app_rpt.c: Refactor handle_link_data() and handle_remote_data() #474
Conversation
| if (l->mode > 1) | ||
| continue; /* dont report local modes */ | ||
| if (l->linklist[0]) { | ||
| buffer_size += (strlen(l->linklist) + strlen(l->name) + 3); /*+3: 2 for the commas, 1 for mode, 1 extra as the first pass has no comma*/ |
There was a problem hiding this comment.
I'm confused about needing "3". You need 1 for the mode and 1 more for either a comma (if more nodes follow) or the end-of-line (if this was the last node).
Note: if looks like RPT_ALINKS includes an extra character so we do need 3 (mode, keyed/unkeyed, and comma/EOL)
There was a problem hiding this comment.
There is a comma appended to the end of the full list except for the first pass:
T123, "stuff from the linklist" , T345, "stuff from the next link list" -> 2 commas if more than 1 link (between T123 and T345) other wise only 1 comma.
Maybe my comment about 1 extra comma for the first pass is not clear enough, we will be 1 char larger than we need.
update sprintf to snprintf rpt_link.c: Calculate buffer size and malloc buffers. add function __get_buffer_size() to calculate buffer size required. add node count to __mklinklist() add macro for OBUFSIZE and BUFSIZE calculation elminate finddelim() call as __mklinklist now counts links
mkmer
force-pushed
the
Refactor-link-message
branch
from
ff1ee3c to
9f1050a
Compare
|
I will squash into a single commit once we are done reviewing it. |
fix up count in __mklinklist
|
Fixing up finddelim may not be possible. I think some of the instances are searching through non null terminated values (but I can't be sure right now). With numbers like 3 in the limit while strs[100]. |
|
I'll take a look at the finddelim call sites tomorrow. |
|
FYI : in So, when needed, we can use |
|
Looks like chan_echolink.c, chan_tlb.c, and chan_voter have their own copy/version of |
I saw that when I was making the move. |
mkmer
added
the
code quality
mkmer
force-pushed
the
Refactor-link-message
branch
from
0d63fd5 to
76c4416
Compare
This is a follow on to PR #470 which should not be delayed. Intent is to clean up the stuff I notice while hunting down the truncated message.
Eliminate copy of malloc
char*to fixed lengthchar*update
sprintf()tosnprintf()rpt_link.c: Calculate buffer size and malloc buffers. add function __get_buffer_size() to calculate buffer size required. add node count to __mklinklist()
add macro for OBUFSIZE and BUFSIZE calculation
eliminate
finddelim()call as__mklinklist()now counts links.