Doesnt work in Electron apps that have used .asar to package themselves

[pfrazee]

See electron/electron#2708

The startDaemon() call never returns error or success. I'm fairly sure the issue is as in 2708: the relative paths fail.

[pfrazee]

I solved this by not using .asar. I havent figured out why Electron folks bother with the .asar, so I'm not sure what the downside is for my change.

[jbenet]

(My comment is Off topic)

bother with .asar

I think it's supposed to be a bit better than tar, less error prone (stream
cutoff errors) and packed, etc. similar to our desires for .car

On Wed, Aug 10, 2016 at 17:45 Paul Frazee notifications@github.com wrote:

I solved this by not using .asar. I havent figured out why Electron folks
bother with the .asar, so I'm not sure what the downside is for my change.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#94 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAIcoUDfbVh6U8Qmrem3iiS97D21ytfYks5qekZ5gaJpZM4Jhm6p
.

[pfrazee]

Yeah but there's no compression in .asar, and osx is already bundled into a
dmg. Maybe it's for windows and linux? not sure

On Wed, Aug 10, 2016 at 8:47 PM, Juan Benet notifications@github.com
wrote:

(My comment is Off topic)

bother with .asar

I think it's supposed to be a bit better than tar, less error prone (stream
cutoff errors) and packed, etc. similar to our desires for .car

On Wed, Aug 10, 2016 at 17:45 Paul Frazee notifications@github.com
wrote:

I solved this by not using .asar. I havent figured out why Electron folks
bother with the .asar, so I'm not sure what the downside is for my
change.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#94 (comment),
or mute the thread
<https://github.com/notifications/unsubscribe-auth/
AAIcoUDfbVh6U8Qmrem3iiS97D21ytfYks5qekZ5gaJpZM4Jhm6p>
.

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#94 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABNhU2A78kzo4pA_gWi8JuWB6TkZLmJpks5qen85gaJpZM4Jhm6p
.

[pfrazee]

Aha:

To mitigate issues around long path names on Windows, slightly speed up require and conceal your source code from cursory inspection, you can choose to package your app into an asar archive with little changes to your source code.

http://electron.atom.io/docs/tutorial/application-packaging/

[pfrazee]

It turns out, they'd even documented this specific issue: http://electron.atom.io/docs/tutorial/application-packaging/#executing-binaries-inside-asar-archive

[daviddias]

@haadcode do you run into this issue with orbit? How did you solve it?

[haadcode]

@diasdavid fixed it in the same way as @pfrazee did: not using asar.

[thisconnect]

It seems that the solution is to exclude go-ipfs-dep from node_modules during packaging. I got it working with electron-packager + the modification in PR #155 this PR is not a final solution yet. Only tested on MacOS

electron-packager ./ --overwrite --out ./dist --asar.unpackDir=node_modules/go-ipfs-dep/**

[daviddias]

@thisconnect thanks for working on a interim solution and documenting it on the README 👍 https://github.com/ipfs/js-ipfsd-ctl#packaging

[thisconnect]

Can this issue be closed?