Skip to content

Implement issue #42: use upstream RapidJSON v1.1#46

Merged
kenrobbins merged 11 commits into
python-rapidjson:masterfrom
lelit:rj-v1.1-as-submodule-v2
Dec 8, 2016
Merged

Implement issue #42: use upstream RapidJSON v1.1#46
kenrobbins merged 11 commits into
python-rapidjson:masterfrom
lelit:rj-v1.1-as-submodule-v2

Conversation

@lelit

@lelit lelit commented Aug 31, 2016

Copy link
Copy Markdown
Contributor

This adapts the module to use upstream RapidJSON v1.1, using a submodule referencing the official tag.

Upon first checkout (or after the merge), you should execute a git submodule update --init. Alternatively you can use git clone --recursive to do the checkout and the submodule initialization in a single step.

lelit added 11 commits August 31, 2016 13:16
@lelit
First steps toward using current upstream RapidJSON 1.1
e31b836
@lelit
Fix tests, actually dump the Decimal instance
42006c7
@lelit
Fix case of generated exagesimal UTF-8 codes
633f3b5
@lelit
Use the RawNumber API to load numbers
bfe97c8 
This allows loading the (virtually) unlimited range of Python's integers
and floats (when using the Decimal data type). Since the conversion from
the textual representation to the numeric is now entirely done by the
Python runtime, the flag precise_float does not makes sense anymore and
it's been removed.
@lelit
Replace pytest's deprecated section name
17f2ee6
@lelit
Remove tests and documentation related to the option precise_float
7ebe742
@lelit
Adapt doctest, upstream RapidJSON uses uppercase letters for hex codes
6b4991e
@lelit
Remove unused variable
941f774
@lelit
Remove unused import
c512a6c
@lelit
Fix recognition of float value, ignoring the minus sign
43b54ea
@lelit
Fix grouping of benchmarks results
6b4f27b 
Use the same grouping rule both when comparing with other engines or
not.
@lelit

lelit commented Sep 1, 2016

Copy link
Copy Markdown
Contributor Author

FYI, I'm leaving for a week of vacation, so I won't be reachable until sept 11...

@lelit

lelit commented Sep 13, 2016

Copy link
Copy Markdown
Contributor Author

In lelit/python-rapidjson@fa63d17 I added a native_numbers option to dumps() and loads() to force use of the underlying C library arithmetic, somewhat faster, in particular for the latter function:

--------------------------------------------------------------------------- benchmark '512 numbers array: deserialize': 2 tests ----------------------------------------------------------------------------
Name (time in us)                                             Min                 Max                Mean            StdDev              Median               IQR            Outliers(*)  Rounds  Iterations
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_loads_numbers[Native numbers-512 numbers array]      30.2480 (1.0)       60.2280 (1.0)       30.7363 (1.0)      0.7514 (1.0)       30.6820 (1.0)      0.1680 (1.0)          400;902   25039           1
test_loads_numbers[Wide numbers-512 numbers array]       113.2490 (3.74)     177.5800 (2.95)     114.2523 (3.72)     2.3087 (3.07)     113.9890 (3.72)     0.2550 (1.52)         209;440    7326           1
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------- benchmark '512 numbers array: serialize': 2 tests ---------------------------------------------------------------------------
Name (time in us)                                            Min                 Max               Mean            StdDev             Median               IQR            Outliers(*)  Rounds  Iterations
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_dumps_numbers[Native numbers-512 numbers array]     50.6330 (1.0)      115.7490 (1.06)     51.5841 (1.0)      1.8838 (1.56)     51.3230 (1.0)      0.2320 (1.10)        330;1624   16585           1
test_dumps_numbers[Wide numbers-512 numbers array]       66.5600 (1.31)     109.6080 (1.0)      67.2445 (1.30)     1.2087 (1.0)      67.1300 (1.31)     0.2100 (1.0)          255;618   12171           1
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Let me know if it seems ok to merge that here (in particular wrt the option name).

@lelit

lelit commented Sep 27, 2016

Copy link
Copy Markdown
Contributor Author

@kenrobbins: is there anything left I can/should do to promote this PR into your master branch?

@kenrobbins kenrobbins merged commit 5405613 into python-rapidjson:master Dec 8, 2016
@lelit lelit mentioned this pull request Dec 9, 2016
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@lelit @kenrobbins