Is anyone still working on this module? The user count and last updated date arent very promising.

It seems the esi-module is finally doing some progress, but nothing visible for D7. So, should I expect this module to do something in the upcoming future :)?

Comments

pounard’s picture

pounard
Primary language French
commented

Oh I could get back to work on this one, I just have no time. What I'm missing is probably good ideas of features, would like to help?

bibo’s picture

bibo commented
Status: Active » Closed (works as designed)

@pounard

Sorry for being slow. I do have some ideas, but it seems that the esi-module has just recently gained some greater momentum with people like mikeytown2. I do appreciate this module of course, but it might be better to concentrate efforts there and wait for the D7-version, or what do you think?

Anyway, a few features I need (and may implement in my own custom module): better communication between varnish and the webserver. For example, varnish can tell the server with a custom header that it's "there" and ready to do esi. And the webserver may respond with a header like "X-ESI-Process: 1" that tells varnish to do esi processing. Also, some pages may need to be cached per user (or rather, per session), be it esi or not. I know some good vcl snippets for those. Also page cache tagging, but that's not really about esi.

pounard’s picture

pounard
Primary language French
commented

For Varnish headers and communication, I'm interested! Any information, data, explaination or just pointing me to the right documentation is welcome. My biggest problem is time now.

About the ESI module maintainers, I already proposed to merge effort, I exchanged some mails with one of the maintainers, but they never gave me (at least read) access of any kind to their development code at the time: their responsability, their loss. Actually when I read some descriptions of their module I see some of my ideas getting in there. I guess they don't want to share neither code, neither credits. Plus, their git repository still do not include any D7 code. They chose to fully develop it hidden from public, so any argument about architecture or features is already dead even before we have a chance to see the code: I don't want to be the slave of their ideas and internal development workflow. They came to me, I gave them a chance to merge, they rejected their own offer; End of story.

I'm still eager to maintain this module, and I will do it, but I'm afraid that right now, this would happen only if I need it for business reasons: real personal involvment will have to wait a few weeks! If it happens that some relatively reasonable sized features that can be done come to my mind, I may do it before that.