S3Drive
Visit Live Discord
1
We're pending major release with multiple new features being added. It will also include photo backup from Android/iOS to S3. Different sync modes including two-way sync will be available ~April 2023 - https://s3drive.canny.io/feature-requests/p/syncbackup-options-two-way-one-way
1 Thanks for your help!
1
https, where as IP addr will default to http.
Filename encryption is on our Roadmap and we have a working prototype already. https://s3drive.canny.io/feature-requests/p/filenamefilepath-encryption (ETA ~April 2023).
We're making further research to understand standards or well established implementation in that area, so we can stay compatible.
The sharing functionality is based on S3 presign URLs, their limitation is that the signature can't be valid longer than 7 days, so every 7 days new link would have to be generated. We're researching how to overcome this limitation. For instance we could combine this with a link shortener, so there is single link that doesn't change, but under the hood we would regenerate the destination link as needed.
The encrypted share link has the master key at the end after the # character and looks like this:
https://s3.us-west-004.backblazeb2.com/my-s3drive/.aashare/hsnwye5bno3p/index.html?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=004060aad6064900000000044%2F20230214%2Fus-west-004%2Fs3%2Faws4_request&X-Amz-Date=20230214T095014Z&X-Amz-Expires=604800&X-Amz-SignedHeaders=host&X-Amz-Signature=abdcd875e2106ee54c6a1d1851617c7e694e121464c5ca9023526ce2836be595#GKSGYX4HGNAd4nTcXb/GIA==
What it does it tries to load the encrypted asset as usual, but it's not aware per se if an asset is encrypted. In the background JavaScript tries to fetch the asset and replaces the one on the screen with decrypted version. It looks like it has failed on your side. Can you go to the console (right-click -> inspect element) to see if there is anything abnormal (that is error in the Console or different than 200 status code in any of the network requests).
1. I'm happy to hear that folder name and file name is on the near roadmap for completion. Usability problem solved
2. The S3 presign URL concept is interesting. Can a link be done for a time period shorter than a week? I assume the user would need to have their program opened and some kind of timer to check if any links need to be signed again. Then the shortened link would need to be updated. It's a good idea actually but would want to make sure it's done in a trustless or trust minimized way. Also be careful deploying resources that you have to pay for as then you'd have ongoing costs. It depends what your monetization plan is long term I guess.
3. If the link is presigned, am I right to assume that there is no ability to revoke the sharing?
4. I'm actually not familiar. What is PhotoSync?
5. The link appears to work fine in Chrome but in Firefox it stops loading the page faster. If I click refresh, the image appears as it should. Strange but seems like all is fine. The things I would typically be sharing are not photos so it's not really a priority issue for me.
6. I'd love it if S3 drive had a different favicon so I can more easily distinguish it. I'd love it more if it was mounted as a drive letter in my file explorer. 
Actually, I think what I'd personally use most is if I could right click, go to "send to" and have S3 drive as a destination. I'd have to think about if I'd want it in the main folder or prompt me where to put it. I'd probably choose to have a popup to prompt me where I want it saved within S3 drive.
1. I'm happy to hear that folder name and file name is on the near roadmap for completion. Usability problem solved
2. The S3 presign URL concept is interesting. Can a link be done for a time period shorter than a week? I assume the user would need to have their program opened and some kind of timer to check if any links need to be signed again. Then the shortened link would need to be updated. It's a good idea actually but would want to make sure it's done in a trustless or trust minimized way. Also be careful deploying resources that you have to pay for as then you'd have ongoing costs. It depends what your monetization plan is long term I guess.
3. If the link is presigned, am I right to assume that there is no ability to revoke the sharing?
4. I'm actually not familiar. What is PhotoSync?
5. The link appears to work fine in Chrome but in Firefox it stops loading the page faster. If I click refresh, the image appears as it should. Strange but seems like all is fine. The things I would typically be sharing are not photos so it's not really a priority issue for me.
6. I'd love it if S3 drive had a different favicon so I can more easily distinguish it. I'd love it more if it was mounted as a drive letter in my file explorer. Actually, I think what I'd personally use most is if I could right click, go to "send to" and have S3 drive as a destination. I'd have to think about if I'd want it in the main folder or prompt me where to put it. I'd probably choose to have a popup to prompt me where I want it saved within S3 drive. 
1
1. I'm happy to hear that folder name and file name is on the near roadmap for completion. Usability problem solved
2. The S3 presign URL concept is interesting. Can a link be done for a time period shorter than a week? I assume the user would need to have their program opened and some kind of timer to check if any links need to be signed again. Then the shortened link would need to be updated. It's a good idea actually but would want to make sure it's done in a trustless or trust minimized way. Also be careful deploying resources that you have to pay for as then you'd have ongoing costs. It depends what your monetization plan is long term I guess.
3. If the link is presigned, am I right to assume that there is no ability to revoke the sharing?
4. I'm actually not familiar. What is PhotoSync?
5. The link appears to work fine in Chrome but in Firefox it stops loading the page faster. If I click refresh, the image appears as it should. Strange but seems like all is fine. The things I would typically be sharing are not photos so it's not really a priority issue for me.
6. I'd love it if S3 drive had a different favicon so I can more easily distinguish it. I'd love it more if it was mounted as a drive letter in my file explorer. Actually, I think what I'd personally use most is if I could right click, go to "send to" and have S3 drive as a destination. I'd have to think about if I'd want it in the main folder or prompt me where to put it. I'd probably choose to have a popup to prompt me where I want it saved within S3 drive. I'm happy to be an early user and maybe able to contribute somehow along the way. I'm happy to be an early user and maybe able to contribute somehow along the way. 1 1
1
1 1 1
1 
1 2 1 1
1
1 1 1 1 1 1
1rclone for accessing my files or backing up my photos on a day to day basis.... and I am not afraid of CLIs. 
1
garage... and there is no way to provide region in S3Drive.
It seems that we may add additional form field to specify the region.toml file like this: s3_region = "us-east-1".
We auto-detect region from the endpoint URL and have a way to detect custom region from MinIO.... and if it doesn't work we use the most common default which is us-east-1.
1 1
1 1
The Discord server is hard to join from the app since it opens a webview, I'd recommend opening the invite externally so the Discord application can pick it up, or at least the standard web browser.
Trying out the application with MinIO, so far so good, pleased by the rename function that likely performs a copy + delete under the hood. It does what one would expect from an S3 client, no ads either, really liking it. .aainit file is our write test, as well as ETag response validation (which is required for not yet released syncing features), as some providers (talking mostly about iDrive E2 with SSE enabled) don't generate valid ETags. BTW. Would you like S3Drive to support read-only mode?
Regardless, we will try to improve clarity of this operation, so user feels more confident that we're not doing some shady write/reads.
Speaking of Trash itself, likely this week starting on Android first there will be a Settings option to disable Trash feature altogether (which is a soft-delete emulation, but slow and pointless if bucket already supports versioning). Versioning UI with restore options will come little bit later. .aainit file it's fine, but I'd prefer if the app saved the test results locally then deleted the file. I want to be able to write files so I wouldn't use a read-only mode, and we can always create read-only access keys if we want to be sure that's how the app will behave! I'm very interested by the share link expiry slider or date picker though, I never share for 7 days, it's either a smaller duration or permanent.
Cool, I don't mind not having the versioning UI yet, but had to delete my file versions + the trash versions to cleanup my bucket so… yeah, trash is cool but I assume most people who want that have versioning enabled. I assume you already have quite a few buckets on various providers to test your features, but I can provide a MinIO one if it could be of interest.
There was a 2nd folder with an HTML page in it, not sure what it was about but same thing I'd say, that's probably the least expected action from an S3 browser… While I audited the actions and indeed didn't find anything malicious, that could get me assassinated by my colleagues if I ever connected a more important bucket to the app. 
.aainit file it's fine, but I'd prefer if the app saved the test results locally then deleted the file. I want to be able to write files so I wouldn't use a read-only mode, and we can always create read-only access keys if we want to be sure that's how the app will behave! I'm very interested by the share link expiry slider or date picker though, I never share for 7 days, it's either a smaller duration or permanent.
Cool, I don't mind not having the versioning UI yet, but had to delete my file versions + the trash versions to cleanup my bucket so… yeah, trash is cool but I assume most people who want that have versioning enabled. I assume you already have quite a few buckets on various providers to test your features, but I can provide a MinIO one if it could be of interest.
There was a 2nd folder with an HTML page in it, not sure what it was about but same thing I'd say, that's probably the least expected action from an S3 browser… While I audited the actions and indeed didn't find anything malicious, that could get me assassinated by my colleagues if I ever connected a more important bucket to the app. s3 ls even though headObject couldn't retrieve it as a valid S3 entry. I am curious if you came across of something similar.
... but seriously, it's actually one of the coolest things I've been working on in my 10+ years career and seeing already such positive reception we've no intentions to stop. 
1 I love that one: https://xkcd.com/927/
I love that one: https://xkcd.com/927/ .aainit file it's fine, but I'd prefer if the app saved the test results locally then deleted the file. I want to be able to write files so I wouldn't use a read-only mode, and we can always create read-only access keys if we want to be sure that's how the app will behave! I'm very interested by the share link expiry slider or date picker though, I never share for 7 days, it's either a smaller duration or permanent.
Cool, I don't mind not having the versioning UI yet, but had to delete my file versions + the trash versions to cleanup my bucket so… yeah, trash is cool but I assume most people who want that have versioning enabled. I assume you already have quite a few buckets on various providers to test your features, but I can provide a MinIO one if it could be of interest.
There was a 2nd folder with an HTML page in it, not sure what it was about but same thing I'd say, that's probably the least expected action from an S3 browser… While I audited the actions and indeed didn't find anything malicious, that could get me assassinated by my colleagues if I ever connected a more important bucket to the app. 1.aainit file being nuked (delete file itself + all its versions) once the init is done and raw presigned URL sharing
The Discord server is hard to join from the app since it opens a webview, I'd recommend opening the invite externally so the Discord application can pick it up, or at least the standard web browser.
Trying out the application with MinIO, so far so good, pleased by the rename function that likely performs a copy + delete under the hood. It does what one would expect from an S3 client, no ads either, really liking it. 
Fair enough, we'll plan this in our items along with the slider which needs to be done anyway.
1headObject and get the envelope AES keys.... so it must be a toggle with some warning. It would then simply return the Blob that's stored on S3, regardless of what's inside. 
1
.aainit file it's fine, but I'd prefer if the app saved the test results locally then deleted the file. I want to be able to write files so I wouldn't use a read-only mode, and we can always create read-only access keys if we want to be sure that's how the app will behave! I'm very interested by the share link expiry slider or date picker though, I never share for 7 days, it's either a smaller duration or permanent.
Cool, I don't mind not having the versioning UI yet, but had to delete my file versions + the trash versions to cleanup my bucket so… yeah, trash is cool but I assume most people who want that have versioning enabled. I assume you already have quite a few buckets on various providers to test your features, but I can provide a MinIO one if it could be of interest.
There was a 2nd folder with an HTML page in it, not sure what it was about but same thing I'd say, that's probably the least expected action from an S3 browser… While I audited the actions and indeed didn't find anything malicious, that could get me assassinated by my colleagues if I ever connected a more important bucket to the app. .s3drive_bucket_read_test) and verify the response instead of trying to write a file.
Slider now works, so it's possible to set expiry time shorter than maximum of 7 days. There is an option to use raw preshared URLs.
We've also introduced basic Version UI. It is now possible to preview the revisions. In a next update we will allow opening, preview, deleting and restoring to particular version.
Thank you for these suggestions, they were great and helped us to validate it all !
... and as always we're open for a feedback.
1
... and we're already finding some versioning implementation differences between providers. E.g. MinIO version restore works differently than Backblaze/Wasabi. There are couple bugs (or features?) when deleting entries that we've found on Backblaze... we'll have to deal somehow with all that... but that's fine which extends versioning features, modifies behaviour of copy/rename/move/trash function to seamlessly support versioning and don't duplicate version, also there is an option to restore specific versions or delete object versions to save some space. There is also improved error handling, so in case of an Object Lock there will be an error message and relevant Log which can be viewed by long-tapping the app version number.
Versioning capabilities are configurable in the Settings.
There is a space usage indicator in the left panel / drawer.
There is a new Recent section, which displays all recently modified files.
Search function is improved, so the index is built faster and also includes implicit folder search (e.g. if there is a file: folder/file.txt, but folder/ entry doesn't explicitly exists, it is still searchable)
There is an option to hide files starting with: .
As usual there are couple other performance improvements and bugfixes.
We would love to hear how are you finding new changes and if version management during file operations is what you would expect. 
1 1Hide "." files
Show all files, including starting with the dot.
Hide files starting with the dot character
To:
Hide dotfiles
Show all files, including ones starting with a dot.
Hide files starting with the dot character. 1 1 1
1
In theory this could work. Basically some private key hidden behind the API could generate a verifiable token which would be stored on the user's bucket (and likely device for little bit of redundancy) and could be then verified. The issue that I see so far is that not every user would want their bucket to be written to with some "licensing" file. 
feature_flags int that computes to an array of pro features with bitwise operations, easy on your API and authentication gateway or whatever you do behind the scenes. 
community ! 
S3Drive version (1.2.5) on Web which uses improved rendering engine. Once it loads once all subsequent loads shall be much quicker and the UI shall be way more responsive. I would be glad to hear if you find it an improvement 1
s3.<region>.amazonaws.com 1
1
null response, which is somewhat expected. I would expect to get the SSL related error instead. 1OS Error: CERTIFICATE_VERIFY_FAILED: self signed certificate. And I indeed have self-signed certificate but I followed your instructions from https://github.com/s3drive/app/issues/19 (https://proxyman.io/posts/2020-09-29-Install-And-Trust-Self-Signed-Certificate-On-Android-11) and my browser on Andriod recognizes this certificate (if I go to minio browser, my Chrome is fine with the cert). But S3Drive continues to fail with the same error.
I'm using the latest version. 
Multiple clients (S3Drive being one of them) and multiple back-ends, all speaking 
️ the same language.
If you need my input on anything, just let me know !
It make sense to apply these improvements before focusing on background upload, so there is a stable base. 1 1 2
1
1 
1# Obscure password
echo "YourPlaintextPassword" | rclone obscure -
# Add it to Rclone config, config file location: `rclone config file`
[s3drive_remote]
type = s3
provider = Other
access_key_id = <access_key_id>
secret_access_key = <secret_access_key>
endpoint = <endpoint>
region = <region>
[s3drive_crypt]
type = crypt
filename_encoding = base64
remote = s3drive_remote:<bucket_name>
password = <obscuredPassword>
filename_encryption = standard
directory_name_encryption = true
suffix = none
Then you can use: s3drive_crypt as your remote encrypted location.
Please note that whilst we support both encrypted and unencrypted files in the same location, Rclone doesn't seem to like the mix and won't display existing unencrypted files for the encrypted remote. In such case it's better to either keep everything encrypted globally or have dedicate paths with encrypted-only or unencrypted-only files. 1
1 1
1
1
I will try out ️filename_encoding = base64
suffix = none
By default the Rclone's encoding is base32: https://github.com/rclone/rclone/blob/88c72d1f4de94a5db75e6b685efdbe525adf70b8/backend/crypt/crypt.go#L140 unless overriden by the config creator. filename_encoding = base64
suffix = none
By default the Rclone's encoding is base32: https://github.com/rclone/rclone/blob/88c72d1f4de94a5db75e6b685efdbe525adf70b8/backend/crypt/crypt.go#L140 unless overriden by the config creator. 
1filename_encoding = base64
suffix = none
By default the Rclone's encoding is base32: https://github.com/rclone/rclone/blob/88c72d1f4de94a5db75e6b685efdbe525adf70b8/backend/crypt/crypt.go#L140 unless overriden by the config creator. 
️
I upvoted${aws:username} by anything you want, be it a variable or a fixed bucket name, there unfortunately isn't any group name variable 
One policy for a group and then just add users
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"s3:*"
],
"Resource": [
"arn:aws:s3:::${aws:username}"
]
},
{
"Effect": "Allow",
"Action": [
"s3:*"
],
"Resource": [
"arn:aws:s3:::${aws:username}/*"
]
}
]
}
I will write this to my wiki and make a Post on Lemmy and Reddit. I am surely not the only one steuggeling with this
1 
Contributor role, it isn't much but still a nice way to recognize individuals who go out of their way to help the project out, what do you think about it? 
I read you need to have your profil public in your settings 
️ 1 1AppImage you can find deb package in the releases: https://github.com/s3drive/app/releases if that's any use for you. 1
️ 1 
️
rclone password dump gives obscured password. You need to use your original text password. Alternatively you'll need to use "password reveal" on your obscured password.
https://forum.rclone.org/t/how-to-retrieve-a-crypt-password-from-a-config-file/20051
We're not supporting Rclone 2nd password, but it's part of our roadmap: https://s3drive.canny.io/feature-requests/p/support-2nd-rclone-crypt-password
We're supporting default Rclone salt: https://forum.rclone.org/t/how-to-correctly-setup-the-salt-for-the-crypt-remote/4273/2
I've created additional two roadmap items to support your use case:
https://s3drive.canny.io/feature-requests/p/add-support-for-custom-rclone-salt
https://s3drive.canny.io/feature-requests/p/add-option-to-restore-rclone-password
Please vote on them, so the priority is pushed higher.
If you have any more issues with S3Drive, please create a support item: https://discord.com/channels/1069654792902815845/1102236355645419550
Thanks 
I have got one I have got one 1
1 1
1
1
2 1
1 1 5
1 1
1
1
1️
11.5.3 - https://s3drive.app/changelog
Please try now install newest DMG from our website. It should resolve your issues.
What message did you get exactly from the app? Was it update more recent version available or perhaps that your version has expired? 1 1
Now I have duplicates, but I can easily remove them
This is only a cosmetic bug
Now I have duplicates, but I can easily remove them
This is only a cosmetic bug 1
1
1 1
3FCKGW-RHQQ2... license.
This would either mean that you would have to generate some activation key on our website from time to time and paste it to the app... or once you activate features in your app with some activation key you would have to deactivate it before you could use it on some other Window's client.
2 2 
tocloud, // Upload to remote, delete remotely if file was deleted locally
tocloud_keepdeleted, // Won't remove file remotely if it was deleted locally
tocloud_compat, // If file is removed remotely, local won't know that, it will be reuploaded on a next ocassion
In principle:
"To remote" will upload file to remote and delete it remotely if it was deleted locally. If file is deleted remotely it won't get re-uploaded again.
"To remote (don't delete remotely)" - the same as "To remote", except it will keep file on the remote even if it was deleted locally.
The above 2 options require bucket versioning support.
The "compatibility mode" doesn't require versioning API, however that makes it not aware of any file changes in between, so it's simply blind one way copy instead of sync.
I hope that helps little bit. We'll build documentation once we sort couple challenges related to E2E encryption with syncing, as depending how we manage to solve these problems it may influence the available options.
[
{
"bucketName": "acme-internal-files",
"keyId": "EVLJ2eXJukWUR9U17dyQqq6NPTi9mUu6scqpLCau",
"applicationKey": "X9EiaepygvDK2S0fmMmFayehHoETDOphNP1r96PI",
"endpoint": "https://s3.us-west-004.backblazeb2.com",
"region": "us-west-004",
"host": "s3.us-west-004.backblazeb2.com",
"port": 443,
"useSSL": true,
"encryptionKey": "cG90YXRv",
"rclonePlaintextKey": true,
"filepathEncryptionEnabled": true,
"rcloneDerivedKey": [
116,
85,
199,
26,
177,
124,
134,
91,
132,
...
]
}
]
This may be a good start.
We plan to implement QR code login, but the QR size limitation makes QR not a solution for all use cases.
There are other means, e.g. QR code could transfer the "place holder ID" which would be then used to fetch the required details, but then again this setup would require more moving parts.
We're very much open on this. 
[
{
"bucketName": "acme-internal-files",
"keyId": "EVLJ2eXJukWUR9U17dyQqq6NPTi9mUu6scqpLCau",
"applicationKey": "X9EiaepygvDK2S0fmMmFayehHoETDOphNP1r96PI",
"endpoint": "https://s3.us-west-004.backblazeb2.com",
"region": "us-west-004",
"host": "s3.us-west-004.backblazeb2.com",
"port": 443,
"useSSL": true,
"encryptionKey": "cG90YXRv",
"rclonePlaintextKey": true,
"filepathEncryptionEnabled": true,
"rcloneDerivedKey": [
116,
85,
199,
26,
177,
124,
134,
91,
132,
...
]
}
]
This may be a good start.
We plan to implement QR code login, but the QR size limitation makes QR not a solution for all use cases.
There are other means, e.g. QR code could transfer the "place holder ID" which would be then used to fetch the required details, but then again this setup would require more moving parts.
We're very much open on this. host / gateway, or if you want to set the encryption key, both encryptionKey and generated: rcloneDerivedKey must be provided.
If there is a need we could certainly simplify the format, so things gets smartly derived if not present. host / gateway, or if you want to set the encryption key, both encryptionKey and generated: rcloneDerivedKey must be provided.
If there is a need we could certainly simplify the format, so things gets smartly derived if not present. encryptionKey field is required to setup the encryption.
Speaking of decryption, it's an open format. Naturally you can use S3Drive (on any platform) to access encrypted data (you'll need to access bucket with data and setup E2E with the same password that was initially used for encryption).
You can also mount data using as network drive (that's possible from S3Drive after clicking on tray icon).
Alternatively you can access data using rclone command, as we're 1:1 compatible with their encryption: https://rclone.org/crypt/#file-encryption
In that case please visit our docs to understand how you can set up rclone command: https://docs.s3drive.app/advanced/#setup-with-rclone
Then you would be able to use commands like copy: https://rclone.org/commands/rclone_copy/ or sync: https://rclone.org/commands/rclone_sync/ or couple others depending on your needs.
There are couple options out there. 
3[
{
"bucketName": "bucket-photos",
"keyId": "keyId",
"applicationKey": "applicationKey",
"endpoint": "https://s3.pl-waw.scw.cloud",
"encryptionKey": "cG90YXRv"
}
]
This would configure all necessary things and enable encryption with password: potato, the encryptionKey is base64 encoded plaintext password. [
{
"bucketName": "bucket-photos",
"keyId": "keyId",
"applicationKey": "applicationKey",
"endpoint": "https://s3.pl-waw.scw.cloud",
"encryptionKey": "cG90YXRv"
}
]
This would configure all necessary things and enable encryption with password: potato, the encryptionKey is base64 encoded plaintext password. 
zenity, qarma and kdialog.
https://github.com/miguelpruivo/flutter_file_picker/issues/1282#issuecomment-1551924613
I will add this item to our internal items and try to play around in Xubuntu. In the meantime would you be happy to try out the Flathub version? https://flathub.org/en-GB/apps/io.kapsa.drive (Please note that it awaits 1.6.4 release which will be likely available later today or tomorrow). zenity or kdialog on your OS and see if it solves the issue?
It it does we will add it as a dependency to .AppImage.
https://forum.juce.com/t/native-filechooser-not-used-on-linux-xfce/26347
zenity or kdialog on your OS and see if it solves the issue?
It it does we will add it as a dependency to .AppImage.
https://forum.juce.com/t/native-filechooser-not-used-on-linux-xfce/26347 zenity in our releases.zenity in our releases. zenity in our releases. 
1
1
1
1
1content://, since we operate on network resources, what we get with S3 is just a network URL, that we don't store locally (except the video cache) and pass directly to the video player. Since data isn't stored on Android device locally I don't think there is a method to expose it as a content URI.
If I understand little bit more about your use case I might be able to come up with some other approach. 
discount for Lifetime license 
.
There are only 3 codes to redeem, so whoever is fastest gets S3Drive cheaper.
Happy buying ! 
️
glxinfo | grep "direct rendering" ? glxinfo | grep "direct rendering" ? glxinfo | grep "direct rendering" ? 
1
1
1
1 1
1.7.1 sync feature to be able to interact with the local FS. 1file.png to test/ would rename it to testfile.png and the file is not moved the directory 1.7.0, we've now prioritized this and shall be able to release a hotfix at some point today.
1
Do you have any issue with folder name change?
Do you have any issue with folder name change? 1
1
️
If I consumed 4TB Storage what is the cost? Thanks in advance 
If I consumed 4TB Storage what is the cost? Thanks in advance 1
1libmpv2 as an alternative, but don't really have capacity at the moment to test things out.
Ideally movies should play out as normal, as MPV dependency is required by media library that we use: https://pub.dev/packages/media_kit libmpv2 as an alternative, but don't really have capacity at the moment to test things out.
Ideally movies should play out as normal, as MPV dependency is required by media library that we use: https://pub.dev/packages/media_kit 2 1libmpv version. We're working to have it resolved promptly, please bear with us.git clone --recursive git@github.com:flathub/io.kapsa.drive.git
cd io.kapsa.drive
flatpak-builder --user --install --force-clean build-dir io.kapsa.drive.json
... however it does require some prior environment setup, like:
flatpak install flathub org.freedesktop.Sdk//23.08
flatpak install flathub org.freedesktop.Platform
flatpak install org.freedesktop.Sdk.Extension.vala/x86_64/23.08
We will be providing full guide, "how to compile Flatpak". 
git clone --recursive git@github.com:flathub/io.kapsa.drive.git
cd io.kapsa.drive
flatpak-builder --user --install --force-clean build-dir io.kapsa.drive.json
... however it does require some prior environment setup, like:
flatpak install flathub org.freedesktop.Sdk//23.08
flatpak install flathub org.freedesktop.Platform
flatpak install org.freedesktop.Sdk.Extension.vala/x86_64/23.08
We will be providing full guide, "how to compile Flatpak". 1 ! 1./S3Drive-x86_64.AppImage
(kapsa:2730352): Gdk-CRITICAL **: 09:39:57.636: gdk_window_get_state: assertion 'GDK_IS_WINDOW (window)' failed
package:media_kit_libs_linux registered.
flutter: *** sqflite warning ***
You are changing sqflite default factory.
Be aware of the potential side effects. Any library using sqflite
will have this factory as the default for all operations.
*** sqflite warning ***
method call InitAppWindow
method call InitSystemTray
SystemTray::set_system_tray_info title: (null), icon_path: /tmp/.mount_S3DrivJ2GgY2/data/flutter_assets/assets/logos/logo_42.png, toolTip: (null)
method call CreateContextMenu
value_to_menu_item type:label, label:Show
value_to_menu_item type:label, label:Hide
value_to_menu_item type:label, label:Start drive mount
value_to_menu_item type:label, label:Stop drive mount
value_to_menu_item type:label, label:Start WebDav
value_to_menu_item type:label, label:Stop WebDav
value_to_menu_item type:label, label:Support
value_to_menu_item type:label, label:Visit Website
value_to_menu_item type:label, label:About
value_to_menu_item type:label, label:Changelog
value_to_menu_item type:label, label:Logs
value_to_menu_item type:label, label:Version 1.7.11
method call SetContextMenu
Just a question, did you try running Flatpak format? https://github.com/flathub/io.kapsa.drive/./S3Drive-x86_64.AppImage
(kapsa:2730352): Gdk-CRITICAL **: 09:39:57.636: gdk_window_get_state: assertion 'GDK_IS_WINDOW (window)' failed
package:media_kit_libs_linux registered.
flutter: *** sqflite warning ***
You are changing sqflite default factory.
Be aware of the potential side effects. Any library using sqflite
will have this factory as the default for all operations.
*** sqflite warning ***
method call InitAppWindow
method call InitSystemTray
SystemTray::set_system_tray_info title: (null), icon_path: /tmp/.mount_S3DrivJ2GgY2/data/flutter_assets/assets/logos/logo_42.png, toolTip: (null)
method call CreateContextMenu
value_to_menu_item type:label, label:Show
value_to_menu_item type:label, label:Hide
value_to_menu_item type:label, label:Start drive mount
value_to_menu_item type:label, label:Stop drive mount
value_to_menu_item type:label, label:Start WebDav
value_to_menu_item type:label, label:Stop WebDav
value_to_menu_item type:label, label:Support
value_to_menu_item type:label, label:Visit Website
value_to_menu_item type:label, label:About
value_to_menu_item type:label, label:Changelog
value_to_menu_item type:label, label:Logs
value_to_menu_item type:label, label:Version 1.7.11
method call SetContextMenu
Just a question, did you try running Flatpak format? https://github.com/flathub/io.kapsa.drive/ 
1
1
1
1
1
1️ 1 
1
2
1
And how would you list the deleted of deleted files?
And how would you list the deleted of deleted files? bash evidence and going to ask publicly on MinIO's Github.
The issue isn't complex at all. Basically you have trash on Windows/Linux/macOS whatever. If you delete files from your computer, they land in Trash. We can say they're versioned as their latest version is available for restore.
From a UI point of view, after deletion, you wouldn't expect for these deleted entries to appear in a location from where they were originally deleted. They're now in Trash (available for further deletion or restore) and shouldn't be present anywhere else.
MinIO shows the folder hierarchy in the original location despite that it was all deleted and it's correct place is Trash. I don't think it's a correct behavior from purely "files&directories" UI point of view. PUT folder/
PUT folder/file.txt
then: LIST folder might not return you the folder/file.txt
Apparently they mention it in here: https://min.io/docs/minio/container/operations/checklists/thresholds.html#id6
I am not sure, but we may have to change the way we create folders to overcome this issue. E.g. instead of folder/ we would rather create folder/.empty to not cause conflicting keys. ️
I read about the workaround. Thanks
10y old
1
1 1
1x-amz-object-lock-mode: ObjectLockMode
The Object Lock mode that you want to apply to this object.
Valid Values: GOVERNANCE | COMPLIANCE
x-amz-object-lock-retain-until-date: ObjectLockRetainUntilDate
The date and time when you want this object's Object Lock to expire. Must be formatted as a timestamp parameter.
x-amz-object-lock-legal-hold: ObjectLockLegalHoldStatus
Specifies whether a legal hold will be applied to this object. For more information about S3 Object Lock, see Object Lock.
Valid Values: ON | OFF
It's a matter of providing sane settings UI where these settings can be applied.
Depending on the requirements there could be multiple layers with override rules. For instance user could specify settings on the bucket level which would then be overridden by the settings on the folder level, then on the sub-folder level (and so on) down until the file level.
We're open for suggestions how this should/could work. 1
1 1 2 1 
️
S3Drive is now available on Flathub: https://flathub.org/apps/io.kapsa.drive which is an alternative to our deb and AppImage packages: https://github.com/s3drive/app/releases
cc @helios6509 cc @morethanevil 1 2 1
1
1
Most important features implemented:
Allow sharing files from other apps and upload it to S3Drive (Android),
Allow opening file externally, editing and saving it back to S3Drive - change is detected automatically,
Add WebDav server for desktop,
Import all buckets feature / autodiscovery,
Folder ZIP download,
Folder upload,
Flathub release: https://flathub.org/apps/io.kapsa.drive
Work in progress:
Background media backup,
Upload constraints (only wifi, only charger, battery not low),
Custom backup location,
Add a storage class option,
Allow to share file from external app to S3Drive (iOS)
Detailed changelog: (https://s3drive.app/changelog)
[1.5.3] - 2023-09-22
Search by name, size, file type, date modified,
Import all buckets feature,
Folder ZIP download,
Fix bucket listing for Web,
Folder upload for Web
[1.5.2] - 2023-09-14
Allow opening file externally, editing and saving it back to S3Drive,
Fix alphabetic order of folders and filenames,
Use “application/octet-stream” content type for encrypted paths,
Allow refreshing results on mobile by pull,
Add refresh button on desktop/web
[1.5.1] - 2023-09-08
Allow sharing files from other apps and upload it to S3Drive,
Add WebDav server for desktop,
Fix version “prefix issue” match during hard delete, trash delete and rename operations,
Fix remaining storage indicator invalid value
[1.5.0] - 2023-08-21
Managed account with no bandwidth / API rate limit,
Storage quota - display account plan capacity
[1.4.9] - 2023-08-13
Fix swipe left/right on Android/iOS,
Preserve folder path when quitting preview,
Display port in profiles,
Improve filename decryption for paths with mixed encrypted/unencrypted segments 1
1 2 1
1
/public/ directory which is configured as world-readable.
Whenever I share a file from this bucket, I just want https://domain.tld/bucket/public/file.ext 1 1
btw, if I encrypt it using S3Drive, will I be able to stream the video or I need to download the file first to play video?AndroidManifest.xml and this is how they do it (plus implementation of course).
https://github.com/bitfireAT/davx5-ose/blob/273deecbe49b9f0c5ae753353ad0f8a514c4c401/app/src/main/AndroidManifest.xml#L288-L296
Thank you for your hard work, using S3 Drive and liking it a lot!
3
1 1s3fs and run the WebDav server yourself as you say, but you could also achieve the same with our native rclone mount which is likely going to be more performant than a POSIX compatible s3fs.
goofys, whereas s3fs offers maximum POSIX compatibility at huge cost. E.g. listing directory with 1000 files will take up to 1000! requests with s3fs, however it will take just one with rclone / goofys. 1s3fs in order to be POSIX compatible in some cases needs to issue 1000x times more requests. This has dramatic performance consequences.
You're right you won't be able to specify group permissions for a dir. (Actually it may work, but such data won't be preserved when you remount). 1 1 1-rw-rw-r-- which is pretty default Rclone setting, but in principle we could add settings configuration to change it.rclone command it's probably not a high priority.
When you go to the app logs, you will see exactly what commands app executed on rclone binary. You can replicate the same on your server if you wish.
1goofys and even had an AWS client-side encryption compatible prototype, ultimately we've replaced it with Rclone, however experience with goofys and its codebase was pretty good. goofys and even had an AWS client-side encryption compatible prototype, ultimately we've replaced it with Rclone, however experience with goofys and its codebase was pretty good. 
1
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"s3:List*"
],
"Resource": "*"
}
]
}
bucket can be set up without problems despite the read check.
Upload/download naturally wouldn't work, but that's expected. (Please note that these error responses come from 1.6.1 version which is due to be released. In older version errors might be rendered differently).
Drive mount does also seem to mount properly and listing works.
What's your permission set and S3 provider which gets you to: "Access denied"? I would be happy to try that out. Thanks ! 
1 1
1 1
1
rclone ls driveName: shall give you some files in the listing.
Alternatively you can use about command, e.g.: rclone about driveName: and you should get e.g.:
Total: 7 GiB
Used: 611.346 MiB
Free: 6.403 GiB
If you're not getting these results, STOP and try setting back-end again using: rclone config as mentioned here: https://docs.s3drive.app/setup/import_rclone/
If things are working good for you at this stage, then use: rclone config dump command in order to extract all configs, then manually select, copy and paste the relevant Google Drive config into S3Drive (click new "+" and import). 1️1.65.1 to 1.65.2 in a next release end of this week / Monday, if this doesn't resolve this issue, then we'll escalate. 1
1
They have said they'll be making desktop and mobile apps at some point but i won't hold by breath!
here's their kickstarter from 5 years ago if you're interested! https://www.kickstarter.com/projects/cubbit/cubbit-reinventing-the-cloud
Thanks again, really glad I found S3Drive! 1
Thanks for listening!
1
It is a rare case
1
){kind=link}
which cat | grep libc
strings /usr/lib/x86_64-linux-gnu/libstdc++.so.6 | grep GLIBCXX command is used it shows: GLIBCXX_3.4.32?
https://stackoverflow.com/a/77075793/2263395
Couple weeks ago, there was an upgrade to our build machine, which may or may not affected the newest glibc required. If possible I would still advise to use Flatpak release which is supposed to solve these issues.
In the meantime we'll try to confirm if we can somehow bundle glibc of fallback to older version.
1
We may add additional option from the settings menu to slightly improve chances of people finding it.
1
1
1
.bin extension it does look to me as if you weren't using the filename encryption for these files.
This is the config that works with S3Drive:
https://discord.com/channels/1069654792902815845/1069654792902815848/1135157727216279585
If filename encryption is off: https://rclone.org/crypt/#crypt-filename-encryption (default is standard), then the .bin suffix gets added.
We don't really support stripping the .bin suffix, that's why in the config that we recommend we suggest (Discord link above) we suggest to disable it: https://rclone.org/crypt/#crypt-suffix
Given that you already have some data, perhaps we could reconsider support for: .bin stripping for users which used the default setting before running it all up in compatibility with S3Drive.
We're almost there and it works on our debug releases, but for some reason it doesn't work on live release. We're tracing the issue down.
We hope to have it solved this month, we then plan to implement couple other tricks & hacks and gotchas suggested by the community to make it more "background" like, because these two modes differ a lot between iOS and Android.
Stay tuned for next updates ! 1
How can we make it?
1
Maybe for exiting the app a double tap can be used.
Maybe you can move the logs in the about page as an item like changelog, webpage etc.
1
Issue solved
last modified field date that we display early on gets directly from S3 and is technically last modified on the remote side, but in fact the real local modification date is stored as: x-amz-meta-mtime header. 
1
At the moment app starts with full UI, in the future we will fix it, so once it starts it's minimized and stays in tray which I believe would be preferred for most users.
3 1
1seahorse package and the necessity of unlocking keyrings.
https://stackoverflow.com/a/77338413/2263395
As soon as I find a clear solution I will let you know. 
️
https://docs.aws.amazon.com/general/latest/gr/rande.html
️️
[
{
"AllowedHeaders": [
"*"
],
"AllowedMethods": [
"GET",
"PUT",
"POST",
"DELETE"
],
"AllowedOrigins": [
"https://web.s3drive.app"
],
"ExposeHeaders": []
}
] 1 <CORSConfiguration xmlns="http://s3.amazonaws.com/doc/2006-03-01/">
<CORSRule>
<ID>S3Drive</ID>
<AllowedOrigin>https://web.s3drive.app</AllowedOrigin>
<AllowedMethod>GET</AllowedMethod>
<AllowedMethod>HEAD</AllowedMethod>
<AllowedMethod>POST</AllowedMethod>
<AllowedMethod>PUT</AllowedMethod>
<AllowedMethod>DELETE</AllowedMethod>
<MaxAgeSeconds>3600</MaxAgeSeconds>
<ExposeHeader>etag</ExposeHeader>
<ExposeHeader>x-amz-meta-x-amz-key</ExposeHeader>
<ExposeHeader>x-amz-meta-x-amz-iv</ExposeHeader>
<ExposeHeader>x-amz-meta-x-amz-cek-alg</ExposeHeader>
<ExposeHeader>x-amz-meta-x-amz-wrap-alg</ExposeHeader>
<ExposeHeader>x-amz-meta-x-amz-key-v2</ExposeHeader>
<ExposeHeader>x-amz-meta-x-amz-tag-len</ExposeHeader>
<ExposeHeader>x-amz-meta-x-amz-unencrypted-content-length</ExposeHeader>
<ExposeHeader>x-amz-version-id</ExposeHeader>
<ExposeHeader>x-amz-meta-key</ExposeHeader>
<ExposeHeader>x-amz-meta-iv</ExposeHeader>
<ExposeHeader>x-amz-meta-chunk</ExposeHeader>
<ExposeHeader>x-amz-meta-cek-alg</ExposeHeader>
<ExposeHeader>x-amz-meta-s3drive</ExposeHeader>
<AllowedHeader>*</AllowedHeader>
</CORSRule>
</CORSConfiguration>
Please note that in your configuration there are at least couple things missing which may affect some features.
For instance HEAD HTTP operation or: x-amz-version-id and x-amz-meta-s3drive (optional) headers. 1 2 1
1
rclone v1.63.0 and configured my storage like:
[storj]
type = s3
provider = Other
access_key_id = <redacted>
secret_access_key = <redacted>
endpoint = https://gateway.storjshare.io
[storj_crypt]
type = crypt
filename_encoding = base64
remote = storj:my-photos
password = <redacted>
filename_encryption = standard
directory_name_encryption = true
suffix = none
I've then copied test pdf like this: rclone copy test.pdf storj_crypt: and I get valid object with ETag and openable within S3Drive.
Can you post your Rclone storj remote configuration? You've initially posted config, but that's just crypt: https://discord.com/channels/1069654792902815845/1159814485515710525/1159827592732479590
Amazon S3 (or S3 compatible) has: MD5. 2
1LE2123.
1MinioError: ListObjectsV2 search parameter maxKeys not implemented 1
1content-length header in the response (e.g. during file download/open). We rely on it to display transfer progress results as well as to make certain predicaments related to encryption. In theory we could implement content-length workaround, but it would take us little bit while.
We're going investigate first whether it is possible to enable that header. I know that Cloudflare has some logic behind the content-length header which in some cases is provided and in some isn't... we'll have a look on it as well, however if you're going to reach out Cloudflare it is something you can ask about as well. Thanks 
1
content-length issue on mobile and desktop clients. Web client fix will have to wait little bit longer (because we've no control of content-length header in the browser).
Basically if: accept-encoding HTTP request header includes gzip Cloudflare seem to skip content-length altogether.
We're testing couple things right now, but if things go well we'll be able to release it promptly.
Related: https://community.cloudflare.com/t/no-content-length-header-when-content-type-gzip/492964 1 1
Thanks for your patience.
1.5.7 will be released live early next week, but you can comfortably stay at your pre-release build. 1
1
