Opened 17 years ago
Last modified 15 months ago
#1062 in-progress enhancement
add copyright headers to googlefs; cleanup code.
| Reported by: | mmu_man | Owned by: | mmu_man |
|---|---|---|---|
| Priority: | low | Milestone: | Unscheduled |
| Component: | File Systems/GoogleFS | Version: | R1/pre-alpha1 |
| Keywords: | Cc: | ||
| Blocked By: | Blocking: | ||
| Platform: | All |
Description
The code is messy; just a personal reminder. Don't even look at it :p
Change History (10)
comment:1 by , 17 years ago
| Status: | new → assigned |
|---|
comment:2 by , 17 years ago
comment:3 by , 17 years ago
Can always rename it... I don't think they fear our competition though :D
comment:4 by , 17 years ago
I wasn't talking about the naming issue. The problem is that we are using their search results and changing the way they're displayed (we even remove the ads). AFAIK, they don't allow manipulating the resulting HTML. The search results must look exactly the same as in a web browser. At least, we discussed this on the admin list.
comment:5 by , 17 years ago
the ad removal is just an unattended result of the simplistic html parser (it also skips non-html pages like PDF). I started looking at their XML API but it was restricted and they removed it anyway... so much for the altruistic period :) They probably can't do anything against the code itself (free speech) just for shipping bins... I don't think they ever sued over Firefox search side-bars which do parse the html, but I'll remove it if deemed necessary.
comment:6 by , 11 years ago
It's been 6 years, have you finished this yet? ;) is it even still needed/useful?
comment:7 by , 11 years ago
Well yeah, I probably have a patch for that lying in the svn working copy I moved away when switching to git...
comment:8 by , 9 years ago
| Milestone: | R1 → Unscheduled |
|---|
comment:9 by , 3 years ago
Current status:
- Smap violations (I fixed some)
- The current Google result page is way too complicated to parse
- I tired switching to html.duckduckgo.com which is a lot simpler to parse, but requires https, and I don't think we want to implement that in the kernel?
So, if this has any future, it should be moved to userlandfs.
And the "legal" issue still stands, it's not clear if we're allowed to use the web search results in this way. Some info about this here: https://duckduckgo.com/api which explains why DDG can't provide an API for search results (basically it's because they outsource the search to other companies).
comment:10 by , 15 months ago
There is this project that deploys Google File System partially in C++, this one under apache license: https://github.com/Gan-Tu/cppGFS2.0
The closest thing to the GFS is this project, also under apache license: https://github.com/quantcast/qfs
I believe this project under MIT license is more suitable for Haiku: http://ori.scs.stanford.edu/
I'm just not sure if we want to keep GoogleFS in our repository because of the legal issues it might have.