Fix memory corruption crash

[H3rnand3zzz]

Read commit messages. For crash payload, contact me.

[sjaeckel]

Thanks for the PR!

I agree with the changes, but I don't see a potential double free here.

The issue we're seeing is an out-of-bounds access of the data returned by gpgme_data_release_and_get_mem(). The call only allocates len bytes, but profanity is accessing len+1 bytes.

Didn't we already have a similar issue in the same part of the code not long ago? Maybe those parts should be reviewed!?

Please update the commit message, PR title&description accordingly.

[H3rnand3zzz]

Thanks for the PR!

I agree with the changes, but I don't see a potential double free here.

The issue we're seeing is an out-of-bounds access of the data returned by gpgme_data_release_and_get_mem(). The call only allocates len bytes, but profanity is accessing len+1 bytes.

Didn't we already have a similar issue in the same part of the code not long ago? Maybe those parts should be reviewed!?

Please update the commit message, PR title&description accordingly.

Thank you for the review. In fact, there was a consistent crash caused by some payload, which glib claimed to be "double free or corruption". I don't see double free here as well, so probably it's a false positive or rarther "memory corruption" issue, which breaks the execution anyway. I renamed to "memory corruption".

Line from where of GDB:
#5 0x00007f48a05e8657 in malloc_printerr (str=str@entry=0x7f48a06f45f8 "double free or corruption (!prev)") at malloc.c:5651

[sjaeckel]

I renamed to "memory corruption".

please also reword the commit message. GH is ephemeral, the Git history will most likely survive this service.

[H3rnand3zzz]

I renamed to "memory corruption".

please also reword the commit message. GH is ephemeral, the Git history will most likely survive this service.

I thought that I did, probably messed it up on some stage :) Fixed, thanks.

[jubalh]

Mmhh now we have a merge commit in there. Can you just rebase on master instead?
Also I would prefer if the 2nd commit (re cleanup) could be a bit more descriptive in the heading. Something like Cleanup p_ox_gpg_decrypt or something like that would already suffice. Then when skimming through the list of changes when looking for something specific I don't need to click on the entry to see which part is actually cleaned up/touched.

[jubalh]

Thanks!