{"id":783,"date":"2014-06-09T07:00:00","date_gmt":"2014-06-09T07:00:00","guid":{"rendered":"https:\/\/blogs.msdn.microsoft.com\/oldnewthing\/2014\/06\/09\/improving-the-performance-of-cf_hdrop-by-providing-file-attribute-information\/"},"modified":"2014-06-09T07:00:00","modified_gmt":"2014-06-09T07:00:00","slug":"improving-the-performance-of-cf_hdrop-by-providing-file-attribute-information","status":"publish","type":"post","link":"https:\/\/devblogs.microsoft.com\/oldnewthing\/20140609-00\/?p=783","title":{"rendered":"Improving the performance of CF_HDROP by providing file attribute information"},"content":{"rendered":"

\nThe CF_HDROP<\/code> clipboard format is still quite popular,\ndespite its limitation of being limited to files.\nYou can’t use it to represent\n\nvirtual content<\/a>,\nfor example.\n<\/p>\n

\nFor all of you still using\nCF_HDROP<\/code>,\nyou can improve the performance of drag\/drop operations by\nadding a little more information to your data object.\n<\/p>\n

\nObserve that the\nCF_HDROP<\/code>\nclipboard format is just a list of paths.\nSome drop targets care about whether the paths refer to directories\nor to files,\nand since CF_HDROP<\/code> does not provide this information,\nthe drop targets are forced to access the disk to get the answer.\n(This can be expensive for network locations.)\n<\/p>\n

\nTo help this case, you can add a\nCFSTR_FILE_ATTRIBUTES_ARRAY<\/code> to your\ndata object.\nThis contains the file attribute information for the items in your\nCF_HDROP<\/code>,\nthereby saving the drop target the cost of having to go find them.\n<\/p>\n

\nTake our\n\ntiny drag-drop sample<\/a>\nand make the following changes:\n<\/p>\n

\nclass CTinyDataObject : public IDataObject\n{\n  ...\n  enum {\n    \/\/ DATA_TEXT,<\/strike><\/font>\n    DATA_HDROP,<\/font>\n    DATA_NUM,\n    DATA_INVALID = -1,\n  };\n  ...\n};\nCTinyDataObject::CTinyDataObject() : m_cRef(1)\n{\n  SetFORMATETC(&m_rgfe[DATA_HDROP<\/font>], CF_HDROP<\/font>);\n}\nstruct STATICDROPFILES\n{\n DROPFILES df;\n TCHAR szFile[ARRAYSIZE(TEXT(\"C:\\\\Something.txt\\0\"))];\n} const c_hdrop = {\n  {\n    FIELD_OFFSET(STATICDROPFILES, szFile),\n    { 0, 0 },\n    FALSE,\n    sizeof(TCHAR) == sizeof(WCHAR), \/\/ fUnicode\n  },\n  TEXT(\"C:\\\\Something.txt\\0\"),\n};<\/font>\nHRESULT CTinyDataObject::GetData(FORMATETC *pfe, STGMEDIUM *pmed)\n{\n  ZeroMemory(pmed, sizeof(*pmed));\n  switch (GetDataIndex(pfe)) {\n  case DATA_HDROP<\/font>:\n    pmed->tymed = TYMED_HGLOBAL;\n    return CreateHGlobalFromBlob(&&c_hdrop, sizeof(c_hdrop),<\/font>\n                              GMEM_MOVEABLE, &pmed->hGlobal);\n  }\n  return DV_E_FORMATETC;\n}\n<\/pre>\n
\nOkay, let’s look at what we did here.\n<\/p>\n
\nFirst, we make our data object report a\nCF_HDROP<\/code>.\nWe then declare a static DROP­FILES<\/code> structure\nwhich we use for all of our drag-drop operations.\n(Of course, in real life, you would generate it dynamically,\nbut this is just a Little Program.)\n<\/p>\n
\nThat’s our basic program that drags a file.\n<\/p>\n
\nNote that\n<\/p>\n
\nyou are much better off letting the shell create the data object,\n<\/p><\/blockquote>\n
\nsince that data object will contain much richer information\n(and this entire article would not be needed).\nHere’s a sample program which\n\nuses the Get­UI­Object­Of­File<\/code> function\nto do this in just a few lines<\/a>.\nIt’s much shorter than having to cook up this\nCTiny­Data­Object<\/code> class.\nI’m doing it this way on the assumption that your program\nis deeply invested in the less flexible CF_HDROP<\/code>\nformat,\nso changing from CF_HDROP<\/code> to some other format\nwould be impractical.\n<\/p>\n
\nOkay, so that’s the program we’re starting from.\nLet’s add support for precomputed attributes.\n<\/p>\n
\nclass CTinyDataObject : public IDataObject\n{\n  ...\n  enum {\n    DATA_HDROP,\n    DATA_ATTRIBUTES,<\/font>\n    DATA_NUM,\n    DATA_INVALID = -1,\n  };\n  ...\n};\nCTinyDataObject::CTinyDataObject() : m_cRef(1)\n{\n  SetFORMATETC(&m_rgfe[DATA_HDROP], CF_HDROP);\n  SetFORMATETC(&m_rgfe[DATA_ATTRIBUTES],\n               RegisterClipboardFormat(CFSTR_FILE_ATTRIBUTES_ARRAY));<\/font>\n}\nFILE_ATTRIBUTES_ARRAY c_attr = {\n 1, \/\/ cItems\n FILE_ATTRIBUTE_ARCHIVE, \/\/ OR of attributes\n FILE_ATTRIBUTE_ARCHIVE, \/\/ AND of attributes\n { FILE_ATTRIBUTE_ARCHIVE }, \/\/ the file attributes\n};<\/font>\nHRESULT CTinyDataObject::GetData(FORMATETC *pfe, STGMEDIUM *pmed)\n{\n  ZeroMemory(pmed, sizeof(*pmed));\n  switch (GetDataIndex(pfe)) {\n  case DATA_HDROP:\n    pmed->tymed = TYMED_HGLOBAL;\n    return CreateHGlobalFromBlob(&amp;c_hdrop, sizeof(c_hdrop),\n                              GMEM_MOVEABLE, &pmed->hGlobal);\n  case DATA_ATTRIBUTES:\n    pmed->tymed = TYMED_HGLOBAL;\n    return CreateHGlobalFromBlob(&c_attr1, sizeof(c_attr1),\n                              GMEM_MOVEABLE, &pmed->hGlobal);<\/font>\n  }\n  return DV_E_FORMATETC;\n}\n<\/pre>\n
\nOkay, let’s look at what we did here.\n<\/p>\n
\nWe added a new data format,\nCFSTR_FILE_ATTRIBUTES_ARRAY<\/code>,\nand we created a static copy of the\nFILE_ATTRIBUTES_ARRAY<\/code>\nvariable-length structure that contains the attributes of our\none file.\nOf course, in a real program, you would generate the structure\ndynamically.\nNote that I use a sneaky trick here:\nSince the\nFILE_ATTRIBUTES_ARRAY<\/code>\n\nends with an array of length 1<\/a>,\nand I happen to need exactly one item,\nI can just declare the structure as-is and fill in the one slot.\n(If I had more than one item, then I would have needed more typing.)\n<\/p>\n
\nTo make things easier for the consumers of the\nFILE_ATTRIBUTES_ARRAY<\/code>,\nthe structure also asks you to report the logical OR and logical AND\nof all the file attributes.\nThis is to allow quick answers to questions like\n“Is everything in this CF_DROP<\/code> a file?”\nor\n“Is anything in this CF_DROP<\/code> write-protected?”\nSince we have only one file, the calculation of these OR and AND\nvalues is nearly trivial.\n<\/p>\n
\nOkay, so there isn’t much benefit to adding file attributes\nto a drag of a single file from the local hard drive,\nsince the local hard drive is pretty fast,\nand the file attributes may very well be cached.\nBut if you’ve placed thousands of files from a network drive\nonto the clipboard,\nthis shortcut can save a lot of time.\n(That was in fact the customer problem that inspired this Little Program.)<\/p>\n","protected":false},"excerpt":{"rendered":"
The CF_HDROP clipboard format is still quite popular, despite its limitation of being limited to files. You can’t use it to represent virtual content, for example. For all of you still using CF_HDROP, you can improve the performance of drag\/drop operations by adding a little more information to your data object. Observe that the CF_HDROP […]<\/p>\n","protected":false},"author":1069,"featured_media":111744,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1],"tags":[25],"class_list":["post-783","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-oldnewthing","tag-code"],"acf":[],"blog_post_summary":"
The CF_HDROP clipboard format is still quite popular, despite its limitation of being limited to files. You can’t use it to represent virtual content, for example. For all of you still using CF_HDROP, you can improve the performance of drag\/drop operations by adding a little more information to your data object. Observe that the CF_HDROP […]<\/p>\n","_links":{"self":[{"href":"https:\/\/devblogs.microsoft.com\/oldnewthing\/wp-json\/wp\/v2\/posts\/783","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devblogs.microsoft.com\/oldnewthing\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devblogs.microsoft.com\/oldnewthing\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/oldnewthing\/wp-json\/wp\/v2\/users\/1069"}],"replies":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/oldnewthing\/wp-json\/wp\/v2\/comments?post=783"}],"version-history":[{"count":0,"href":"https:\/\/devblogs.microsoft.com\/oldnewthing\/wp-json\/wp\/v2\/posts\/783\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/oldnewthing\/wp-json\/wp\/v2\/media\/111744"}],"wp:attachment":[{"href":"https:\/\/devblogs.microsoft.com\/oldnewthing\/wp-json\/wp\/v2\/media?parent=783"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/oldnewthing\/wp-json\/wp\/v2\/categories?post=783"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/oldnewthing\/wp-json\/wp\/v2\/tags?post=783"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}