[erlang-patches] Print column numbers when compiling

Henrik Nord henrik@REDACTED
Tue Aug 28 09:10:48 CEST 2012


I will refetch this and let it cook in 'pu'

/Henrik

On 08/27/2012 07:12 PM, Anthony Ramine wrote:
> Hi,
>
> Please refetch my branch again, it now includes commits that patch the
> following test suites:
>
>   * erl_lint
>   * qlc
>   * erl_scan
>   * ms_transform
>
> Regards,
>
> --
> Anthony Ramine
>
> Le 25 août 2012 à 11:55, Anthony Ramine a écrit :
>
>> I've rewritten erl_lint_SUITE to handle column line numbers, doing so
>> I found a bug in the handling of disallowed_nowarn_bif_clash which I fixed
>> by creating a new "file" token attribute in erl_scan.
>>
>> Errors illegal_bin_pattern and unsized_binary_not_at_end column numbers
>> could be improved.
>>
>> Test format_warn for error format_error hasn't been rewritten yet.
>>
>> Please refetch my branch.
>>
>> 	git fetch https://github.com/nox/otp.git compile-column-numbers
>> 	https://github.com/nox/otp/tree/compile-column-numbers
>> 	https://github.com/nox/otp/compare/9bbda97f63ba4ee7cd58c266ee69af1059352189...compile-column-numbers
>>
>> The following changes since commit 9bbda97f63ba4ee7cd58c266ee69af1059352189:
>>
>>   Merge branch 'raimo/tools/remove-fprof-tuple-funs/OTP-10091' into maint (2012-05-23 16:15:58 +0200)
>>
>> are available in the git repository at:
>>
>>
>>   https://github.com/nox/otp.git compile-column-numbers
>>
>> for you to fetch changes up to 9fc46876d9e963e39d5743536a15c1f9ee97fd64:
>>
>>   Rewrite the erl_lint_SUITE to handle column numbers (2012-08-25 11:42:36 +0200)
>>
>> ----------------------------------------------------------------
>> Anthony Ramine (9):
>>       Export type erl_scan:location/0
>>       Allow setting of initial position in epp
>>       Create a new "column" option in compile
>>       Fix printing of errors with column numbers
>>       Test column number reporting in error_SUITE
>>       Fix type compile:err_info/0
>>       Fix messages ordering with column numbers
>>       Create a new "file" token attribute in erl_scan
>>       Rewrite the erl_lint_SUITE to handle column numbers
>>
>> lib/compiler/doc/src/compile.xml   |    5 +
>> lib/compiler/src/compile.erl       |   25 +-
>> lib/compiler/test/error_SUITE.erl  |   15 +-
>> lib/stdlib/doc/src/epp.xml         |    2 +
>> lib/stdlib/src/epp.erl             |   44 +-
>> lib/stdlib/src/erl_lint.erl        |   27 +-
>> lib/stdlib/src/erl_scan.erl        |   48 +-
>> lib/stdlib/test/erl_lint_SUITE.erl | 4028 +++++++++++++++++++-----------------
>> 8 files changed, 2250 insertions(+), 1944 deletions(-)
>>
>> --
>> Anthony Ramine
>>
>>
>>
>>
>> Le 24 août 2012 à 12:32, Anthony Ramine a écrit :
>>
>>> Errata: get_line/1 doesn't exist, I'm talking about get_attribute/2, sorry.
>>>
>>> --
>>> Anthony Ramine
>>>
>>> Le 24 août 2012 à 12:22, Anthony Ramine a écrit :
>>>
>>>> Most OTP (at least erl_lint, erl_parse and syntax_tools) already knows that locations can be a line or a line and a column; furthermore it is said in erl_parse itself that one should always use get_line/1 and set_line/3 instead of depending on having an integer there.
>>>>
>>>> If it is really that heavy of a change, why do we have erl_scan:set_attribute/3 and erl_scan:attributes_info/{1,2}? They obviously were coded to mitigate these issues.
>>>>
>>>> Can't this option be included as part of an experimental feature and bugs be fixed as they emerge? It's not like it is mandatory to use it, nor like there is not already experimental and dubious code in OTP (parameterized modules I'm looking at you).
>>>>
>>>> Any parse transform that break because of this change will break because it does AST manipulation wrong and their users will just have to not use the "column" option for it to continue to work.
>>>>
>>>> I'm not deterred at all from getting this patch into OTP and I would like to know exactly which tests should I update to make you guys happy.
>>>>
>>>> Regards,
>>>>
>>>> --
>>>> Anthony Ramine
>>>>
>>>> Le 22 août 2012 à 17:30, Lukas Larsson a écrit :
>>>>
>>>>> Hi,
>>>>>
>>>>> We have now looked at this patch and there are a couple of issues remaining which have to be solved.
>>>>>
>>>>> The patch changes the representation of the abstract format which has a couple of implications:
>>>>>
>>>>> * Any parse transform which uses the line number position might brake when this new option (this could include OTP parse transforms)
>>>>> * Any tool which depend on having an integer as the line number position will break (syntax tools might be one such tool)
>>>>>
>>>>> In order to mitigate this backwards incompatibility, we suggest making the line/column number an opaque datatype which is only accessible via an api and update the documentation to reflect this.
>>>>>
>>>>> In addition all of the different warnings/errors in erl_lint and friends have to be checked with the new column number so that they point to the correct position.
>>>>>
>>>>> Also the erl_lint, epp and any other suite which deal with error messages/warnings have to be updated to run tests with and without this option.
>>>>>
>>>>> I'm sure that I've forgotten some things which have to be done. This is a major change and will take some time to get correct.
>>>>>
>>>>> I hope I have not deterred you from getting this patch into the Erlang/OTP. If you have any questions about anything I will do what I can to help you.
>>>>>
>>>>> Lukas
>>>>>
>>>>> On 10/06/12 13:24, nox wrote:
>>>>>> I added a commit to correctly order locations in compile:messages_per_file/1
>>>>>> which didn't take into account column numbers.
>>>>>>
>>>>>> git fetch https://github.com/nox/otp.git compile-column-numbers
>>>>>>
>>>>>>
>>>>>> The following changes since commit 9bbda97f63ba4ee7cd58c266ee69af1059352189:
>>>>>>
>>>>>> Merge branch 'raimo/tools/remove-fprof-tuple-funs/OTP-10091' into
>>>>>> maint (2012-05-23 16:15:58 +0200)
>>>>>>
>>>>>> are available in the git repository at:
>>>>>>
>>>>>>
>>>>>> https://github.com/nox/otp.git compile-column-numbers
>>>>>>
>>>>>> for you to fetch changes up to d932ad582676f1593d917f222b1d8dec71fe88b3:
>>>>>>
>>>>>> Fix messages ordering with column numbers (2012-06-10 13:01:50 +0200)
>>>>>>
>>>>>> ----------------------------------------------------------------
>>>>>> Anthony Ramine (7):
>>>>>>    Export type erl_scan:location/0
>>>>>>    Allow setting of initial position in epp
>>>>>>    Create a new "column" option in compile
>>>>>>    Fix printing of errors with column numbers
>>>>>>    Test column number reporting in error_SUITE
>>>>>>    Fix type compile:err_info/0
>>>>>>    Fix messages ordering with column numbers
>>>>>>
>>>>>> lib/compiler/doc/src/compile.xml  |    5 +++++
>>>>>> lib/compiler/src/compile.erl      |   25 +++++++++++++++++----
>>>>>> lib/compiler/test/error_SUITE.erl |   15 +++++++++++--
>>>>>> lib/stdlib/doc/src/epp.xml        |    2 ++
>>>>>> lib/stdlib/src/epp.erl            |   44 +++++++++++++++++++++++++++++--------
>>>>>> lib/stdlib/src/erl_scan.erl       |    2 +-
>>>>>> 6 files changed, 77 insertions(+), 16 deletions(-)
>>>>>>
>>>>>> --
>>>>>> Anthony Ramine
>>>>>> _______________________________________________
>>>>>> erlang-patches mailing list
>>>>>> erlang-patches@REDACTED
>>>>>> http://erlang.org/mailman/listinfo/erlang-patches
>>>>>>
> _______________________________________________
> erlang-patches mailing list
> erlang-patches@REDACTED
> http://erlang.org/mailman/listinfo/erlang-patches

-- 
/Henrik Nord Erlang/OTP




More information about the erlang-patches mailing list