[tor-bugs] #13339 [Tor]: Merge GSoC project - Consensus Diffs

[Tor Bug Tracker & Wiki]

#13339: Merge GSoC project - Consensus Diffs
-----------------------------+-------------------------------------------
     Reporter:  mvdan        |      Owner:
         Type:  enhancement  |     Status:  needs_revision
     Priority:  major        |  Milestone:  Tor: 0.2.6.x-final
    Component:  Tor          |    Version:
   Resolution:               |   Keywords:  gsoc merge tor-client prop140
Actual Points:               |  Parent ID:
       Points:               |
-----------------------------+-------------------------------------------

Comment (by nickm):

 Replying to [comment:12 mvdan]:
 > Some more comments, for nickm this time :)
 >
 > * So in tor_rmdir, you say that we should check against S_IFLINK (that
 it is false) when checking for S_IFDIR. I thought that stat ran on the
 symlink, not on its target? In other words, we couldn't have both be true
 at the same time. If it's a symlink and it's not a folder, we delete it.
 In the opposite case, we enter the dir since we're not following a
 symlink.

 Maybe we should lstat it then?  I worry that this could be used along with
 a symlink to rm_rf something we didn't intend to.  Probably nothing to
 worry about given this use case, but a bit alarming.

 > * smartlist_slice_t in consdiff.h instead of consdiff.c - this is mainly
 because back when I added it I wasn't sure whether I'd need it in other
 places. But also because I was thinking that the structure and its
 functions might be useful to other parts of Tor. After all, slicing a list
 is a very generic tool, and it sounds simple and cheap to me. Of course it
 would need polishing to be made public, so I think I'll just move it to
 consdiff.c for now.

 Okay, unless you think that exposing it for testing could help test
 individual functions.

 > Regarding what both you and dgoulet have said about the add_bytes
 function - GSoC was nearing its end and I was getting a bit confused by
 all the "serve bytes/files" code in dirserv.c, so what I did was take a
 function that kind of did what I wanted and tweaked it for my purposes.
 This has two obvious problems, the first that it duplicates code, but the
 second that I'm not exactly sure whether what I did is correct or not.
 >
 > In other words, I would appreciate it if you could help me here. I
 remember nickm mentioning that this code has barely been touched in years,
 and that it's a tad confusing since it's gotten so complex. So I'd rather
 not do it alone :)

 Okay; I'll try to have a look.

 > Oh, and I'll definitely be attending the winter dev meeting in Valencia
 in March. Will any of you guys be there?

 I sure will!

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/13339#comment:14>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online