Make Chrome profile directory configurable

[llemaitre19]

Hi,

As far I know, their is no way to manually set Chrome --user-data-dir option through Indium.

If indium-chrome-use-temporary-profile is set, a new temporary profile is created at each session, and login/password, history, favorites, ... are lost.

If I try to manually add --user-data-dir=... to indium-chrome-executable I get this error :

Cannot find chrome/chromium binary (%s) in PATH

So adding a --indium-chrome-option could be usefull if there is not already another way to do that.

Note that if indium-chrome-use-temporary-profile is only present to prevent Chrome from reusing existing instance without debugging flag, just setting a fixed alternative profile directory for development allows to bypass this problem without using a new profile at each session.

[NicolasPetton]

@moki76 As a first step and workaround, would it be ok with you if the profile directory used was always the same (in a known temp directory), but not configurable?

[llemaitre19]

Yes of course, in fact I do not need the directory to be configurable for my own usage.
Thanks

[NicolasPetton]

@moki76 The next build on MELPA should fix the issue.

[llemaitre19]

Thank you for the fix. Unfortunately it is not exactly what I looking for because the fixed temporary directory is only persistent for the Emacs session, but as soon as Emacs is closed and reopened a new one is generated and the previous profile informations are lost...

[NicolasPetton]

@moki76 I exposed the variable indium-chrome-data-dir with that use-case in mind.

If you want to set a permanent profile, you can set it to something like /tmp/indium-chrome.

And if you also want to persist across reboot, you can set it to a directory within your .emacs.d:

(setq indium-chrome-data-dir (expand-file-name "indium-chrome" user-emacs-directory))
(make-directory indium-chrome-data-dir t)

Does that answer your needs? If not, I'll reopen the issue.

[llemaitre19]

My elisp bad understanding (defvar vs defcustom use) leads me to make a wrong assertion, sorry...

It works now as expected, thanks @NicolasPetton !

[NicolasPetton]

No problem, I'm glad it works!