New issue
Advanced search Search tips

Issue 895237 link

Starred by 2 users
Status: WontFix
Owner:
agrieve@chromium.org
Closed: Oct 15
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 3
Type: Bug



Sign in to add a comment

Supersize diff shows positive numbers as 0

Project Member Reported by brat...@opera.com, Oct 15 
I moved a string from data to rodata and got this diff:

Section Legend: t=.text, r=.rodata, R=.data.rel.ro, d=.data, b=.bss, x=.dex, m=.dex.method, p=.pak.translations, P=.pak.nontranslated, o=.other
  Index | Running Total | Section@Address | Δ PSS (Δ size_without_padding) | Path
  ------------------------------------------------------------
  - 0)       -137 (3425.0%) d@0x0        -137 (137->0)      sql/recovery.cc
                 sql::Recovery::BeginRecoverDatabase::kCreateMetaItemsSql
  + 1)          0 (-0.0%) r@0x711377   137 (0->137)       sql/recovery.cc
                 sql::Recovery::BeginRecoverDatabase::kCreateMetaItemsSql
  + 2)        -23 (575.0%) d@0x0        -23 (0->0)         {no path}
                 ** aggregate padding of diff'ed symbols
  + 3)          0 (-0.0%) r@0x0        23 (0->0)          {no path}
                 ** aggregate padding of diff'ed symbols
  ~ 4)         -3 (75.0%) o@0x0        -3 (81194->81191)  $APK/META-INF/CHROMIUM.SF
                 META-INF/CHROMIUM.SF
  ~ 5)         -4 (100.0%) o@0x0        -1 (79210->79209)  $APK/META-INF/MANIFEST.MF
                 META-INF/MANIFEST.MF

The second column should have said +137 on the second line I think, but instead it says 0. Similar with the fourth line that should have said +23.


This is from 
https://logs.chromium.org/logs/chromium/buildbucket/cr-buildbucket.appspot.com/8932869482359676064/+/steps/Show_Supersize_Diff/0/logs/s_____View_Supersize_Diff____/0
coming from
https://ci.chromium.org/p/chromium/builders/luci.chromium.try/android-binary-size/74870

Comment 1 by agrieve@chromium.org, Oct 15

Status: WontFix (was: Untriaged)
The second column is "Running Total", so it was zero because -137 + 137 = 0.

The per-symbol size is the Δ PSS column.

I'm certainly open to suggestions for how to improve this UI. These symbol diffs are often hard to read.

Comment 2 by brat...@opera.com, Oct 15 

Ah, that makes sense! I happened to interpret it as "change that matters for running binaries" or something equally nonsensically.

I guess accumulated instead of running would mean the same but that is a longer word so not a clear improvement. 

Another suggestion would be to align headers and data better. Can use a multi line header to make some fields less wide. But not very important. All the data is there to see.
►

Sign in to add a comment