I did one more commit to this module with 2 additions

- A stub for a ::rivet::exit_thread function has been added. For a non threaded MPM this command is supposed to behave like the core 'exit' command (thus terminating a whole process) whereas with a threaded MPM such as 'worker' it should cause a single thread to exit. For this purpose the exception mechanism used for ::rivet::abort_page was extended.

- Completed handling of RivetUserConf scripts. These scripts are not cached though and we should probably devise a policy for them. I'm facing 2 problems with them:

° I could not figure out a way to cache consistently these scripts by identifying rivet_server_conf with data coming from RivetUserConf directives and using it a key to an hash table. RivetDirConf allowed me to do so using a 'path' field in the cmd_parms structure

° If an .htaccess file applies Rivet_UserConf gets called on every request (enough for deprecate the habit of using them) and there is no obvious way to understand if the file was actually updated except maybe for explicitly read the atime of the file and store it somewhere.

 -- Massimo

On 09/29/2014 03:08 PM, Harald Oehlmann wrote:
Thanks !
Harald

Am 29.09.2014 um 15:03 schrieb Massimo Manghi:
Yes, in principle we have everything we need to provide support for
Windows (what is the most common MPM for windows? winnt?) The worker
bridge might be very close to work with it as it doesn't assume anything
it needs was inherited through a fork. As long as APR is really
'portable' and predictable in its behavior across OS it should work.
Mostly at least.

I recently had to wipe out my windows 7 installation. Even though I
seldom fired it I found out it had been attacked by multiple malware and
other stuff. I reinstalled it in a 32-bit virtualbox. I need a jumpstart
to install a minimal development environment where I can build Tcl,
apache and mod_rivet....

On 29-09-2014 12:53, Harald Oehlmann wrote:
Congratulations, Massimo, sounds great.
That also means that the way to a windows port is getting closer ?

Thank you,
Harald



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]




---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to