Preface: as of May 2023, the system introduced on this page has been superseded by the newer PC-Phone USB Sync—a standalone backup/sync app that runs on both Android 8-and-later
, as well as Windows, macOS, and Linux
. Read about and fetch the new app at its separate
, and tap Details in the follow-up page's
for more info.
Published October, 2021 This page is a continuation of the Android 11 saga started
and left
. It both reviews the
for keeping content on a new Samsung flagship phone, and details a
that was developed to manage content with
on this device.
The work-around chronicled here is a partly manual process which requires a new script and a dedicated proxy drive, and may at first seem less convenient than the USB access or microSD cards available earlier. But it passes as a fix of sorts, takes roughly the same amount of user interaction, and can be used with any Android 11 and later device that has been similarly handicapped by both its hardware and software makers (at this writing, Android 12 is still
).
A note up front: this page does not reflect a vested interest of any kind in Samsung or its products. The phone focus here stems from a first-hand evaluation, and is rooted in the fact that this company has been a leading Android vendor for years. While other phones' mileage may vary, the reality is that Samsung's moves have almost as much weight on the Android world as those of Android itself.
Samsung Cripples Its Flagships by Removing microSD
In August 2021, Samsung released its newest flagship Android phone, the
. This device is clearly interesting from a hardware perspective—it's both phone and mini tablet, with an amazing screen that brings content creation into scope, and makes it difficult to go back to the standard candy-bar form factor and its porthole
.
In a word, this phone is disruptive. Its cameras have room for improvement, and time will tell if smartphone users accustomed to placing their phones in situations as perilous as a back pocket will warm to the new folding paradigm. But this device already feels like the future.
The Bad News
At least at first glance, however, the Fold3 also seems to be unusable as a host for non-trivial content collections managed by portable programs like the
content-sync system. Despite its new-age aesthetics, this device:
Has no support for microSD cards, unlike the Note models it seems aimed to
Ships with Android 11, which restricts USB access to its non-POSIX Java
The combo of these two yields a content-island device, with neither removable storage nor open USB access. You can't pop a card out and manage it on your PC, and you can't plug in a USB drive and access it from
-based programs run on the phone. Hence, there is no direct way to store and manage large collections of content on this device using Python programs like Mergeall. This is true for the large 200G collection maintained by Mergeall's proprietor, but much smaller collections will fare the same.
The Damage
To be clear, USB drive access isn't fully gone in Android 11: it's just been locked down so it can be accessed only through Android's proprietary Java framework. Programs coded in Java, such as file explorers, can still do so, though many or most undoubtedly required significant changes to work in 11 (and some
). This, however, qualifies as a kiss of death for portable POSIX-based programs like those coded in Python, which rely on device-neutral pathnames and system calls to work across multiple platforms.
Tools and techniques that allow Python code to call out to Java code have long been available, and might in principle make it possible to interface with Android's Java frameworks to access USB content. However, while findings are still tentative, this seems likely to be:
Inconvenient in the extreme. In principle, every IO operation might have to use special-case code for content on USB drives, on Android alone. Such massive rewrites are easily preclusive for interoperable programs which employ POSIX calls for USB content that work perfectly well elsewhere, and worked perfectly well on Android until 11. It's not hard to see that going proprietary means limiting users' choices.
Impossible in Android Python apps. In testing so far, none of the usual Python-to-Java bridges has proved to be usable for code run in existing apps. In Pydroid 3, for example,
can be installed but raises a SystemError exception when imported sans tkinter, and the app manifest is beyond the reach of Python programs. Developing standalone apps that embed Python and tkinter just to skirt such issues seems radical overkill.
Hence, as of Android 11, there is no direct way to access USB-drive content from Python code. And there's no clear way to use any POSIX-based software to update on-phone content through the USB port. Samsung's removal of microSD support appears to seal content's last escape hatch, and finish the work that Android 11 began.
The Options
difficult not to see this as a vendor closing the doors that provided alternatives to cloud storage. With either removable microSD or unlocked USB, content could be stored and maintained on Android 11 phones, in either shared storage or app-specific storage. To be sure, both options have substantial tradeoffs in Android 11:
Shared storage (e.g., /sdcard/folder) is open but slow. It's usable by just as many apps as it was before, but has been throttled down radically in Android 11. Per the coverage
, shared storage typically runs 10X-100X slower than it did on earlier Androids, which can make on-phone content management tedious or implausible.
App-specific storage (e.g., /sdcard/Android/data/com.termux/folder) is faster, but has multiple problems. It makes content off-limits to other apps (e.g., Pydroid 3 can't access content in Termux's storage, and vice versa); it's prone to accidental deletion on app uninstall (sans the new manifest
, which apps likely won't have); it's a read-only resource if Unix permissions are copied to it (see the
); and accessing its content requires file explorers, viewers, and editors with enhanced app-storage permissions (see the status
).
Despite their downsides, though, both options are at least viable, and reflect unfortunate but workable sacrifices for on-phone content storage in Android 11, made arguably more palatable by devices like the Fold3.
But with neither microSD nor USB, internal on-phone storage seems a moot point. There's no direct way for a Python script to get the data onto a phone, and no way for a Python script to incrementally update the data once there. As we saw
, network-based solutions are far too slow or buggy to be practical for non-trivial content; and file-explorer copies by USB, when they work at all, may botch modification times, permissions, and more, and won't help for later content-change synchronizations in any event.
Conclusions 1.0
Thus, the Fold3, like most of Samsung's other recent
, initially appears to be unusable for storing and maintaining large content collections with POSIX software. It may be viable for smaller content sets for which full drag-and-drop copies suffice to propagate changes, and for which cloud options are feasible—and this, alas, may be one of the real goals of this device. But larger collections, at least on first inspection, are simply out of scope on this phone.
More fundamentally, this phone ships with 256G or 512G internal storage, and no good means to populate it, apart from occasional media downloads, exhaustive and perilous brute-force copies, and software that slavishly complies with Android's increasingly authoritarian requirements. That's a kind of device, but it's not a PC replacement, and it's not going to attract a lot of quality programs. While other vendors still make phones with microSD support, the fact that the leading Android hardware maker is now crafting content islands seems a likely and regrettable harbinger of Android's closed future.
A Mergeall Work-Around to the Rescue: Delta Sets
All the forgoing being said, the sky hasn't fallen in full.
For users of the Fold3—and all other content islands like it—the Mergeall team has come up with a new way to use your phone's USB port and Mergeall to update on-phone content to match that on a PC (where PC means any Windows, macOS, or Linux device). It's a partly manual procedure that uses custom code and a new script that batches changes, and requires a dedicated external drive to serve as a proxy for the phone. But it does not require POSIX USB access or microSD card, and does not rely on phone rooting or any of the impractically slow network-based alternatives to USB which we explored
.
Work-Around Overview
This fix is largely automated by scripts we'll meet in moment, but as a summary, here are its core concepts and steps:
Reserve a proxy This fix requires a USB drive to hold a copy of your content which will serve as a stand-in for the phone. This drive is called a "proxy" because it takes your phone's place when syncing from your PC. Setup devices This fix also requires Python, Mergeall's source code, and the scripts' package on PC and phone, along with some simple edits to config files to specify content paths. Copy content To get started, you'll run a script on your PC and phone to initially copy your content to a proxy drive, and from there to your phone. This requires a manual copy step on the phone because POSIX scripts can't access USB, but you'll copy just one content zipfile. Sync changes For all later syncs, you'll run a script on your PC and phone to merge all the changes made on your PC into the content copies on both the proxy and your phone. Again, this requires a manual copy step on the phone, but for just a single changes zipfile. Beyond this, additional scripts are provided for verifying and exporting content stored on your phone, though these are optional and occasional tasks.
The Deltas Script
Much of the "magic" behind this work-around is the new Mergeall script,
. This is a variation of the main
script. It collects changes (a.k.a. deltas) required to make the TO folder the same as FROM as usual, but has been modified to store these changes in a folder as a separate set, instead of applying them to TO immediately. The deltas script was initially coded to collect and archive recent changes to data already burned to optical disk, but it can be an advantage anytime a deferred changes set is useful.
In particular, delta sets have two crucial features for our purposes here. First, because they contain only recent changes, delta sets are generally very small, and can be copied much quicker than full content. Second, because deltas are stored in the same format as Mergeall backups, the main mergeall.py script's -restore mode can be used to apply them on demand to any content copy that's the same as the copy used to compute the deltas.
This combination of features makes deltas ideal for USB-handicapped phones. Specifically, deltas can be:
Collected by comparing PC to proxy with deltas.py
Applied to the proxy's copy with mergeall.py
Copied to the phone quickly in a single manual step
Applied to the on-phone copy with mergeall.py too
The net effect brings the phone up to date with the PC's changes indirectly, and requires neither POSIX USB access nor microSD card support. In addition, zipping deltas can make them immune to platform and filesystem interoperability issues in transit, and spare them from the vagaries (and outright failures) of raw copies in Android file-explorer apps.
The Work-Around
The best news is that nearly all of the above is automated by free and open-source Python scripts, which you can run with either a simple command line, a file-icon or shortcut click, or an easy step in an IDE GUI. Rather than going into further details here, though, this page will defer to these scripts' own documentation and content for more coverage. Get the Android Deltas Sync fix and its usage docs here:
In closing, it's important to note that this work-around is probably not for casual phone users. Though automated, it requires running command lines or script files, and is nowhere as nice as the former two-merge solution that can be launched with a GUI on both sides of the fence and is still usable on nearly every other platform—including Windows, macOS, Linux, and Androids 10 and earlier.
Android 11 remains the odd platform out in this story (along with the iOS content prison it seems determined to emulate). While 11's restrictions make content-management tasks harder, this fix's scripts at least make them feasible for those who understand that the perils of cloud storage easily justify extra measures to ensure the privacy of your digital property. This is, after all, Mergeall's main point.
