Request feature to use cached IDEs

[glex85]

Hello.

I'd like to ask a feature to enable 'using cached IDE files' instead of downloading them from server.
It would be helpful if you simply add -C option for the curl command to download IDEs.
Would you please check if you can support it?

Thanks in advance!

[ChrisCarini]

Hi @glex85

Perhaps I'm getting myself confused, but it seems like this is two different feature requests:

1. "I'd like to ask a feature to enable 'using cached IDE files' instead of downloading them from server."
2. "It would be helpful if you simply add -C option for the curl command to download IDEs."

Or, perhaps they are the same, and I am just not understanding how they are related. The -C option for curl is for resuming transfers at a specific offset - can you help me understand how that relates to caching IDEs instead of downloading them?

Also, caching IDEs has been on my radar for >4 years, and about 1.5 months ago I did an analysis to understand how that would fit in this action; see here for details: #4 (comment)

The short of ^ is that, while it seems possibly useful to add caching, it might have undesired/unexpected side-effects for repos using this action that they may not expect (i.e. running out of repo-cache / early cache-invalidation because IDEs are large payloads that would fill the 10GB cache faster, evicting other cached artifacts).

Please do help me better understand your request, however, as perhaps I am misunderstanding something!

Best,
Chris

[glex85]

@ChrisCarini Thank you for your answer!
Firstly, I may misunderstood about -C option for CURL. So, you can just ignore this :)

Yes, I'd like to ask to provide way to use cached IDE. In my case, I have my own runner for github action. So, I want to include IDE in my runner and skip 'downloading IDE step' and use stored IDE in my runner. Could you help me to do this?

Thanks in advance.

[ChrisCarini]

Hi @glex85 - happy new year!

Firstly, I may misunderstood about -C option for CURL. So, you can just ignore this :)

Sounds good. 👍

So, I want to include IDE in my runner and skip 'downloading IDE step' and use stored IDE in my runner. Could you help me to do this?

Can you help provide more details here please? Are you looking to:

1. skip just the download portion (but not the unzipping portion)
2. skip both download + unzipping
3. something else

[glex85]

Hi @glex85 - happy new year!

Firstly, I may misunderstood about -C option for CURL. So, you can just ignore this :)

Sounds good. 👍

So, I want to include IDE in my runner and skip 'downloading IDE step' and use stored IDE in my runner. Could you help me to do this?

Can you help provide more details here please? Are you looking to:

1. skip just the download portion (but not the unzipping portion)
2. skip both download + unzipping
3. something else

Hello. Thank you for not missing my request.

If 'this action' able to unzip artifacts which was already downloaded on my runner, 1st option will be better or 2nd option is fine, too.

Happy new year, too :)

[ChrisCarini]

Hi @glex85 - either is possible, however the 'contract' of each would be different, hence me asking.

Specifically, if...

skip just the download portion (but not the unzipping portion)

... then the action would need you to have the zip files in a specific place for it to know to unzip.

FWIW, I'm not opposed to this option, however I do see it as a bit 'wasteful' of resources as you'd have to continue to unzip the same artifacts over and over again on your runner (also, would you be cleaning up the unzipped directories each time, or would that be an expectation you'd have of this action? or?). Granted, it does not take a 'long' time to unzip each IDE, however this would still mean slightly longer CI times for your runner, which is something folks might care about.

skip both download + unzipping

... then the action wouldn't need to care where you have the IDEs unzipped, but instead you'd have to specify the unzipped IDE directories through a different means (specifically, a file at /tmp/ide_directories.txt).

Given the above, do you still prefer option 1 (skip just download, but still unzip the artifact(s))?

[glex85]

Hi @glex85 - either is possible, however the 'contract' of each would be different, hence me asking.

Specifically, if...

skip just the download portion (but not the unzipping portion)

... then the action would need you to have the zip files in a specific place for it to know to unzip.

FWIW, I'm not opposed to this option, however I do see it as a bit 'wasteful' of resources as you'd have to continue to unzip the same artifacts over and over again on your runner (also, would you be cleaning up the unzipped directories each time, or would that be an expectation you'd have of this action? or?). Granted, it does not take a 'long' time to unzip each IDE, however this would still mean slightly longer CI times for your runner, which is something folks might care about.

skip both download + unzipping

... then the action wouldn't need to care where you have the IDEs unzipped, but instead you'd have to specify the unzipped IDE directories through a different means (specifically, a file at /tmp/ide_directories.txt).

Given the above, do you still prefer option 1 (skip just download, but still unzip the artifact(s))?

Hello.

I'd still prefer to 1st option, because it has much more advantages of management side for me.
I got to know the 1st option can consume more time to unzipping files, but I don't think It is critical.
I can put all IDEs on specific place you offered and I'd also prefer to cleaning up the unzipped directories each time.

Thank you.

[glex85]

@ChrisCarini Hello. Would you let me know how's going on this request? 2nd option also fine with me. Thanks in advance.