• Get environment variable pointing to the XDG Runtime folder
• Loop through the gvfs folder to find/add shares

I´m not perfectly sure if all folders have the same name convention, but unaltered they don't make good names to add to the list as is.
I've made the assumption that every entry has a "share=xxx" at the end for now.

NB.
Put the variable declarations near the loop.. just for quick reference
My additions to the code are mostly "kitbashing"
I have no real insight in the blender code and available tools (yet)
Not a C++ coder :-(
(en de leeftijd speelt ook mee)

Removed one check for existance of /run/user/xxx/gvfs folder
If it doesn't exist no real problem arises.
Benefit however is that when the network is down, it won't have to wait twice for a dir-exists and dir-contents

All in all I find it very difficult to get a good solution.. I'm starting to think that just adding the "gvfs" folder with another label name like "Gnome Shares" or something like that solves a lot of problems.
but then again, inside it shows those stupid folder names like in my case:
'smb-share:server=imac2009-capitan.local,share=data_1/'
'smb-share:server=imac2009-capitan.local,share=data_2/'
'smb-share:server=imac2009-capitan.local,share=datapartition/'

On the other hand, now it looks nice and just lists:
data_1
data_2
datapartition
but with the risk that starting blender and opening the file browser will take a lot of time IF the user decides to disconnect the network while the shares are still mounted.

NB. I did find another mountpoint that should not be in the system folder list: /boot/efi

forgot to include the #include "BLI_fileops_types.h" this last diff.

And while looking over the code for the nth time.. I know the addition of 6 to the label pointer is dangerous and at least foolish.. but like it said in the code.. I'm still assuming on every platform the directory entries look the same.. if not.. than my filtering out the share name is obsolete anyhow ; -)

As noted by Campbell, using STRPREFIX makes more sense.
Made the filtering a bit smarter/error proof.

Do you think the timeout will be a dealbreaker for implementing when the user has active gvfs shares AND a disconnected network?
Or does this fall under "expected and nothing we can do about"

Rewind accidental changes mixing up with rB4e9911663a22766f6751d33ae84c1ebcc1fae559

Rebase on 2.8, applied minor cleanup to master/2.8

Added BLI_filelist_free. (should have been in there from the start.. seems I forgot to copy/paste that line before, and never noticed it)
It solves the leak.

To be sure I've built a debug version with ASAN.. and I'm getting no leaks anymore.
Also confirmed by Dalai (IRC)

Got to say.. for a newb it was not easy to get ASAN up and running. In the end it's simple.. like everything. learned a lot

Checking for NULL on XDG_RUNTIME_DIR
As if there are people not using systemd ;-)

I took the liberty not to change (foolproof) more of the code (check existance of the gvfs folder inside the runtime_dir..
But if that would be required, I'll surely hear from you again ;-)
Not really sure how deep/foolproof the checking should be.

Thanks,
Roel