Thread: Uninitialized var utilized (src/backend/tsearch/spell.c)
At function NIImportAffixes (src/backend/tsearch/spell.c).
If option "flag" is not handled, variable char flag[BUFSIZE] will remain uninitialized.
regards,
Ranier Vilela
Attachment
> On 9 Oct 2020, at 14:36, Ranier Vilela <ranier.vf@gmail.com> wrote: > At function NIImportAffixes (src/backend/tsearch/spell.c). > > If option "flag" is not handled, variable char flag[BUFSIZE] will remain uninitialized. To help reviewers, your report should contain an explanation of when that can happen. cheers ./daniel
Em sex., 9 de out. de 2020 às 11:08, Daniel Gustafsson <daniel@yesql.se> escreveu:
> On 9 Oct 2020, at 14:36, Ranier Vilela <ranier.vf@gmail.com> wrote:
> At function NIImportAffixes (src/backend/tsearch/spell.c).
>
> If option "flag" is not handled, variable char flag[BUFSIZE] will remain uninitialized.
To help reviewers, your report should contain an explanation of when that can
happen.
When option "flag" is not handled.
if (STRNCMP(pstr, "flag") == 0)
regards,
Ranier Vilela
Ranier Vilela <ranier.vf@gmail.com> writes: > Em sex., 9 de out. de 2020 às 11:08, Daniel Gustafsson <daniel@yesql.se> > escreveu: >> To help reviewers, your report should contain an explanation of when that >> can happen. > When option "flag" is not handled. > if (STRNCMP(pstr, "flag") == 0) I think what he means is that if the file contains no "flag" command before an affix entry then then we would arrive at NIAddAffix with an undefined flag buffer. That's illegal syntax according to a quick scan of the ispell(5) man page, which explains the lack of complaints; but it might be worth guarding against. Aside from failing to initialize some variables that need it, it looks to me like NIImportAffixes is uselessly initializing some variables that don't need it. I'd also be inclined to figure out which values are actually meant to be carried across lines, and declare the ones that aren't inside the loop, just for clarity. regards, tom lane
Em sex., 9 de out. de 2020 às 11:37, Tom Lane <tgl@sss.pgh.pa.us> escreveu:
Ranier Vilela <ranier.vf@gmail.com> writes:
> Em sex., 9 de out. de 2020 às 11:08, Daniel Gustafsson <daniel@yesql.se>
> escreveu:
>> To help reviewers, your report should contain an explanation of when that
>> can happen.
> When option "flag" is not handled.
> if (STRNCMP(pstr, "flag") == 0)
I think what he means is that if the file contains no "flag" command
before an affix entry then then we would arrive at NIAddAffix with an
undefined flag buffer. That's illegal syntax according to a quick scan
of the ispell(5) man page, which explains the lack of complaints; but
it might be worth guarding against.
Aside from failing to initialize some variables that need it, it looks to
me like NIImportAffixes is uselessly initializing some variables that
don't need it. I'd also be inclined to figure out which values are
actually meant to be carried across lines, and declare the ones that
aren't inside the loop, just for clarity.
Thanks Tom, for the great explanation.
regards,
Ranier Vilela