Add CHIP_ERROR object methods

[kpschoedel]

Problem

Having CHIP_ERROR be a class type provides type safety and the option to
trace the source of errors. Since all platforms now use the class type,
transitional code that provided for CHIP_ERROR to be either a class or
integer type is no longer required and can be removed, which simplifies
the class and some notation for using errors.

Change overview

• Remove the CHIP_CONFIG_ERROR_CLASS configuration option.
• Add instance methods to replace the static ChipError functions, which
  were present to allow transitional integer overloads. To avoid
  breaking PRs in flight, the static methods remain present for now,
  marked “DO NOT USE”; they will be removed in a final cleanup PR.

Testing

Existing tests should confirm no change to functionality.

[andy31415]

@kpschoedel - merge conflicts

[github-actions]

Size increase report for "esp32-example-build" from 8511730

File	Section	File	VM
chip-ipv6only-app.elf	.flash.text	-172	-172
chip-all-clusters-app.elf	.flash.text	-36	-36
chip-lock-app.elf	.flash.text	-64	-64
chip-temperature-measurement-app.elf	.flash.text	-60	-60

Full report output

BLOAT REPORT

Files found only in the build output:
    report.csv

Comparing ./master_artifact/chip-ipv6only-app.elf and ./pull_artifact/chip-ipv6only-app.elf:

sections,vmsize,filesize
.debug_info,0,14460
.debug_str,0,666
[Unmapped],0,172
.debug_abbrev,0,116
.debug_line,0,10
.flash.text,-172,-172

Comparing ./master_artifact/chip-pigweed-app.elf and ./pull_artifact/chip-pigweed-app.elf:

sections,vmsize,filesize
.debug_info,0,14460
.debug_str,0,674
.debug_abbrev,0,116
.debug_line,0,10

Comparing ./master_artifact/chip-persistent-storage.elf and ./pull_artifact/chip-persistent-storage.elf:

sections,vmsize,filesize
.debug_info,0,2570
.debug_str,0,424
.debug_abbrev,0,141
.debug_loc,0,17
.debug_line,0,-4

Comparing ./master_artifact/chip-all-clusters-app.elf and ./pull_artifact/chip-all-clusters-app.elf:

sections,vmsize,filesize
.debug_info,0,85029
.debug_abbrev,0,3565
.debug_str,0,417
[Unmapped],0,36
.debug_loc,0,22
.riscv.attributes,0,2
.shstrtab,0,2
.strtab,0,-2
.debug_frame,0,-12
.flash.text,-36,-36
.debug_ranges,0,-40
.debug_line,0,-987

Comparing ./master_artifact/chip-shell.elf and ./pull_artifact/chip-shell.elf:

sections,vmsize,filesize
.debug_info,0,37588
.debug_abbrev,0,2619
.debug_str,0,428
.strtab,0,48
.debug_ranges,0,-8
.debug_loc,0,-55
.debug_line,0,-160

Comparing ./master_artifact/chip-lock-app.elf and ./pull_artifact/chip-lock-app.elf:

sections,vmsize,filesize
.debug_info,0,1016768
.debug_abbrev,0,15787
.debug_line,0,314
[Unmapped],0,64
.strtab,0,1
.shstrtab,0,-1
.debug_ranges,0,-8
.flash.text,-64,-64
.debug_loc,0,-363
.debug_str,0,-1306

Comparing ./master_artifact/chip-temperature-measurement-app.elf and ./pull_artifact/chip-temperature-measurement-app.elf:

sections,vmsize,filesize
.debug_info,0,66531
.debug_abbrev,0,2897
.debug_str,0,420
[Unmapped],0,60
.debug_ranges,0,-8
.flash.text,-60,-60
.debug_line,0,-308
.debug_loc,0,-364

[bzbarsky-apple]

So switching to class-type errors was not a problem in terms of codesize? I guess as long as CHIP_CONFIG_ERROR_SOURCE is false the codesize is fine? @kpschoedel

[kpschoedel]

So switching to class-type errors was not a problem in terms of codesize? I guess as long as CHIP_CONFIG_ERROR_SOURCE is false the codesize is fine? @kpschoedel

Yes, I checked that before embarking on this conversion. With all the relevant things inline the code for a class-wrapped integer is generally identical to a bare one.

[kpschoedel]

I'll now start monitoring PRs for the obsolete functions and submit a final cleanup when it's safe.