Introduction
Internet Explorer classic (IE in the following) was based on ActiveX technology. It was very easy to automate IE for tasks like Webscraping or testing from OLE-aware programming languages like VBA. But Microsoft will end support for IE in the near future and wants users to move to newer browsers like Microsoft Edge.
Microsoft Edge is no longer based on ActiveX technology. Microsoft seems uninterested in creating a drop-in replacement for the IE OLE Object. There are libraries that try to fill this gap using Selenium, see
Seleniumbasic
as an example. But this requires the installation of a Webdriver, which might not be feasible in some environments. The following solution needs no additional software, apart from a Chrome-based browser.
Keep in mind, that all running Edge procceses must be terminated before running the code. Otherwise the tabs are opened in the currently running process, not the one that has been started and subsequent communication between VBA and Edge fails.
CDP Protocol
The code uses the Chrome Devtools Protocol (CDP) to communicate with the browser. A full documentation of the protocol can be found
here
. The code implements only a very narrow set of functions:
Basic functions to set up the communication channel
Navigation to a url
Evaluate arbitrary JavaScript expressions in the context of a page and return the result
But these functions should suffice to do basic Webscraping. The main code is as follows:
Sub
runedge()
Dim
objBrowser
As
clsEdge
Set
objBrowser =
New
clsEdge
Call
objBrowser.start
Call
objBrowser.attach(
"
"
)
Call
objBrowser.navigate(
"
https://google.de"
)
Call
objBrowser.waitCompletion
Call
objBrowser.jsEval(
"
alert(""hi"")"
)
Call
objBrowser.jsEval(
"
document.getElementsByName(""q"")[0].value=""automate edge vba"""
)
Call
objBrowser.jsEval(
"
document.getElementsByName(""q"")[0].form.submit()"
)
Call
objBrowser.waitCompletion
Call
objBrowser.jsEval(
"
document.evaluate("".//h3[text()='Automate Chrome / Edge using VBA - CodeProject']"", document).iterateNext().click()"
)
Call
objBrowser.waitCompletion
Dim
strVotes
As
String
strVotes = objBrowser.jsEval(
"
ctl00_RateArticle_VountCountHist.innerText"
)
MsgBox (
"
finish! Vote count is "
& strVotes)
objBrowser.closeBrowser
End
Sub
The class
clsEdge
implements the CDP protocol. The CDP protocol is a message-based protocol. Messages are encoded as JSON. To generate and parse JSON, the code uses the VBA-JSON library from
here
.
Low-Level Communication with Pipes
The low-level access to the CDP protocol is avaible by two means: Either Edge starts a small Webserver on a specific port or via pipes. The Webserver lacks any security features. Any user on the computer has access to the webserver. This may pose no risks on single user computers or dedicated virtual containers. But if the process is run on a terminal server with more than one user, this is not acceptable. That's why the code uses pipes to communicate with Edge.
Edge uses the third file descriptor (fd) for reading messages and the fourth fd for writing messages. Passing fds from a parent process to child process is common under Unix, but not under Windows. The WinApi call to create a child process (
CreateProcess
) allows to setup pipes for the three common fds (
stdin
,
stdout
,
stderr
) using the
STARTUPINFO
structure, see
CreateProcessA function (processthreadsapi.h)
and
STARTUPINFOA structure (processthreadsapi.h)
. Other fds cannot be passed to the child process.
In order to set up the fourth and fifth fds, one must use an undocumented feature of the Microsoft Visual C Runtime (
MSVCRT
): If an application is compiled with Microsoft C, than one can pass the pipes using the
lpReserved2
parameter of the
STARTUPINFO
structure. See "
Undocumented CreateProcess
" for more details (scroll down the page).
The structure that can be passed in
lpReserved2
is defined in the module
modExec
.
Public
Type STDIO_BUFFER
number_of_fds
As
Long
crt_flags(
0
To
4
)
As
Byte
os_handle(
0
To
4
)
As
LongPtr
End
Type
The structure is defined to pass five fds in the
os_handle
array. The values for the
crt_flags
array can be obtained from
https://github.com/libuv/libuv/blob/v1.x/src/win/process-stdio.c
. The fields of the
struct
must lie contiguously in memory (packed). VBA aligns
struct
fields to 4 byte boundaries (on 32-bit systems). That's why a second
struct
with raw types is defined.
Public
Type STDIO_BUFFER2
number_of_fds
As
Long
raw_bytes(
0
To
24
)
As
Byte
End
Type
After populating the
STDIO_BUFFER struct
, the content is copied using
MoveMemory
to the
STDIO_BUFFER2
struct. The size of 25 bytes is enought to hold
crt_flags
(5 bytes) and the pointers (20 bytes).
History
8
th
July, 2021: Initial version
18
th
August 2021, added support for 64bit Office
3
rd
November 2021, some minor improvements
'
http://ww3.cad.de/foren/ubb/Forum226/HTML/000794.shtml
Private Declare Function CreateToolhelpSnapshot Lib "kernel32" Alias "CreateToolhelp32Snapshot" ( _
ByVal lFlgas As Long, _
ByVal lProcessID As Long) As Long
Private Declare Function TerminateProcess Lib "kernel32" ( _
ByVal hProcess As Long, _
ByVal uExitCode As Long) As Long
Private Declare Function OpenProcess Lib "kernel32" ( _
ByVal dwDesiredAccess As Long, _
ByVal bInheritHandle As Long, _
ByVal dwProcessId As Long) As Long
Private Declare Function ProcessFirst Lib "kernel32" Alias "Process32First" ( _
ByVal hSnapshot As Long, _
ByRef uProcess As PROCESSENTRY32) As Long
Private Declare Function ProcessNext Lib "kernel32" Alias "Process32Next" ( _
ByVal hSnapshot As Long, _
ByRef uProcess As PROCESSENTRY32) As Long
Private Declare Function CloseHandle Lib "kernel32" ( _
ByVal hObject As Long) As Long
Private Type PROCESSENTRY32
dwSize As Long
cntUsage As Long
th32ProcessID As Long
th32DefaultHeapID As Long
th32ModuleID As Long
cntThreads As Long
th32ParentProcessID As Long
pcPriClassBase As Long
dwflags As Long
szexeFile As String * 260
End Type
Private Const PROCESS_TERMINATE = &H1
Private Const GW_HWNDNEXT As Long = 2
Public Sub KillEXE(ByVal strEXEName As String)
Dim lngHandle As Long, retVal As Long, hTask As Long, lResult As Long
Dim pe32current As PROCESSENTRY32
lngHandle = CreateToolhelpSnapshot(2&, 0&)
If lngHandle <> 0 Then
pe32current.dwSize = Len(pe32current)
retVal = ProcessFirst(lngHandle, pe32current)
Do While Not (retVal = 0)
If InStr(1, pe32current.szexeFile, strEXEName, vbTextCompare) > 0 Then
hTask = OpenProcess(PROCESS_TERMINATE, 0&, pe32current.th32ProcessID)
lResult = TerminateProcess(hTask, 1&)
lResult = CloseHandle(hTask)
Exit Sub
End If
retVal = ProcessNext(lngHandle, pe32current)
CloseHandle lngHandle
End If
End Sub
Public Function IsEXErunning(ByVal strEXEName As String)
Dim lngHandle As Long, retVal As Long, hTask As Long, lResult As Long
Dim pe32current As PROCESSENTRY32
lngHandle = CreateToolhelpSnapshot(2&, 0&)
If lngHandle <> 0 Then
pe32current.dwSize = Len(pe32current)
retVal = ProcessFirst(lngHandle, pe32current)
Do While Not (retVal = 0)
If InStr(1, pe32current.szexeFile, strEXEName, vbTextCompare) > 0 Then
hTask = OpenProcess(PROCESS_TERMINATE, 0&, pe32current.th32ProcessID)
'
lResult = TerminateProcess(hTask, 1&)
lResult = CloseHandle(hTask)
IsEXErunning = True
Exit Function
End If
retVal = ProcessNext(lngHandle, pe32current)
CloseHandle lngHandle
End If
End Function
Public Sub test()
Dim Response As VbMsgBoxResult
If IsEXErunning(
"
msedge.exe"
) Then
Response = MsgBox(
"
Achtung, MS EDGE läuft bereits."
& vbCrLf & vbCrLf &
"
Soll der Browser jetzt geschlossen werden?"
, vbYesNo)
If Response = vbNo Then
Exit Sub
Call KillEXE(
"
msedge.exe"
)
End If
End If
End Sub
2) So before the actual call: "Set objBrowser = New clsEdge" we make another query if EDGE is already running:
Dim Response As VbMsgBoxResult
If IsEXErunning(
"
msedge.exe"
) Then
Response = MsgBox(
"
Achtung, MS EDGE läuft bereits."
& vbCrLf & vbCrLf &
"
Soll der Browser jetzt geschlossen werden?"
, vbYesNo)
If Response = vbNo Then
Exit Function
Call KillEXE(
"
msedge.exe"
)
End If
End If
'
Start Browser
Set objBrowser = New clsEdge
Sign In
·
View Thread
Would you please teach me how to use these codes? I still have "Error PeekNamedPipe in readProcCDP Pin".
Sign In
·
View Thread
I still have this:"Error PeekNamedPipe in readProcCDP"
I have used your code but nothing changes.
Sign In
·
View Thread
I included the files and pasted the sample code. I also added the reference to the Scripting for Dictionary. When I try to run the sample code I get an error.
Edge is opened without problem but than the execution of
Quote:
Call objBrowser.attach("")
leads to the error
Quote:
Runetime: 901; Error PeekNamedPipe in readProcCDP
Sign In
·
View Thread
For me a similiar approach worked. Before opening Edge I kill all running instances with this little function (this takes some seconds but I never had that error again):
Function Kill_Edges()
Dim myWmi As
Object
, getProc As Variant
Dim myProcArr As Variant, tarProc As Variant
Set myWmi = GetObject(
"
winmgmts:"
)
getProc =
"
select * from win32_process where name='msedge.exe'"
Set myProcArr = myWmi.ExecQuery(getProc)
On Error Resume Next
For Each tarProc In myProcArr
tarProc.Terminate
0
End Function
Sign In
·
View Thread
is it possible to attach to edge that already opened without run objBrowser.start() ?
modified 3-Mar-22 3:07am.
Sign In
·
View Thread
Just want update base on my experience so far. I think there is no way to attach edge already oppened.
I also have another issue on automate edge by this code or selenium. The issue is when we found website that blocked automate then probably we will found "Bad response status 400" on some action POST.
Appreciate for anyone that can share the problem solving for above issue.
Sign In
·
View Thread
Hello, Chris K23
"- enable logging" in strcall string seems to be used for log. After deleted, console windows disappears.
Will it affect the function of the program? Hope to get your reply,Thanks.
modified 14-Feb-22 19:49pm.
Sign In
·
View Thread
I removed it also and everything works as expected.
By the way, does anybody see something in that black window? If this is the log window, I never see any text showing up...
Sign In
·
View Thread
to hide that window you can set as below on clsEdge
strCall = """C:\Program Files (x86)\Microsoft\Edge\Application\msedge.exe"" --remote-debugging-pipe --disable-automation --disable-logging"
Sign In
·
View Thread
Your code works perfectly thank you so much..
Iam very new to this so sorry if this is basic question
Where can i find the exact syntax for other commands (like finding element by id, class or other parameters) .
because i tried something like
Call objBrowser.jsEval("document.getElementsByClassName(""xxxx"")[0].form.submit()")
but it is not working. Please help
Sign In
·
View Thread
The easiest way to try out commands is the JavaScript console in the developer tools. I think it even has syntax completion.
Sign In
·
View Thread
If objRes.Item(
"
result"
).Item(
"
result"
).Item(
"
type"
) =
"
string"
Or objRes.Item(
"
result"
).Item(
"
result"
).Item(
"
type"
) =
"
boolean"
Or objRes.Item(
"
result"
).Item(
"
result"
).Item(
"
type"
) =
"
number"
Then
jsEval = objRes.Item(
"
result"
).Item(
"
result"
).Item(
"
value"
)
End If
But if he wants to use "getElementsByTagName", this will give him a collection of objects. Is there a way for the jsEval function to return a collection of objects from the CDP message?
Sign In
·
View Thread
With the current code:no. If you need to work with non-elementary objects like DOM elements you could code your logic in JavaScript as a workaround. The JavaScript code could then be evaluated in the browser using jseval and return the plain resulting object back to VBA.
Sign In
·
View Thread
Or you could do it the long way around as I do.
I use jsEval to get the innerHTML of the main document and then set it
to MSHTML.HTMLDoc (in Excel VBA).
I can then use the normal getElementById, getElementsByName etc...
This is in Excel (rough example)
Dim webDoc As MSHTML.HTMLDocument
'
Later in the code when I need it
Set webDoc = objBrowser.jsEval("document.getElementById('
myid
'
).innerHTML;")
'
at
this
point you should be able to use the the following
webDoc.getElementById
webDoc.getElementsByName
I have gotten Option List Elements, pretty much anything I need.
Sign In
·
View Thread
IF you opened Edge with the provided code (through CDP protocol) in the first place, then yes!
It's the serialize/deserialize functions that will do the trick. You can read previous posts to learn more about it.
Sign In
·
View Thread
