Net: Turn some methods and params/variables const #9659

pull jtimon wants to merge 3 commits into bitcoin:master from jtimon:0.14-net-more-const changing 5 files +19 −19

jtimon commented at 10:51 PM on January 31, 2017: contributor

Also remove currently unused interruptMsgProc param from SendMessages. I guess some squashing can be done (but I'm worried that interruptMsgProc is planned to be used in SendMessages later).
Net: CConnman: Make some methods const d45955fa09
Net: Make CNetMsgMaker more const fc7f2ffad4
Net: pass interruptMsgProc as const where possible 0729102f99
jtimon cross-referenced this on Jan 31, 2017 from issue Net: Divide ProcessMessage in smaller functions by jtimon
jtimon cross-referenced this on Jan 31, 2017 from issue Net: Trivial-review: Make SendMessages easier to review by jtimon

in src/net.cpp:None in 9c0d42177a outdated

1870 | @@ -1871,7 +1871,7 @@ void CConnman::ThreadMessageHandler()
1871 |              // Send messages
1872 |              {
1873 |                  LOCK(pnode->cs_sendProcessing);
1874 | -                GetNodeSignals().SendMessages(pnode, *this, flagInterruptMsgProc);
1875 | +                GetNodeSignals().SendMessages(pnode, *this);

TheBlueMatt commented at 11:52 PM on January 31, 2017:

Even if we dont use it now, I'm not sure we want to get rid of passing flagInterruptMsgProc through...there is a strong case for using it in 0.15.

TheBlueMatt commented at 12:03 AM on February 1, 2017: contributor

General concept ack, with the exception of removing the flagInterruptMsg pass (but please do make it const)
jtimon commented at 1:18 AM on February 1, 2017: contributor

Great, it seems I was right on leaving the last commit separated, I'll leave it out.
jtimon force-pushed on Feb 1, 2017
jonasschnelli added the label Refactoring on Feb 1, 2017
in src/net.cpp:None in 0729102f99
```
 753 | @@ -754,7 +754,7 @@ const uint256& CNetMessage::GetMessageHash() const
 754 |  
 755 |  
 756 |  // requires LOCK(cs_vSend)
 757 | -size_t CConnman::SocketSendData(CNode *pnode)
```
TheBlueMatt commented at 9:41 PM on February 1, 2017:

Hmm...while technicall const for the CConnman, I'm not sure this "should be const" given that it is writing data to the peer itself...

jtimon commented at 11:10 PM on February 2, 2017:

Well, writing to the peer itself here is writing to pnode, which isn't const, no?
TheBlueMatt commented at 10:31 PM on February 1, 2017: contributor

utACK 0729102f99241b9716456ad40bf778c09f650b1b (+/- SocketSendData being made const).
laanwj commented at 1:34 PM on February 6, 2017: member

utACK 0729102
laanwj merged this on Feb 6, 2017
laanwj closed this on Feb 6, 2017
laanwj referenced this in commit 09e0c28f85 on Feb 6, 2017
codablock referenced this in commit 7026c35208 on Jan 19, 2018
codablock referenced this in commit 3f1929fb0d on Jan 23, 2018
gladcow cross-referenced this on Mar 1, 2018 from issue Backport compact blocks functionality from bitcoin by gladcow
gladcow referenced this in commit 87387c4306 on Mar 8, 2018
gladcow referenced this in commit 80c2f36c5d on Mar 13, 2018
dagurval cross-referenced this on Jun 8, 2018 from issue More net cherries by dagurval
andvgal referenced this in commit 423788ba89 on Jan 6, 2019
CryptoCentric referenced this in commit 1a1424842b on Feb 27, 2019
bitcoin locked this on Sep 8, 2021

Contributors

Labels

Linked (view graph)

#9579 Net: Trivial-review: Make SendMessages easier to review #9608 Net: Divide ProcessMessage in smaller functions