Recent Commits to buku:master (6)

  • 31/10/2025 22:36

    support default scheme customization

    support default scheme customization
    
    This allows the user to specify a scheme that should be used to try to open a bookmark if the bookmark is missing a scheme. E.g. `www.example.com` will be changed to `http://www.example.com` or `https://www.example.com` automatically before an open/browse is attempted.
    
    Originally this patch simply forced opening with https, but it has since been modified to default to either http or https, at the user's preference. The default scheme will continue to be http, for now.
    
    Forcing https by default was a suggestion based mainly on my own use case. I have 3,678 bookmarks. 3,670 use https and 7 use http. Of those 7, at least a couple are dead links and a couple are http-only to prove a point.
    
    I keep "https only" on in my daily driver browser (Firefox) and my alternate browser (Chromium). I can still view http sites if I click through the warning.
    
    I did a few lazyweb searches and found a presentation from the 26th USENIX in 2017:
    
      https://www.usenix.org/conference/usenixsecurity17/technical-sessions/presentation/felt
    
    In the video on that page they cover how https is more often available, used, and even defaulted to for many websites, somewhere around 50-90% depending on how you measure it and where you get your data (and this is back in 2017).
    
    That presentation is about a paper by Felt et al titled "Measuring HTTPS Adoption on the Web":
    
      https://www.usenix.org/system/files/conference/usenixsecurity17/sec17-felt.pdf
    
    One of their data sources is MAWI, so I just peeked at MAWI's latest charts:
    
      https://mawi.wide.ad.jp/mawi/samplepoint-F/2025/202510241400.html
    
    I did some super simple math. I can't say if this is at all valid, but I just compared the first two numbers where I saw http and https and found http is 14.4% of the total and https is 85.6%. Looks like I was comparing ip.tcp.http packets vs. ip.tcp.https packets. So that's something there to look at, I guess.
    
    (83894509 ip https packets + 14072651 ip http packets ) = 97967160 total packets
    
    14072651 ÷ 97967160 = 0.144 = 14.4% http packets
    
    83894509 ÷ 97967160 = 0.856 = 85.6% https packets
    
    MAWI data snippet:
    
    ```
    Protocol Breakdown
    
         protocol		packets			bytes		bytes/pkt
    ------------------------------------------------------------------------
     total        334771299 (100.00%)     364013578524 (100.00%)   1087.35
     ip           288108851 ( 86.06%)     307918329080 ( 84.59%)   1068.76
      tcp         125995157 ( 37.64%)     190503802981 ( 52.33%)   1511.99
       http        14072651 (  4.20%)      24776980541 (  6.81%)   1760.65
       https       83894509 ( 25.06%)     150210613012 ( 41.27%)   1790.47
    ...✂️...
    ```
    
    Anyway, sorry for all this noise, it's really just breadcrumbs in case someone wants to do actual research.
    
    Also I'm really enjoying buku so far. Thank you for sharing it!
    
    Unit tests pass locally.
    
    No part of this commit log message was AI-generated.
    
    ---
    
    add --default-scheme to manpage
    
    ---
    
    pass default_scheme into browse() as a keyword arg
  • 26/10/2025 13:14

    [jarun#753] migrate to Bootstrap v4

    [jarun#753] migrate to Bootstrap v4
  • 26/10/2025 10:29

    Merge pull request #867 from LeXofLeviafan/improve-bukucrypt

    Merge pull request #867 from LeXofLeviafan/improve-bukucrypt
    
    [API] BukuCrypt improvements
  • 26/10/2025 10:28

    Merge pull request #870 from LeXofLeviafan/python-3.14

    Merge pull request #870 from LeXofLeviafan/python-3.14
    
    Updating supported Python versions
  • 19/10/2025 13:52

    added Python 3.14 + dropped Python 3.9

    added Python 3.14 + dropped Python 3.9
  • 03/10/2025 20:52

    [API] improved BukuCrypt interface

    [API] improved BukuCrypt interface

Log in