[Tarantool-patches] [PATCH luajit v4 2/4] memprof: refactor location parsing

[Igor Munkin]

Misha,

Thanks for the patch! Everything is fine in general, but consider a few
nits below, please.

On 29.09.21, Mikhail Shishatskiy wrote:
> As it is hard to keep function `id_location` easily usable in the
> function `parse_location` with a growing number of fields in the
> <loc> table, the code is refactored to make it more understandable.
> 
> The API of <utils/symtab.lua> module changed with adding a function
> `id(loc)` returning the same id-string, as `id_location` did. This
> function is useful to "stringify" the location and use it as a key
> in the key/value storage.

It's worth to also mention that the function need to be used in other
modules, hence you've made it public in symtab.lua.

> 
> Part of tarantool/tarantool#5814
> ---
> 
> Issue: https://github.com/tarantool/tarantool/issues/5814
> Branch: https://github.com/tarantool/luajit/tree/shishqa/gh-5814-group-allocations-on-trace-by-trace-number
> CI: https://github.com/tarantool/tarantool/tree/shishqa/gh-5814-group-allocations-on-trace-by-trace-number
> 
>  tools/memprof/parse.lua | 26 +++++++++++++-------------
>  tools/utils/symtab.lua  |  4 ++++
>  2 files changed, 17 insertions(+), 13 deletions(-)
> 
> diff --git a/tools/memprof/parse.lua b/tools/memprof/parse.lua
> index 12e2758f..34ff8aef 100644
> --- a/tools/memprof/parse.lua
> +++ b/tools/memprof/parse.lua
> @@ -8,6 +8,8 @@ local bit = require "bit"
>  local band = bit.band
>  local lshift = bit.lshift
>  
> +local symtab = require "utils.symtab"

Minor: Why here? It's better either to group all <requires> in the one
place (worse, IMHO) or just add a new <require> just prior to the part
with constants i.e. below <string.format> caching (better, IMHO).

> +
>  local string_format = string.format
>  
>  local LJM_MAGIC = "ljm"
> @@ -59,22 +61,20 @@ local function link_to_previous(heap_chunk, e, nsize)
>    end
>  end
>  
> -local function id_location(addr, line)
> -  return string_format("f%#xl%d", addr, line), {
> -    addr = addr,
> -    line = line,
> -  }
> -end
> -
>  local function parse_location(reader, asource)
> -  if asource == ASOURCE_INT then
> -    return id_location(0, 0)
> -  elseif asource == ASOURCE_CFUNC then
> -    return id_location(reader:read_uleb128(), 0)
> +  local loc = {
> +    addr = 0,
> +    line = 0,
> +  }
> +  if asource == ASOURCE_CFUNC then
> +    loc.addr = reader:read_uleb128()
>    elseif asource == ASOURCE_LFUNC then
> -    return id_location(reader:read_uleb128(), reader:read_uleb128())
> +    loc.addr = reader:read_uleb128()
> +    loc.line = reader:read_uleb128()
> +  elseif asource ~= ASOURCE_INT then

Minor: It's better to write a separate condition for <ASOURCE_INT> and a
separate block for so-called "default" (i.e. plain "else" block) with an
error. So it looks kinda the following:

| if asource == ASOURCE_INT then
|   -- Do nothing. Add a suppression for luacheck here.
| elseif asource == ASOURCE_CFUNC then
|   loc.addr = reader:read_uleb128()
| elseif asource == ASOURCE_LFUNC then
|   loc.addr = reader:read_uleb128()
|   loc.line = reader:read_uleb128()
| else
|   error("Unknown asource "..asource)
| end

If you were confused by luacheck, don't worry: just add a suppression
for this block.

> +    error("Unknown asource "..asource)
>    end
> -  error("Unknown asource "..asource)
> +  return symtab.id(loc), loc
>  end
>  
>  local function parse_alloc(reader, asource, events, heap)

<snipped>

>  
> -- 
> 2.33.0
> 

-- 
Best regards,
IM