uskudnik / amazon-glacier-cmd-interface

Command line interface for Amazon Glacier
MIT License
374 stars 100 forks source link

Future plans #93

Open offlinehacker opened 11 years ago

offlinehacker commented 11 years ago

@wvmarle and also some other coders have made a lot of progress lately and glacier-cmd is really becoming more and more stable software. But we still don't have a release yet.

Key points we have to ask:

Please express your opinions, so we will be able to make future plans.

wvmarle commented 11 years ago

Q1 should be "What will be included in release?" Then Q2 becomes "when list of Q1 is finished".

I think boto is a great foundation upon which to build our wrapper and command line interface. Primary additions are bookkeeping and upload/download resumption, and overall making the whole thing more user friendly.

offlinehacker commented 11 years ago

I think first we need to write tests, but before that we need to know what we will use from boto(so we will be able to mock correct functions from boto). So it would be nice to check boto's paralell uploads and check if their code can be used.

Me and @uskudnik will have sprint here in Slovenia one day, and we will try to achive and setup testing framework, write tests and use travis-ci for testing. We will also refract code, so it will follow PEP specifications. After that untested code and one not following PEP won't be merged.

gburca commented 11 years ago

What's going on with this project? Is it dead? There doesn't seem to be much progress, and there are some show-stopper issues that have been opened for over a month with no fixes in sight.

I realize this is an open source project, and I don't expect immediate resolution to every little thing since people contribute when and if they're able. However to have major breakage (like that reported in issue #95 for example) go unresolved for a month raises questions about the vitality of this project.

I used to contribute until about 2 months ago when a slew of poorly tested features were added without taking the time to stabilize the master branch. Coupling that with a number of merge issues by some people who were not familiar with git resulted in a development tip that seems to be completely broken for any kind of serious use. Maybe that's just my personal opinion and most people are using this tool without issue.

My fork is stuck at commit 0c80e969b6f which works like a charm. I've used it to upload files that were over 70Gb from a cron job that uploads a number of archives daily, and I've had with no problem with it. I'd like to move up to the latest to pick up the new features and continue to contribute, but not if it means uploads over 13Gb fail (issue #95) and stdin uploads are broken (issue #113) or uploads just mysteriously break (issue #90) or I'll run into all kinds of issues with mmap (issue #94).

This is a great tool, and with a little bit of love could be made to shine and operate reliably as an interface to Glacier.

offlinehacker commented 11 years ago

I will explain what is the problem with project, besides me not having any free time lately, and having some personal problems.

The major problem is we don't have tests. We didn't took test-driven development approach, but we just coded until the point where so much effort has to be put in to test everything that i lost my motivation somewhere in betwene and this is also partially my fault. You can add features one by one, but code is going to become very unpredictable. I belive that some version of code works reliably, but is it going to work reliably when somebody adds some feature, or is it going to break and introduce new bugs. Test-driven development eliminates most of these problems. As i said i made a mistake to not scream before how badly we need tests. Boto support for glacier might not be very good, but they have proper tests, so +1 for them...

...we are still waiting for a brave man to write propper tests for glacier-cmd and i can say i am not yet brave enough.

Of course we can release your code wvmarle if it is working stable, but for future as you said with a bit of love we can write tests for all our current funcionality, and make this code quality as it should be.

On Sat, Dec 8, 2012 at 8:12 PM, Gabriel Burca notifications@github.comwrote:

What's going on with this project? Is it dead? There doesn't seem to be much progress, and there are some show-stopper issues that have been opened for over a month with no fixes in sight.

I realize this is an open source project, and I don't expect immediate resolution to every little thing since people contribute when and if they're able. However to have major breakage (like that reported in issue

95 https://github.com/uskudnik/amazon-glacier-cmd-interface/issues/95for example) go unresolved for a month raises questions about the vitality

of this project.

I used to contribute until about 2 months ago when a slew of poorly tested features were added without taking the time to stabilize the master branch. Coupling that with a number of merge issues by some people who were not familiar with git resulted in a development tip that seems to be completely broken for any kind of serious use. Maybe that's just my personal opinion and most people are using this tool without issue.

My fork is stuck at commit 0c80e96https://github.com/uskudnik/amazon-glacier-cmd-interface/commit/0c80e969b6fwhich works like a charm. I've used it to upload files that were over 70Gb from a cron job that uploads a number of archives daily, and I've had with no problem with it. I'd like to move up to the latest to pick up the new features and continue to contribute, but not if it means uploads over 13Gb fail (issue #95https://github.com/uskudnik/amazon-glacier-cmd-interface/issues/95) and stdin uploads are broken (issue #113https://github.com/uskudnik/amazon-glacier-cmd-interface/issues/113) or uploads just mysteriously break (issue #90**) or I'll run into all kinds of issues with mmap (issue https://github.com/uskudnik/amazon-glacier-cmd-interface/issues/90#94https://github.com/uskudnik/amazon-glacier-cmd-interface/issues/94 ).

This is a great tool, and with a little bit of love could be made to shine and operate reliably as an interface to Glacier.

— Reply to this email directly or view it on GitHubhttps://github.com/uskudnik/amazon-glacier-cmd-interface/issues/93#issuecomment-11162288.