binary-size issues

[Fergal Daly]

Fuchsia folks,

this change repeatedly failed the fuchsia-binary-size bot but then it started passing, so I thought it must be something transient. I took a closer look and TBH, I don't see much difference between the pass and the fail. The baseline in the pass was a little higher so maybe that helped.

According to bloaty this mostly seems to be TFlite code that was previously not included into fuchsia. So I think this i a real binary size increase.

What should we do? This is for a new web-platform API (language detection and soon translation). Should we #ifdef this out for fuchsia?

F

[Zijie He]

Thank you for reaching out.

The binary size check includes multiple steps in the middle, and since it uses compression, even a bit may impact the end result. And as you notice, the baseline also changes due to other submitted changes, we calculate the diff in the tryjob.

In general we do not want fuchsia to be different from other platforms, unless something totally irrelevant to the fuchsia components we are shipping (devtools UI is a good example). So if it's a web-platform API, I'd be very happy to *not* exclude fuchsia, both for the external users and the maintenance cost. Since we are only talking about some 24K, you may ignore the error with `Fuchsia-Binary-Size={reason}` tag next time (See the instruction).

But it would be great if you can CC me on the changes if you need to bypass the checks.

--
You received this message because you are subscribed to the Google Groups "fuchsia-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fuchsia-dev...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/fuchsia-dev/CAAozHLk0pkQ6G2sfUuTcskqgZro2aAredWUZ4ayDLv5%3DTsc0Kw%40mail.gmail.com.