Hello, I’m using D2007, DA 6.0.47.841 and TROIPHTTPServer
I’ve a lot of AV in TIPHTTPRequestHeaders.Clear in my servers, this is the madExcept CallStack
operating system : Windows 2003 R2 Service Pack 2 build 3790
system up time : 137 days 14 hours
program up time : 2 hours 32 minutes
processors : 4x Intel(R) Xeon(R) CPU X3210 @ 2.13GHz
physical memory : 2629/4094 MB (free/total)
free disk space : (C:) 76,42 GB
display mode : 1024x768, 32 bit
process id : $1524
allocated memory : 221,72 MB
largest free block : 638,93 MB
exec. date/time : 2017-01-17 10:34
compiled with : Delphi 2006/07
madExcept version : 4.0.16
callstack crc : $ea9a618b, $2a90808c, $0fa9fe4b
count : 2
exception number : 2
exception class : EAccessViolation
exception message : Access violation at address 007B9054 in module ‘server.exe’. Read of address 00000045.
we had similar issue and I think it was fixed a long time ago
if you are compiling your project with latest RO version, this error will be reproduced?
another workaround: replace TROIPHTTPServer with Indy-based HTTP server
hmm, weird.
looks like your code isn’t thread-safe.
can you describe your application in several words? does it use background threads, events?
if it use events, it should have personal channel+message for usual calls and for events.
can you create a simple testcase that reproduces this problem?
client-side should use one channel+message for usual calls and one channel+message for events.
note: this is valid for usual (i.e. non-super) channels. one instance of super channel can be used for all requests (usual calls and events)
“Assign its Message property to the message you use to communicate with the server (note that since events are dispatched based on the ClientID, this must be the same Message component you also use for your regular communication)”
it could be a problem in SDK or a problem with your code.
you have reproduced this with both different servers (synapse- and indy-based).
we have never had such issues for plain indy http server so it could be something with your code if you do something with server-side headers.
by other way, it could be some corner case, but I need to testcase for reproducing this.
also I recommend to retest this issue with the latest [trial] RO version. it has a chance, that everything will work as expected
I tried to retest this issue with latest (trial) RO version but it’s impossible without legacy components. uROEncryption, IDADataSets.SQL, etc, etc. It would take me days to change all my program code for a test
The setup of last trial installation don’t use the folder especified in the installation, always use the default folder. This causes errors with my old version installation.