[PATCH] stdio: Remove memory leak from multibyte convertion (BZ#25691)

[Adhemerval Zanella]

On 19/03/2020 18:14, Florian Weimer wrote:
> * Adhemerval Zanella via Libc-alpha:
> 
>> Co-authored-by: Adhemerval Zanella  <adhemerval.zanella@linaro.org>
> 
> I suggest to put the test into a separate commit.

Alright, I will split it.

> 
> Is (BZ#25691) recognized by the commit hooks as a bug reference?
> I thought it hat to be [BZ #25691] or (bug 25691).

Good question, I am not sure. I will use [BZ #25691].

> 
>> diff --git a/stdio-common/Makefile b/stdio-common/Makefile
>> index 95af0c12d7..0a8d66b846 100644
>> --- a/stdio-common/Makefile
>> +++ b/stdio-common/Makefile
>> @@ -66,6 +66,7 @@ tests := tstscanf test_rdwr test-popen tstgetln test-fseek \
>>  	 tst-scanf-round \
>>  	 tst-renameat2 tst-bz11319 tst-bz11319-fortify2 \
>>  	 scanf14a scanf16a \
>> +	 tst-printf-bz25691
>>  
>>  
>>  test-srcs = tst-unbputc tst-printf tst-printfsz-islongdouble
>> @@ -75,10 +76,12 @@ tests-special += $(objpfx)tst-unbputc.out $(objpfx)tst-printf.out \
>>  		 $(objpfx)tst-printf-bz18872-mem.out \
>>  		 $(objpfx)tst-setvbuf1-cmp.out \
>>  		 $(objpfx)tst-vfprintf-width-prec-mem.out \
>> -		 $(objpfx)tst-printfsz-islongdouble.out
>> +		 $(objpfx)tst-printfsz-islongdouble.out \
>> +		 $(objpfx)tst-printf-bz25691-mem.out
>>  generated += tst-printf-bz18872.c tst-printf-bz18872.mtrace \
>>  	     tst-printf-bz18872-mem.out \
>> -	     tst-vfprintf-width-prec.mtrace tst-vfprintf-width-prec-mem.out
>> +	     tst-vfprintf-width-prec.mtrace tst-vfprintf-width-prec-mem.out \
>> +	     tst-printf-bz25691.mtrace tst-printf-bz25691-mem.out
>>  endif
>>  
>>  include ../Rules
>> @@ -100,6 +103,8 @@ endif
>>  tst-printf-bz18872-ENV = MALLOC_TRACE=$(objpfx)tst-printf-bz18872.mtrace
>>  tst-vfprintf-width-prec-ENV = \
>>    MALLOC_TRACE=$(objpfx)tst-vfprintf-width-prec.mtrace
>> +tst-printf-bz25691-ENV = \
>> +  MALLOC_TRACE=$(objpfx)tst-printf-bz25691.mtrace
>>  
>>  $(objpfx)tst-unbputc.out: tst-unbputc.sh $(objpfx)tst-unbputc
>>  	$(SHELL) $< $(common-objpfx) '$(test-program-prefix)' > $@; \
> 
> Ah, there's a pattern rule matching *-mem.out.  Looks okay.
> 
> 
>> diff --git a/stdio-common/tst-printf-bz25691.c b/stdio-common/tst-printf-bz25691.c
> 
>> +    /* The input size value is to force a heap allocation on temporary
>> +       buffer.  */
> 
>> +    /* The input size value is to force a heap allocation on temporary
>> +       buffer.  */
> 
> I think these comments are now outdated.  Maybe add “(in the old
> implementation)”?

kk.

> 
>> diff --git a/stdio-common/vfprintf-internal.c b/stdio-common/vfprintf-internal.c
>> index 3be92d4b6e..d34e31746c 100644
>> --- a/stdio-common/vfprintf-internal.c
>> +++ b/stdio-common/vfprintf-internal.c
> 
> That part still looks okay to me.  Thanks.
>