Jump to content

GrandAdmiralThrawn

Member
  • Posts

    28
  • Joined

  • Last visited

  • Donations

    0.00 USD 
  • Country

    Austria

Posts posted by GrandAdmiralThrawn

  1. On 9/19/2021 at 1:44 PM, UCyborg said:

    Moon Tester Tool can be used by a monkey to install an extension if its only issue is that it doesn't target Pale Moon. In either case, all extensions come in a form of .xpi file (basically ZIP archive) which contains text install.rdf file, containing basic information about the extension such as GUID, name, version, which applications it targets etc. The Moon Tester Tool basically rewrites the file so the extension targets Pale Moon.

    Information about the max version of the target application is ignored unless strict compatibility checking directive is present/enabled. Some extensions contain the directive that the XPI content should be extracted upon installation, for these extensions you should find the original XPI file.

    AFAIK, New Moon still accepts extensions targeting Firefox, so if some extension doesn't work, parts of its logic may have to be rewritten to accommodate for changes in the browser core.

    I supposed you'd have to find the old documentation for XUL coding and info regarding whatever MCP folks changed that broke the extension to be able to fix, the developer tools baked in the browser may also give hints.

    I modified Destroy The Web some time ago to remove dependency on deprecated FUEL, which is long gone from Basilisk and will eventually disappear from Pale Moon as well.

    Alright, thank you very much! I've installed the Moon Tester Tool extension and will give it a shot with a newer version of New Moon after having worked through the remaining npapi exensions. Thing is, more and more sites are starting to break on my 29.1.1 version (actually MyPal at the moment), so I'm hoping that upgrading to a newer version of either MyPal or New Moon will do some good. :)

    Your help is appreciated!

  2. I have a quick question to which I haven't found a clear solution browsing through the first and last 10 pages of this thread. If it's possible, is there an easy, generic step-by-step guide on how to make old FireFox extensions work with New Moon (and/or Pale Moon) after version 29.1.x? I tried to replace as many extensions as I could with native Pale Moon ones, but there are still some left I can't do without. ;)

    There's some mention of an "install.rdf" trick, but I'm not sure how to pull it off. Not all my FF extensions have such a file in their subfolder within my profile folder. Before I mess something up (I'll make backups, but still) I thought I'd ask here first.

    Thank you!

  3. On 4/16/2021 at 7:34 AM, HHo2016 said:

    In case anyone wants to download all the updates in the CSV file...

    I created a Python download script to bulk download those updates - tested on Python 3.9.2 but should also work for older Python 3 versions after installing the "requests" package ('pip install requests' shall suffice): [...]

    Thank you very much! However, the code did not work on my system using Python 3.4. It appears there are some syntax errors (at least when running it with 3.4) plus it tries to access the download subfolder without creating it first, which fails upon file write. I've adapted your code, fixing all the breaking stuff, adding the folder creation and a few more human-readable strings for what I believe are more common HTTP return codes. One new required Python module: "os".

    It's not very efficient I guess, trying to create the subfolder on every iteration, but hey, I am no Python programmer at all. But this did work for me (Python 3.4 on some old CentOS 6.10 Linux):

    import csv
    import os
    import requests
    
    # Add/replace the filenames in csv_files below with the CSV files you want to download
    csv_files = ['NT_5.1.2600-x86-All.csv', 'NT_5.1.2600-x86-IE-All.csv', 'NT_5.2.3790-x64-All.csv', 'NT_5.2.3790-x64-IE.csv', 'NT_5.2.3790-x86-All.csv', 'NT_5.2.3790-x86-IE.csv']
    
    for csvfile in csv_files:
        print('Starting file download for' , csvfile)
        with open(csvfile) as csv_file:
            csv_reader = csv.reader(csv_file, delimiter=',')  # Read comma-delimited CSV
            line_count = 0
            for row in csv_reader:  # Iterate over rows
                if line_count == 0:   # Output CSV header
                    print('Column names are:' , ' '.join(row))
                    line_count += 1
                else:
                    # Create download folder, but do nothing if it exists already
                    try:
                        os.mkdir('./' + csvfile[:-4] + '_DL/')
                        print('Subfolder' , csvfile[:-4] , '_DL/' , 'created')
                    except FileExistsError:
                        pass
                    print('\tDownloading' , row[0] , '...')
                    url = row[1]
                    r = requests.get(url)  # Get WU file
                    with open('./' + csvfile[:-4] + '_DL/' + row[0], 'wb') as f:
                        f.write(r.content)  # Write WU file to disk
                    # Let's define some human-redable HTTP status codes to output
                    if r.status_code == 200:
                        status = ', OK'
                    elif r.status_code == 403:
                        status = ', FORBIDDEN'
                    elif r.status_code == 404:
                        status = ', NOT FOUND'
                    elif r.status_code == 500:
                        status = ', INTERNAL SERVER ERROR'
                    elif r.status_code == 502:
                        status = ', BAD GATEWAY'
                    elif r.status_code == 504:
                        status = ', GATEWAY TIMEOUT'
                    else:
                        status = ''
                    print('\tCompleted with status code' , r.status_code , status)
                    # Write error log, should an error have occurred
                    if r.status_code != 200:
                        with open('./error.log', 'wb') as errLog:
                            errLog.write(('FAIL: ' + row[0] + ', ERROR CODE: ' + str(r.status_code) + status + '\nFAILED URL: ' + row[1] + '\n\n').encode())
                    line_count += 1
            print('Processed' , line_count , 'lines.')
    print('Download Completed.')

    I haven't introduced any additional, Linux/POSIX-specific stuff, so this should work on Windows just as well (or so I hope).

    Edit: Further extended the code. It now writes an error log to file error.log including file name and URL should any errors occur during the downloading process. That way you won't miss any download errors.

  4. On 5/30/2020 at 4:23 PM, Montana Slim said:

    "I just wanted to report, that somewhere in the last 48 hours, Reddit appears to have partially broken on all XP browsers, including New Moon, MayPal and Chromium ports. It seems several smaller functions don't work anymore, but the most prominent thing is that new posts don't get dynamically shown when scrolling down anymore. The amount of posts the user can see seems to depend on the viewport, or rather on the vertical resolution."

    Did you try switching the url to old.reddit.com ?

    For instance, instead of https://www.reddit.com/r/windowsxp ....use: https://old.reddit.com/r/windowsxp

    I tried it just now, and it's pretty ugly (well, in my opinion), but does work! So that's a good workaround for now, thanks! :)

    @VistaLover Thanks for your detailed report. I'm rather concerned about that Chinese browser. I've tested it before in a Russian re-release, supposedly with "less spyware", but I don't have it on my system anymore, so I haven't tested it. At least there is an option for those who are ok with it.

    PS.: I dislike almost all modern "responsive" designs (as they don't properly "respond" to me having a 100dpi desktop monitor, and are wasting my screen real estate), and that includes infinite scrolling. I'm not a Reddit user myself, but a person who used to comment on my blog and now sits on my IRC server reported this to me, so I thought it'd be a good idea to report it here. :)

  5. I just wanted to report, that somewhere in the last 48 hours, Reddit appears to have partially broken on all XP browsers, including New Moon, MayPal and Chromium ports. It seems several smaller functions don't work anymore, but the most prominent thing is that new posts don't get dynamically shown when scrolling down anymore. The amount of posts the user can see seems to depend on the viewport, or rather on the vertical resolution. Let me show you a screenshot:

    reddit-issue.thumb.png.2c068ba96417f8edc8b047db1fa855a3.png

    The two grey fields at the bottom should automatically turn into posts, and more posts should be loaded so I can scroll further down. But this doesn't work anymore, so the user is stuck with this. The smaller the screen resolution, the less content they can actually see.

    In this case, I used MyPal, but New Moon behaves in the exact same way.

    Hope this can be fixed!

    Thank you!

  6. Thank you!

    Also, it'd be interesting to hear of any leaks of updates sourced from paid support (which are not available to the public). There should be some companies which still get them after all. Would violate the license agreement to release them I presume, but still.

    I would've tried to actually pay Microsoft for them myself, but unfortunately, this service is not available for private individuals. ;)

  7. 4 hours ago, Usher said:

    AFAIK, you should search for Windows Server 2003 x64 updates.

    I'm sorry, but just 2 posts before I said that I had tracked Server 2003 x64 updates since the EoL of XP x64. You can find my corresponding list [here]. Also, as said, maybe I've just overlooked something. But my searches don't turn up any TLS 1.2 update for Windows SSL on Server 2003 x64 or XP x64.

    Of course, I can use it with New Moon, but that's not what this is about. I want this for IE8 to upgrade certain softwares which use the IE8 engine to access secure websites (some license activation systems do this, e.g. the one that comes with Adobe Photoshop CS6).

    Edit @FranceBB: You're right about ProxHTTPSProxyMII of course. Just saying, so you don't think I'm ignoring your suggestions. :) But if there's a way to get this done with some Windows Update on XP x64, I'd prefer that way. If at all possible.

  8. TLS v1.2 doesn't work. Like I said, the corresponding Windows update simply doesn't exist for XP x64 (or Server 2003 x64). There is no x64 version of POSReady2009 after all, so my Windows SSL is stuck at TLS v1.0.

    Maybe there is a corresponding update for companies which paid up to get continued support for Server 2003 x64, but no such update was ever leaked to the public as far as I know.

  9. Hello again,

    First: My apologies for double-posting.

    I just tried this on a Windows XP Professional x64 Edition SP2 VM after removing the "Update Root Certificates" Windows component, and this is what it looks like:

    certupdater-vert.png.ffe0986bf7ee4dac97b11541dde1d697.png

    I took a look at the certificate store before and after, and certificates were indeed added and removed. The intermediate certificate authority revocation list stayed the same though. But it appears to be working as intended, I guess? Nothing related was logged into the system logs, no errors either.

    Just a minor suggestion: I think it should say "Root Certificates", not "Roots Certificates".

    Thanks for your work!

  10. Heinoganda,

    I would like to use your updater, but before doing so, I would like to ask whether it works for Windows XP x64 SP2 (NT 5.2) as well? System's English with no MUI installed, but some components differ from regular 32-bit XP. It never got the TLS 1.2 update from the POSReady2009 branch of XP for instance. Also, it has system32/syswow64, like other 64-bit Windows systems, and the registry hive is split between 32/64-bit in a similar fashion.

    Can I still use your program on XP x64, or is this completely untested?

    Thank you!

  11. On 1/17/2020 at 9:05 PM, Mathwiz said:

    GitHub relies way too much on the UA, and none of our browsers provides GitHub with a suitable UA out of the box (or .7z). Well, Serpent 52 does, but only in "native" move (not in FF-compatible mode). So Iset an SSUAO with Serpent 52's native UA string, Mozilla/5.0 (Windows NT 6.1; rv:52.0) Goanna/4.4 Basilisk/52.9.0, for github.com (to be consistent and safe, I also set the same SSUAO for githubassets.com and githubusercontent.com, although I don't think those are strictly necessary).

    Note that FF 60.9 also works, but the MCP's latest browsers now default to FF 68.9, which doesn't work (unless you're running a Quantum browser).

    It's nonsense like this that makes me somewhat sympathetic to what Google is proposing to do with user agents. Web sites should base their functionality on the capabilities of the browser, not its version or what OS it's running on.

    [...]

    Thank you very much! I set the UA for those three domains as you said and set general.useragent.compatMode.firefox from true to false. Then I went back to github.com, and yes, everything works now! :)

    Edit: It works on New Moon as well, so I don't have to switch browsers just for github. That was really helpful.

  12. I tried Basilisk 55/Serpent x64, New Moon 28.9.0a1 x64 and KM-Goanna on XP x64 today, but no matter what, I cannot edit my own issues on github.com. Also, the editor is pretty broken so I can't properly write issues either (e.g. the source code button doesn't work).

    The fun part is, that this works flawlessly with Palemoon 28.8.0 on CentOS Linux 64-bit! I have no idea why the OS would make a difference here.

    I tried faking my user agent to FF 71.0, but this didn't change it either, it just got rid of the "your browser is no longer supported" message on github. Tried FF 52.9.0 ESR as well, no luck. Ancient Chromium 49.x browsers have the same issue on XP x64.

    What could I try?

  13. On 10/6/2018 at 6:19 PM, Dibya said:

    How it is running on my pc compiled with vs2017? Please wait I will check if extended XP is not uninstalled properly.

    I compiled myself an ffmpeg version targeting XP 32-bit and also XP x64 last weekend. Built on CygWin, removing bcrypt/schannel and avssource stuff (because I don't have that on CygWin), and replacing the crypto functions with OpenSSL. I can confirm that this works without any Windows API modifications; No Extended XP or anything on my machines, it's all regular Microsoft stuff:

    ffmpeg 4.0.2 64-bit on XP x64

     

     

     

     

    (click to enlarge)

  14. It's not just that. 10+ years ago, Videos were not always encoded in compute-heavy H.264 either. And websites generally used less CPU resources. Plus, modern video de- and encoding relies on SSE and SSE2 a lot for a speed boost (H.264/AVC as well as other modern codecs), additionally to multithreading.

    Edit: roytam1 was faster.

  15. 15 hours ago, sdfox7 said:

    @dencorso

    You might still be in to something with NT 5.x. I assume that New Moon still is compatible with Server 2003 R2 since it shares code with XP. Of course there's no guarantee New Moon works with vanilla Server 2003.

    Perhaps New Moon XP/03? Pronounced "XP-oh-three"?

    Or, "XP-oh-three-are-2"? It kind of rolls off the tongue.

    It most likely is compatible with Server 2003, as it works on XP x64 (which shares its kernel version 5.2 and its APIs with Server 2003). That's what the 64-bit build is for anyway: XP x64 and Server 2003 x64, as there are no 64-bit NT 5.1 systems, that's all NT 5.2.

    Edit: I just tested the 32-bit and 64-bit versions of New Moon on Windows Server 2003 x64 Standard. Works fine, as expected.

×
×
  • Create New...