599361 - Stop uploading symbols to the symbol server for mobile desktop builds

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[(not currently active) Ted Mielczarek]

Currently we are uploading crashreporter symbols for all the mobile desktop builds. Also, mobile builds against mozilla-central, e10s, and tracemonkey. This means we're storing 3x as many symbol files as Firefox, for builds that I would guess nobody is using as their daily browser. We should just stop uploading these symbols, I don't think they're contributing usefully to crash reporting or development. (Actual crashes on Windows/Mac/Linux are likely to be picked up in the Firefox crash reports anyway. We'd continue to get crash reports on Maemo, which is what matters, and eventually Android hopefully.)

Comment 1

[John Ford [:jhford] CET/CEST Berlin Time]

This is to stop uploading symbols to the symbol server.

Comment 2

[John Ford [:jhford] CET/CEST Berlin Time]

Attachment: turn off windows/mac symbols

this patch disables windows and mac symbol generation and uploading.

The flags to enable/disable uploading symbols also controls whether we build them.  We can change that, but will require a larger patch and testing.

Comment 3

[John Ford [:jhford] CET/CEST Berlin Time]

Attachment: and for project branches

Comment 4

[(not currently active) Ted Mielczarek]

If we're not uploading them, not building them doesn't matter. I think you also want to disable linux-i686 here.

Comment 5

[Chris AtLee [:catlee]]

you touched it, you get it!

Comment 6

[John Ford [:jhford] CET/CEST Berlin Time]

(In reply to comment #4)
> If we're not uploading them, not building them doesn't matter. I think you also
> want to disable linux-i686 here.

turning it off for linux-i686 means that we stop building *and* uploading to stage, which I need for bug 511282.

If we can limit scope for now to stopping windows and mac, we can land something sooner and remove 2/3 of the extraneous uploads.  If we want to split building and uploading symbols, that will be a slightly larger patch.

Comment 7

[(not currently active) Ted Mielczarek]

Ah, makes sense. Seems fine.

Comment 8

[Aki Sasaki (not active)]

The rest of this should have been handled in bug 607660.