIO,
+  DOWNLOAD_CONTENT_VIDEO,
+  DOWNLOAD_CONTENT_OCTET_STREAM,
+  DOWNLOAD_CONTENT_PDF,
+  DOWNLOAD_CONTENT_DOC,
+  DOWNLOAD_CONTENT_XLS,
+  DOWNLOAD_CONTENT_PPT,
+  DOWNLOAD_CONTENT_ARCHIVE,
+  DOWNLOAD_CONTENT_UNRECOGNIZED,
+  DOWNLOAD_CONTENT_MAX,
+};
+
+struct MimeTypeToDownloadContent {
+  const char* mime_type;
+  DownloadContent download_content;
+};
+
+static MimeTypeToDownloadContent kMapMimeTypeToDownloadContent[] = {

[benjhayden, 2011/08/01 18:53:03]

Could you add exe, dmg, crx, and bzip?

exe: application/exe, application/x-exe, application/dos-exe, vms/exe,
application/x-winexe, application/msdos-windows, application/x-msdos-program

dmg: application/x-apple-diskimage

crx: application/x-chrome-extension

bzip: application/bzip2, application/x-bz2, application/x-bzip,
application/x-compressed, application/x-stuffit

[cbentzel, 2011/08/01 19:00:33]

Will do. Hoping I don't need to move to binary search or map approach. 

[cbentzel, 2011/08/01 21:06:18]

Do you know how frequently these happen? I mostly see octet-stream used instead.
Done.
Done.
I used x-bzip only, which is the standard mime type. Also mapped to ARCHIVE. 

[benjhayden, 2011/08/01 21:48:12]

No idea. I just copied it from some website. Don't sweat it too much.
SGTM

+  {"application/octet-stream", DOWNLOAD_CONTENT_OCTET_STREAM},
+  {"binary/octet-stream", DOWNLOAD_CONTENT_OCTET_STREAM},
+  {"application/pdf", DOWNLOAD_CONTENT_PDF},
+  {"application/msword", DOWNLOAD_CONTENT_DOC},
+  {"application/vnd.ms-excel", DOWNLOAD_CONTENT_XLS},
+  {"application/vns.ms-powerpoint", DOWNLOAD_CONTENT_PPT},
+  {"application/zip", DOWNLOAD_CONTENT_ARCHIVE},
+  {"application/x-gzip", DOWNLOAD_CONTENT_ARCHIVE},
+  {"application/x-rar-compressed", DOWNLOAD_CONTENT_ARCHIVE},
+  {"application/x-tar", DOWNLOAD_CONTENT_ARCHIVE},
+};
+
+}  // namespace
+
+void RecordDownloadMimeType(const std::string& mime_type_string) {
+  DownloadContent download_content = DOWNLOAD_CONTENT_UNRECOGNIZED;
+
+  // Look up exact matches.
+  for (size_t i = 0; i < arraysize(kMapMimeTypeToDownloadContent); ++i) {
+    const MimeTypeToDownloadContent& entry =
+        kMapMimeTypeToDownloadContent[i];
+    if (mime_type_string == entry.mime_type) {
+      download_content = entry.download_content;
+      break;
+    }
+  }
+
+  // Do partial matches.
+  if (download_content == DOWNLOAD_CONTENT_UNRECOGNIZED) {
+    if (StartsWithASCII(mime_type_string, "text/", true)) {
+      download_content = DOWNLOAD_CONTENT_TEXT;
+    } else if (StartsWithASCII(mime_type_string, "image/", true)) {
+      download_content = DOWNLOAD_CONTENT_IMAGE;
+    } else if (StartsWithASCII(mime_type_string, "audio/", true)) {
+      download_content = DOWNLOAD_CONTENT_AUDIO;
+    } else if (StartsWithASCII(mime_type_string, "video/", true)) {
+      download_content = DOWNLOAD_CONTENT_VIDEO;
+    }
+  }
+
+  // Record the value.
+  UMA_HISTOGRAM_ENUMERATION("Download.ContentType",
+                            download_content,
+                            DOWNLOAD_CONTENT_MAX);
+}
+
 // Download progress painting --------------------------------------------------
 
 // Common bitmaps used for download progress animations. We load them once the