843343 - Version-numbers clipped on Firefox Release Dashboard "Major Version Breakdown", after toggling between Share(ADI) and Downloads

Status: RESOLVED INCOMPLETE

(Core :: SVG, defect)

Description

[Daniel Holbert [:dholbert]]

Attachment: screenshot

STR:
 1. Visit https://metrics.mozilla.com/pentaho/content/pentaho-cdf-dd/Render?solution=metrics2&path=%2Ftwitter%2Ffirefox&file=fx_release.wcdf
 2. Log in with LDAP. (requires special access, granted by metrics team or IT)
 3. On the "Major Version Breakdown" graph, press the "Share (ADI)" & "Downloads" buttons, to toggle between those states.

ACTUAL RESULTS:
 The version-number labels get clipped. Looks like an invalidation issue -- if I change tabs or click the titlebar, the bug fixes itself.

m-c regression range:
{
Last good nightly: 2013-01-16
First bad nightly: 2013-01-17
Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=d8be4bc4fba8&tochange=712eca11a04e
}

Comment 1

[Daniel Holbert [:dholbert]]

On mozilla-inbound, the regression range is the same datestamps.
* 20130116023801 is good
* 20130117001951 is bad
which yields this regression range:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=e201a9901088&tochange=0b0765410460

Intersecting that with the m-c regression range yields this slightly-smaller range:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=e201a9901088&tochange=712eca11a04e

Also: for reference, my current nightly UA is:
  Mozilla/5.0 (X11; Linux x86_64; rv:22.0) Gecko/20130220 Firefox/22.0

Comment 2

[Daniel Holbert [:dholbert]]

So hg bisect narrowed this down to this cset:
  https://hg.mozilla.org/mozilla-central/rev/93822863127e
  Bug 719659 - Add -std=gnu99 to CFLAGS [...]

That made no sense, so I double-checked w/ targeted builds of that cset & its parent.  The parent is fine; that cset shows this bug.

Waldo suggests in IRC that this is likely an existing code bug that's being tickled by different compile options (and I tend to agree).

Comment 3

[Daniel Holbert [:dholbert]]

The "-std=gnu99" part of 93822863127e is sufficient to trigger this. (without the "-fgnu89-inline" part)

Comment 4

[Robert Longson [:longsonr]]

The site is gone and I imagine the compiler is much different in 10 years.